Evidence · AFNI

3dvolreg

3dvolreg moves every volume of a series onto a base volume by a rigid-body transformation. fMRIPrep’s BOLD reference is the median of 20 of a run’s volumes once 3dvolreg has aligned them. LARMORX has two ports of it, both in larmorx itself. The replica gives AFNI 26.3.05’s bytes, every file 3dvolreg writes, in all 137 compared cases of its record that write outputs, out of 190, fMRIPrep’s calls included. The clean room is held to its specification’s limits instead of AFNI’s bits, and stays within them in all 185 compared cases of its record, out of 251. It runs fMRIPrep’s calls 2.8 to 5.2 times as fast as AFNI on one thread, and 9 to 15 times on 14.

Ports
  • replica larmorx Apache-2.0, with CC BY 4.0's attribution for the Medical College of Wisconsin's code: translated from AFNI 26.3.05's source; in pip install larmorx
  • clean room larmorx Apache-2.0, crate larmorx-afni: written from specs/3dvolreg.md, AFNI's help, Cox and Jesmanowicz's paper (1999) and runs of AFNI 26.3.05 and of the replica, used as black boxes (2026-11-08)
Status
replica validated (marked before the statuses' definition); clean room experimental
Records
the replica's: standard comparison (AFNI 26.3.05) · the clean room's: standard comparison (AFNI 26.3.05) · benchmark afni-volreg
Reference
3dvolreg · API page afni-volreg

What is compared

3dvolreg moves every volume of a series onto a base volume (the first, by default) with a rigid-body transformation: three rotations and three shifts. For each volume it solves a weighted least-squares problem linearised about the current estimate: the volume against the base and the base’s derivatives with respect to the six parameters, weighted by the base’s own values. It resamples, and repeats until the parameters stop changing. -twopass first registers blurred copies, starting from a coarse search over shifts and rotations, and refines from there. The resampling is done by four shears, each a shift of rows, on a volume zero-padded by -zpad voxels. The interpolation is Fourier (-Fourier, the most accurate, fMRIPrep’s), heptic (the default), quintic, cubic, linear or nearest-neighbour. Besides the registered series it writes, on request, the parameters (-1Dfile, -dfile), the matrices (-1Dmatrix_save), the maximum displacement of a brain voxel per volume (-maxdisp1D) and fields of displacements (-savedisp). It can apply saved matrices, shift each slice in time first (-tshift), and register to another dataset’s grid or motion (-gridparent, -rotparent).

fMRIPrep calls it for every BOLD reference. niworkflows’ RobustAverage (the gen_avg node of hmc_boldref_wf) takes 20 of a run’s volumes (on ds000005’s first run, volumes 20 to 39 of 240), clips them and divides them by the global signal’s drift. It gives them to 3dvolreg with -Fourier -twopass -zpad 4, and takes the median of the registered volumes as the run’s reference. sdcflows’ PEPOLAR estimator makes its reference averages the same way.

The two ports run beside AFNI 26.3.05’s 3dvolreg and AFNI 25.2.09’s on the same files with the same arguments, both AFNI releases the repository’s builds, and beside fMRIPrep’s build of AFNI in its image; fMRIPrep’s own output is the file it wrote in its run.

bash
larmorx afni 3dvolreg -Fourier -twopass -zpad 4 -prefix registered.nii.gz \
    -1Dfile motion.1D -1Dmatrix_save motion.aff12.1D -maxdisp1D md.1D sliced.nii.gz
python
import larmorx as lx

r = lx.afni.volreg("sliced.nii.gz", **lx.afni.fmriprep_volreg_options())
r.registered, r.parameters, r.matrices, r.max_displacement

The calls above leave implementation at "auto". The API page maps every option. How to read what follows: how to read the evidence.

Its two ports

3dvolreg’s results depend on AFNI library routines copyrighted by the Medical College of Wisconsin, which AFNI 26.3.05 releases under CC BY 4.0. So the replica is part of LARMORX itself (pip install larmorx), under the Apache License with CC BY 4.0’s attribution (NOTICE, licensing), and runs in the same process. The replica is translated from AFNI’s source. The clean room is written without it, from a description of what the program does (specs/3dvolreg.md), and runs in the same process too.

implementation="auto" | "replica" | "clean-room" chooses between them ("original" in 0.1.0rc1). auto, and larmorx afni 3dvolreg, take the replica. On the command line, larmorx --implementation clean-room afni 3dvolreg ... runs the clean room from the next release; in 0.1.0rc1 it is LARMORX_IMPLEMENTATION=original.

Threads. The replica follows AFNI’s arithmetic step by step. AFNI’s 3dvolreg has no parallel loop on this path, and neither has the replica, so its results do not depend on the thread count. The clean room divides the work inside each volume between threads in fixed pieces, so its results are the same bits at any thread count too.

The verdicts at a glance

The replica is a translation of AFNI’s source, so the question for it is whether it gives AFNI’s bytes: the registered series (header and data), every text file 3dvolreg writes, its standard output and error, and its exit status.

Agreement, replica against the original: AFNI 3dvolreg 26.3.05 (standard comparison, 190 cases of the record), by category and verdict
interpolation methods21
the base volume (-base)7
two passes and the coarse search (-twopass)11
weights and edging8
environment variables6
outputs10
slice timing first (-tshift)8
another dataset's motion or grid (-rotparent, -gridparent)20
-twodup4
messages10
data types10
edge cases23
errors24
expected divergences2
global options8
defects of AFNI3
fMRIPrep's calls8
random testing7

Select a segment to list its cases below.

The replica’s record runs 190 cases against AFNI 26.3.05 (the record). In each of the 137 that write outputs the replica gives AFNI’s bytes, every file: the registered series (header and data) and every text file (-1Dfile, -dfile, -1Dmatrix_save, -maxdisp1D, -savedisp’s datasets). In 21 more neither program writes an output (an existing output, a name with a blank, a directory that does not exist), and both give the same exit status and messages. In 30 AFNI refuses the input and the replica does the same, with AFNI’s message and exit status.

Seven of those are AFNI crashing: an option given last without its value, a base whose least squares cannot be set up, a name of 256 bytes, a Fourier rotation that clears a row through a null pointer. There the replica prints AFNI’s messages up to the crash and stops with an error instead, an expected divergence. So are the two cases where AFNI writes its own dataset format, which LARMORX does not write.

The messages are AFNI’s, byte for byte, in all 180 cases that run both programs. fMRIPrep’s own outputs of its eight calls (RobustAverage’s BOLD reference of six BOLD runs from five fMRIPrep runs, sdcflows’ two PEPOLAR averages) are the replica’s, every value, and so are those of fMRIPrep’s build of AFNI. AFNI 25.2.09 gives 26.3.05’s result in all 189 cases it runs: 3dvolreg did not change between the two releases. The Python API gives the command line’s outputs in 21 of 21 cases.

The cases cover every interpolation method and -final, -twopass with its options and coarse searches (the subsampled search of a volume of 120 or more voxels along every axis included), the bases, weights, edging and clipping, -tshift, -rotparent, -gridparent and -twodup. They cover every output and data type, oblique and sagittal series, NaN, constant and single volumes, the environment’s settings, and every message, exit status and failure case of the command line. Seven cases come from random testing: six from its first run, one from the final build’s check. Each case also runs the replica’s default, fix, where LARMORX corrects AFNI’s deterministic defects and names them (below).

The clean room is not a translation: it is written from the specification (specs/3dvolreg.md), so the question for it is whether it behaves as AFNI does within the specification’s limits (its §14.3). Its exit statuses, files, headers, types, geometry and messages are AFNI’s; its numbers cannot all be AFNI’s bits. Where two axes have the same voxel size (every EPI with square in-plane voxels), the four shears that rotate a volume can be ordered in ways whose norms tie in exact arithmetic. AFNI decides each tie by its last bits, at every resampling, in a way no specification states. The limits allow for either side of every tie and grow with a volume’s rotation, as AFNI’s own spread does when its ties are decided the other way (the spec writer measured it on 16 real runs).

Agreement, clean room against the original: AFNI 3dvolreg 26.3.05 (standard comparison, 251 cases of the record), by category and verdict
interpolation methods22
two passes and the coarse search (-twopass)12
data types12
outputs15
messages13
errors34
the base volume (-base)7
edge cases25
defects of AFNI30
another dataset's motion or grid (-rotparent, -gridparent)24
slice timing first (-tshift)9
-twodup4
weights and edging8
the fit's settings6
environment variables9
global options8
thread counts3
fMRIPrep's calls10

Select a segment to list its cases below.

The clean room’s record runs 251 cases against AFNI 26.3.05 (the record), and holds each case’s fix against the replica’s fix by the same limits. 185 pass, 63 are refused by both with the same messages and exit status, and 3 are expected divergences: AFNI’s own .HEAD/.BRIK format, twice, and a -rotparent attribute that is too short (AFNI-96), which AFNI reads past and the clean room refuses. None fails. 417 of its 1,096 -1Dfile lines are AFNI’s as text, and its outputs are the same bits at 1, 2, 5 and 14 threads. How close it comes to each limit:

How close the clean room comes to each limit, over its record’s 251 cases

Quantitylargest share of its limitcase
the parameters (-1Dfile)0.611original/iso-ras-fourier
the matrices’ rotations0.947original/four-slices
the matrices’ translations0.689original/params
the maximum displacements0.659original/negative
their deltas (_delt)0.626original/negative
-savedisp, DX0.671original/afni99-savedisp
-savedisp, DY0.723original/afni99-savedisp
-savedisp, DZ0.163original/afni99-savedisp
-SAVEDISP’s raw displacements0.538original/SAVEDISP
-SAVEDISP’s normalised displacements, where well conditioned0.0523original/SAVEDISP
the registered volumes0.681original/fmriprep-ds000256-nofeedback
fMRIPrep’s medians0.115original/fmriprep-ds000256-nofeedback

1 = at the limit (specs/3dvolreg.md §14.3, the spec writer’s step 3 of 2026-11-08). The record’s -1Dfile lines: 417 of 1,096 identical with AFNI’s (bit-identity reported, not required). The registered volumes’ largest share is ds000256’s high-motion child run (its worked example below, ‘A child’s high-motion run’).

In fMRIPrep: a real run’s reference volumes

fMRIPrep’s own call on its run of OpenNeuro ds000005 sub-01, given to AFNI 26.3.05, to AFNI 25.2.09, to fMRIPrep’s build of AFNI, to the replica and to the clean room. fMRIPrep keeps only the median of the registered volumes (the next example), so the registered series below is what each program writes for that call. The input panels show the run as stored and the volumes 3dvolreg is given. The charts below the viewer are the motion parameters and maximum displacements, volume by volume.

Worked example

fMRIPrep's registration of a real BOLD run's reference volumes

What this shows: On fMRIPrep’s call the replica gives AFNI’s registered series, every value, and its parameters and matrices as text: in both releases, in fMRIPrep’s build, on 1 and 6 threads. Its default, fix, changes only the maximum displacements, where it corrects AFNI-99; reproduce gives every file AFNI writes. The clean room’s values are not AFNI’s bits, by design: each of the 20 volumes stays within its limit (specs/3dvolreg.md §14.3), the closest at 0.00035 of it.

fMRIPrep’s BOLD reference is the median of some of a run’s volumes, aligned first: niworkflows’ RobustAverage (the gen_avg node of fMRIPrep 25.2.6’s hmc_boldref_wf) takes 20 volumes of OpenNeuro ds000005 sub-01’s first run (file volumes 20 to 39 of 240), clips them and divides them by the global signal’s drift, and gives them to 3dvolreg with -Fourier -twopass -zpad 4. That call, given to AFNI 26.3.05, AFNI 25.2.09, fMRIPrep’s build of AFNI and the replica in fix and in reproduce. Choose a volume with the slider; the input panels show the run before registration. Beside them LARMORX’s clean room (Apache-2.0, written from specs/3dvolreg.md), in fix, its default, at 14 threads and at one and in reproduce, with its parameters’ differences from AFNI’s and every volume against its limit in the second chart and the table below.

Data
OpenNeuro ds000005 (Tom et al., 2007), snapshot 00001, sub-01, task mixedgamblestask run 1 (PDDL-1.0); shown inside fMRIPrep’s BOLD brain mask, dilated by one voxel
Node
fMRIPrep 25.2.6, bold_fit_wf/hmc_boldref_wf/gen_avg (niworkflows’ RobustAverage)
Command
3dvolreg -Fourier -twopass -zpad 4 -prefix o.nii.gz -1Dfile o.1D -1Dmatrix_save o.aff12.1D -maxdisp1D o_md.1D sliced.nii.gz
Programs
AFNI_26.3.05, gcc (Ubuntu 11.4.0-1ubuntu1~22.04.3) 11.4.0 and AFNI_25.2.09, gcc (Ubuntu 11.4.0-1ubuntu1~22.04.3) 11.4.0 (the repository’s builds); AFNI AFNI_25.2.09 (Aug 24 2025) [64-bit] in fMRIPrep’s image; LARMORX 0.0.1 (crate larmorx-afni-replica, built in) at commit c0fa9881ce74; LARMORX’s clean room (crate larmorx-afni, Apache-2.0, experimental), from LARMORX at commit ae74dcb46642
Grid
64 × 64 × 34 × 20 voxels of 3.12 × 3.12 × 4 mm; shown: 53 × 55 × 34 around the brain
Input3dvolreg's input: RobustAverage's sliced volumesclipped to the 99.8th percentile and divided by the global signal's drift (fMRIPrep's Python in its image); int16 with a scale factor
OriginalAFNI 3dvolreg 26.3.05the repository's build; the version the replica follows
Replicareplicalarmorx, Apache-2.0 with CC BY 4.0's attribution · fix (the default), 1 thread · against AFNI 3dvolreg 26.3.05, which it follows
Replicareplicalarmorx, Apache-2.0 with CC BY 4.0's attribution · reproduce, 1 thread · against AFNI 3dvolreg 26.3.05, which it follows
Clean roomclean roomlarmorx, Apache-2.0 · fix (the default), 14 threads · against AFNI 3dvolreg 26.3.05, which it follows

Bit-identical2,785,280 of 2,785,280 voxels equal

the original against itself: versions 25.2.09 and 26.3.05AFNI 3dvolreg 25.2.09 − 26.3.05

Bit-identical2,785,280 of 2,785,280 voxels equal

the original against itself: fMRIPrep's build of 25.2.09 and the repository's 26.3.05AFNI 3dvolreg 25.2.09, fMRIPrep's build − 26.3.05

Bit-identical2,785,280 of 2,785,280 voxels equal

the replica against the originalreplica (fix) − AFNI 3dvolreg 26.3.05
the clean room against the originalclean room (fix) − AFNI 3dvolreg 26.3.05scale ±0.00272

Axial slice 16 of 34, volume 17 of 20; file volume 36, the one AFNI finds farthest from the base (the first volume given). The interactive viewer needs JavaScript; the metrics below are complete without it.

The six motion parameters and the maximum displacements over the 20 volumes, the replica against the original: AFNI 3dvolreg 26.3.05
  • AFNI 26.3.05
  • replica, fix
  • replica, reproduce (AFNI's bytes)
roll (°, about I-S)0.0176−0.0018pitch (°, about R-L)0.005−0.0436yaw (°, about A-P)0.01850dS (mm)0.0897−0.0213dL (mm)0.0178−0.0014dP (mm)0.0145−0.0831maximum displacement (mm, -maxdisp1D)0.1570maximum displacement from the previous volume (mm, _delt)0.1230

3dvolreg’s own parameters (-1Dfile): rotations in degrees about AFNI’s axes, shifts in mm (S, L, P positive), from the base (the first volume) to each volume. Lines that coincide are the same values. Move the pointer over a chart to read every value at one volume.

The text outputs compared, over every volume (the files without their comment line)

Comparisonlargest |Δ| rotation (°)largest |Δ| shift (mm)largest |Δ| max displacement (mm)identical as text
replica (fix) − AFNI 3dvolreg 26.3.05AFNI-99 corrected000.011-1Dfile, matrices
replica (reproduce) − AFNI 3dvolreg 26.3.05000-1Dfile, matrices, -maxdisp1D, _delt
replica: fix − reproduce000.011-1Dfile, matrices
replica: 6 threads − 1 thread000-1Dfile, matrices, -maxdisp1D, _delt
AFNI 3dvolreg 25.2.09 − 26.3.05the version step000-1Dfile, matrices, -maxdisp1D, _delt
AFNI 3dvolreg 25.2.09, fMRIPrep’s build − 26.3.05000-1Dfile, matrices, -maxdisp1D, _delt

Values as the files write them (-1Dfile: 4 decimals; -maxdisp1D: 3). The shaded rows are AFNI against itself. AFNI-99: the maximum displacement AFNI writes is that of the voxel mirrored along every axis stored posterior to anterior (this run’s second axis); fix, the default, computes it for the voxel itself; reproduce gives AFNI’s numbers.

Differences volume by volume, the clean room against the original: AFNI 3dvolreg 26.3.05 (the six parameters, the maximum displacement and each registered volume’s share of its limit)
  • clean room (fix) − AFNI 3dvolreg 26.3.05
  • the volume's limit (specs/3dvolreg.md §14.3)
roll (°): clean room − AFNI 26.3.050.0201−0.0201pitch (°): clean room − AFNI 26.3.050.0202−0.0202yaw (°): clean room − AFNI 26.3.050.0201−0.0201dS (mm): clean room − AFNI 26.3.050.0404−0.0404dL (mm): clean room − AFNI 26.3.050.0313−0.0313dP (mm): clean room − AFNI 26.3.050.0317−0.0317maximum displacement (mm, -maxdisp1D): clean room (reproduce) − AFNI0.00225−0.00225registered volume: mean |Δ| ÷ (L_v · its largest value)10registered volume: (1 − NCC) ÷ (50 · L_v²)10

Each band is the limit §14.3 sets for that volume (from AFNI’s own values): the parameters’ thresholds, the maximum displacement’s derived bound, and 1 for the registered volumes’ shares. The maximum displacements are those of reproduce (AFNI’s convention; those of fix are AFNI-99’s correction, shown in its own example). Move the pointer over a chart to read every value at one volume.

The clean room’s text outputs compared, over every volume (the files without their comment line)

Comparisonlargest |Δ| rotation (°)largest |Δ| shift (mm)largest |Δ| max displacement (mm)largest share of a parameter's thresholdidentical as text
clean room (fix) − AFNI 3dvolreg 26.3.05AFNI-99 named: the maximum displacements of the voxels themselves0.000100.0110.00517 of 20 -1Dfile lines
clean room (reproduce) − AFNI 3dvolreg 26.3.05the maximum displacements within their derived bounds0.0001000.00517 of 20 -1Dfile lines
clean room − replica, both fix0.0001000.00517 of 20 -1Dfile lines
clean room: 1 thread − 14 threads000–20 of 20 -1Dfile lines

§14.3’s parameter thresholds: |Δ angle| ≤ 0.02° + 0.005·|angle| and |Δ shift| ≤ 0.01·δ + 0.005·|shift| (δ the voxel size along the shift’s axis; fMRIPrep’s call keeps 3dvolreg’s default -rot_thresh and -x_thresh), the median of all |Δ| ≤ 0.001; the maximum displacements within alpha_v·r + tau_v + 0.001 (alpha_v, tau_v the volume’s rotation and shift differences, r half the grid’s diagonal: §14.3 ‘Derived quantities’).

Every comparison, computed in float64 over every voxel of the outputs

Comparison (A − B), verdict and reasonmax |Δ|Values that differRMS of Δmax |Δ| ÷ range1 − NCC
The original against itself: the yardstick
AFNI 3dvolreg 25.2.09 − 26.3.05referenceAFNI 25.2.09 gives AFNI 26.3.05’s bytes on this call (the series and every text file): 3dvolreg did not change between the two releases.identical: 0 of 2,785,280 values differ
AFNI 3dvolreg 25.2.09, fMRIPrep's build − 26.3.05referencefMRIPrep’s build of AFNI (in its image) gives the repository’s AFNI 26.3.05 bytes, the series and every text file.identical: 0 of 2,785,280 values differ
The replica against the original
replica (fix) − AFNI 3dvolreg 26.3.05Bit-identicalThe same values, every one: the registered series, and the parameters and matrices as text; the maximum displacements differ (AFNI-99, corrected by default: the chart below). AFNI-99identical: 0 of 2,785,280 values differ
replica (reproduce) − AFNI 3dvolreg 26.3.05Bit-identicalThe same values, every one, and every text file AFNI writes.identical: 0 of 2,785,280 values differ
The replica against itself: thread counts, fix and reproduce
replica, fix − reproduceBit-identicalThe registered volumes are the same bytes in fix and in reproduce: AFNI-99 changes only the maximum displacements, which the default computes for the voxels themselves (the chart below). AFNI-99identical: 0 of 2,785,280 values differ
replica, 6 threads − 1 threadBit-identicalThe same bytes on any number of threads (AFNI’s 3dvolreg has no parallel loop on this path; neither has the replica’s registration).identical: 0 of 2,785,280 values differ
The clean room against the original and the replica
clean room (fix) − AFNI 3dvolreg 26.3.05Within the original's own variabilityNot AFNI’s bits by design: where two axes have the same voxel size the shear decomposition’s axis orders tie in exact arithmetic, and AFNI decides each tie by its last bits at every resampling, which no specification states (specs/3dvolreg.md §5.3, §16); §14.3’s limits allow for either side of every tie, measured from AFNI’s own spread when its ties are decided the other way. Here: every one of the 20 volumes within §14.3’s limit (the closest, sub-brick 3: mean |Δ| 7e-07 of its largest value against L_v = 0.002, 0.00035 of it; 1 − NCC 9.9e-12, 5e-08 of 50·L_v²); 17 of 20 -1Dfile lines identical, the parameters at most 0.005 of their thresholds (the charts below); 1,670,448 of 2,785,280 values differ, the largest by 0.0148 (1.9e-05 of the range) at a field of view’s edge. §14.30.01481,670,448 of 2,785,280 (60 %)0.0004581.86e−51.92e−12
clean room (reproduce) − AFNI 3dvolreg 26.3.05Within the original's own variabilityNot AFNI’s bits by design: where two axes have the same voxel size the shear decomposition’s axis orders tie in exact arithmetic, and AFNI decides each tie by its last bits at every resampling, which no specification states (specs/3dvolreg.md §5.3, §16); §14.3’s limits allow for either side of every tie, measured from AFNI’s own spread when its ties are decided the other way. Here: every one of the 20 volumes within §14.3’s limit (the closest, sub-brick 3: mean |Δ| 7e-07 of its largest value against L_v = 0.002, 0.00035 of it; 1 − NCC 9.9e-12, 5e-08 of 50·L_v²); 17 of 20 -1Dfile lines identical, the parameters at most 0.005 of their thresholds (the charts below); 1,670,448 of 2,785,280 values differ, the largest by 0.0148 (1.9e-05 of the range) at a field of view’s edge. §14.30.01481,670,448 of 2,785,280 (60 %)0.0004581.86e−51.92e−12
clean room − replica, both fixWithin the original's own variabilityThe two ports’ defaults (fix): the replica’s is AFNI’s series bit for bit (AFNI-99 changes only the maximum displacements), so the same limits hold: every one of the 20 volumes within §14.3’s limit (the closest, sub-brick 3: mean |Δ| 7e-07 of its largest value against L_v = 0.002, 0.00035 of it; 1 − NCC 9.9e-12, 5e-08 of 50·L_v²); 17 of 20 -1Dfile lines identical.0.01481,670,448 of 2,785,280 (60 %)0.0004581.86e−51.92e−12
The clean room against itself: thread counts, fix and reproduce
clean room, fix − reproduceBit-identicalThe registered volumes are the same in fix and in reproduce: AFNI-99, named by fix, changes only the maximum displacements (the chart and table below). AFNI-99identical: 0 of 2,785,280 values differ
clean room, 1 thread − 14 threadsBit-identicalThe thread count changes nothing: the clean room registers the volumes one after another and divides the work inside each (the shears’ planes and rows, the coarse search, the blurs) in fixed pieces, so 1 and 14 threads give the same bits (its record checks 1, 2, 5 and 14).identical: 0 of 2,785,280 values differ

Δ is A − B, computed in float64 over every value of the files (2,785,280), not only those shown; the range is B’s maximum minus its minimum. The records’ thresholds: every output bit-identical to AFNI 26.3.05’s (the registered series, -1Dfile, -1Dmatrix_save, -maxdisp1D), with its exit status and messages (the replica); AFNI’s exit status, files, headers and messages, the motion parameters within per-volume limits derived from each volume’s own parameters, and the outputs computed from them within the bounds those limits give (the clean room, specs/3dvolreg.md §14.3). The rows the original against itself are AFNI’s own variability on this input: its previous release and fMRIPrep’s build.

The registered volumes against their limits, volume by volume, the clean room against the original: AFNI 3dvolreg 26.3.05 (fix; specs/3dvolreg.md §14.3)

Sub-brickθ_v (°)L_vmean |Δ| ÷ maxshare of L_v1 − NCCshare of 50·L_v²largest |Δ| ÷ max
sub-brick 000.00200000
sub-brick 10.01090.0022.1e−70.0001051.13e−125.65e−95.31e−6
sub-brick 20.02370.0021.16e−75.79e−52.7e−131.35e−92.42e−6
sub-brick 30.01950.0026.96e−70.0003489.93e−124.96e−81.6e−5
sub-brick 40.01030.0021.24e−76.21e−53.24e−131.62e−92.69e−6
sub-brick 50.02180.0026.48e−70.0003241.27e−116.34e−81.87e−5
sub-brick 60.02750.0021.63e−78.17e−56.45e−133.23e−93.96e−6
sub-brick 70.02750.0021.7e−78.5e−56.86e−133.43e−94.01e−6
sub-brick 80.04120.0022.87e−70.0001442.23e−121.11e−87.46e−6
sub-brick 90.04430.0022.2e−70.000111.24e−126.21e−95.49e−6
sub-brick 100.05560.0021.2e−76.02e−52.99e−131.5e−92.65e−6
sub-brick 110.03890.0022.13e−70.0001071.16e−125.79e−95.23e−6
sub-brick 120.03970.0022.47e−70.0001241.62e−128.09e−96.22e−6
sub-brick 130.03270.0021.87e−79.36e−58.76e−134.38e−94.61e−6
sub-brick 140.03870.0022.28e−70.0001141.36e−126.82e−95.83e−6
sub-brick 150.03180.0021.08e−75.39e−52.25e−131.13e−92.15e−6
sub-brick 160.04390.0021.63e−78.16e−56.26e−133.13e−93.99e−6
sub-brick 170.03320.0021.67e−78.35e−56.54e−133.27e−93.98e−6
sub-brick 180.06420.0021.54e−77.69e−55.5e−132.75e−93.48e−6
sub-brick 190.05340.0022.58e−70.0001291.78e−128.92e−96.75e−6

θ_v = |roll| + |pitch| + |yaw| of AFNI’s -1Dfile line (degrees); L_v = max(2e-3, 1e-3·θ_v); required: mean |Δ| ≤ L_v times the volume’s largest |AFNI| value and NCC ≥ 1 − 50·L_v² (the spec writer’s step 3 of 2026-11-08: the limit grows with the rotation, as AFNI’s own spread under the ties does); the largest single difference is reported, not required (Fourier’s wrap-around at the field of view’s edges).

Provenance of this example
  • Data: OpenNeuro ds000005 (Tom et al., 2007), snapshot 00001, sub-01, PDDL-1.0
  • Inputs: RobustAverage’s sliced volumes (sha256 2cce88990bcfbee8…); the raw BOLD run (sha256 609ae0a21822318f…); fMRIPrep’s BOLD reference (sha256 ab49738164116b9c…); the sliced volumes stored LPS (sha256 71708c457976d2d1…)
  • Runs: figures cache entry afni-volreg/fmriprep-ds000005-run-01, computed 2026-11-07T01:27:24+00:00 with larmorx c0fa9881ce74; AFNI_26.3.05, gcc (Ubuntu 11.4.0-1ubuntu1~22.04.3) 11.4.0; AFNI_25.2.09, gcc (Ubuntu 11.4.0-1ubuntu1~22.04.3) 11.4.0; AFNI_25.2.09 (Aug 24 2025) [64-bit]; larmorx 0.0.1 (crate larmorx-afni-replica, built in); every command, its environment, threads and exit status are in its RUN_INFO.json
  • Runs of the clean room: figures cache entry afni-volreg/fmriprep-ds000005-run-01-original, computed 2026-11-08T08:38:25+00:00 with larmorx ae74dcb46642 (python -m larmorx with LARMORX_IMPLEMENTATION=original: fix, the default at 14 threads and at one, reproduce at 14; on the volumes stored LPS both modes), on the replica’s entry’s inputs; every command, its environment, threads and exit status are in its RUN_INFO.json

Every number above is in afni-volreg.json, exported by python -m larmorx_validation evidence afni-volreg.

The BOLD reference fMRIPrep makes from them

Worked example

The BOLD reference fMRIPrep makes from the registered volumes

What this shows: The replica’s registered volumes give fMRIPrep’s BOLD reference of this run, every value, as AFNI 26.3.05’s do. The clean room’s give a median that differs in 1,979 of its 139,264 values, at 0.00011 of the limit its volumes imply (specs/3dvolreg.md §14.3).

RobustAverage’s last step: the median, voxel by voxel, of the registered volumes of the first example. fMRIPrep’s own reference of the run (the node’s output, from AFNI in its image) beside the median of AFNI 26.3.05’s, AFNI 25.2.09’s and the replica’s registered series, computed as RobustAverage computes it. The first panel is the median of the volumes before registration: choose compare to see what aligning them changes. Beside them the median of LARMORX’s clean room’s volumes, in fix and in reproduce, against the limit its volumes imply (the table below).

Data
OpenNeuro ds000005 (Tom et al., 2007), snapshot 00001, sub-01, run 1 (PDDL-1.0); shown inside fMRIPrep’s BOLD brain mask, dilated by one voxel
Node
fMRIPrep 25.2.6, bold_fit_wf/hmc_boldref_wf/gen_avg: its output, …_bold_average.nii.gz
Median
niworkflows’ RobustAverage: the registered values clipped (at 0 below), np.median over the volumes, written with the sliced volumes’ header (float32)
Grid
64 × 64 × 34 voxels of 3.12 × 3.12 × 4 mm; shown: 53 × 55 × 34 around the brain
Inputthe sliced volumes' median, not registeredwhat the median would be without 3dvolreg (for comparison; not a step of fMRIPrep)
fMRIPrep's outputfMRIPrep's outputfMRIPrep 25.2.6, as it ran: the node's BOLD reference (AFNI in its image)
Replicareplicalarmorx, Apache-2.0 with CC BY 4.0's attribution · its series, median, fix (the default), 1 thread · against AFNI 3dvolreg 26.3.05
Clean roomclean roomlarmorx, Apache-2.0 · its series, median, fix (the default), 14 threads · against AFNI 3dvolreg 26.3.05

Bit-identical139,264 of 139,264 voxels equal

the original against itself: fMRIPrep's AFNI 25.2.09 in its imageAFNI 3dvolreg 26.3.05 − fMRIPrep's output

Bit-identical139,264 of 139,264 voxels equal

the replica against fMRIPrep's outputreplica − fMRIPrep's output
the clean room against fMRIPrep's outputclean room − fMRIPrep's outputscale ±0.0121

Axial slice 16 of 34; the crosshair at the brain's centre. The interactive viewer needs JavaScript; the metrics below are complete without it.

Every comparison, computed in float64 over every voxel of the outputs

Comparison (A − B), verdict and reasonmax |Δ|Values that differRMS of Δmax |Δ| ÷ range1 − NCC
The original against itself: the yardstick
AFNI 3dvolreg 26.3.05 − fMRIPrep's outputreferenceAFNI 26.3.05 gives the reference fMRIPrep computed with its own AFNI, every value.identical: 0 of 139,264 values differ
AFNI 3dvolreg 25.2.09 − 26.3.05referenceThe same bytes: 3dvolreg did not change between the two releases.identical: 0 of 139,264 values differ
The replica against the original and fMRIPrep's output
replica (fix) − AFNI 3dvolreg 26.3.05Bit-identicalThe same values, every one.identical: 0 of 139,264 values differ
replica (reproduce) − AFNI 3dvolreg 26.3.05Bit-identicalThe same values, every one.identical: 0 of 139,264 values differ
replica − fMRIPrep's outputBit-identicalfMRIPrep’s BOLD reference of this run, every value: the replica’s registered volumes give the median fMRIPrep computed from AFNI’s.identical: 0 of 139,264 values differ
The clean room against the original, fMRIPrep's output and the replica
clean room (fix) − AFNI 3dvolreg 26.3.05's medianWithin the original's own variabilityfMRIPrep’s median of the clean room’s registered volumes: at each voxel the median moves by at most the largest of its volumes’ differences there, so §14.3 holds it to the limit its volumes imply (L_med = 0.002, the run’s largest L_v, and the stored median’s step q = 0.0121): mean |Δ| 2.2e-07 of the largest value, 1 − NCC 1.9e-11: 0.00011 of the limit. 1,979 of 139,264 values differ, the largest by 0.0121 (1.5e-05 of the largest value; reported, not required). AFNI’s median is fMRIPrep’s own byte for byte, and so is the replica’s. §14.30.01211,979 of 139,264 (1.42 %)0.001451.53e−51.94e−11
clean room − fMRIPrep's outputWithin the original's own variabilityfMRIPrep’s median of the clean room’s registered volumes: at each voxel the median moves by at most the largest of its volumes’ differences there, so §14.3 holds it to the limit its volumes imply (L_med = 0.002, the run’s largest L_v, and the stored median’s step q = 0.0121): mean |Δ| 2.2e-07 of the largest value, 1 − NCC 1.9e-11: 0.00011 of the limit. 1,979 of 139,264 values differ, the largest by 0.0121 (1.5e-05 of the largest value; reported, not required). §14.30.01211,979 of 139,264 (1.42 %)0.001451.53e−51.94e−11
clean room − replica, both fixWithin the original's own variabilityfMRIPrep’s median of the clean room’s registered volumes: at each voxel the median moves by at most the largest of its volumes’ differences there, so §14.3 holds it to the limit its volumes imply (L_med = 0.002, the run’s largest L_v, and the stored median’s step q = 0.0121): mean |Δ| 2.2e-07 of the largest value, 1 − NCC 1.9e-11: 0.00011 of the limit. 1,979 of 139,264 values differ, the largest by 0.0121 (1.5e-05 of the largest value; reported, not required). AFNI’s median is fMRIPrep’s own byte for byte, and so is the replica’s. §14.30.01211,979 of 139,264 (1.42 %)0.001451.53e−51.94e−11
The clean room against itself: fix and reproduce
clean room, fix − reproduceBit-identicalThe same median in fix and in reproduce (the registered volumes are the same).identical: 0 of 139,264 values differ

Δ is A − B, computed in float64 over every value of the files (139,264), not only those shown; the range is B’s maximum minus its minimum. The records’ thresholds: every output bit-identical to AFNI 26.3.05’s (the registered series and the text files; here their median, which the record also compares with fMRIPrep’s own), with its exit status and messages (the replica); AFNI’s exit status, files, headers and messages, the motion parameters within per-volume limits derived from each volume’s own parameters, and the outputs computed from them within the bounds those limits give (the clean room, specs/3dvolreg.md §14.3). The rows the original against itself are AFNI’s own variability on this input: its previous release and fMRIPrep’s build.

fMRIPrep’s median of the clean room’s volumes against fMRIPrep’s output, its own reference (specs/3dvolreg.md §14.3)

ComparisonL_medqmean |Δ| ÷ max1 − NCCshare of the limitlargest |Δ| ÷ maxvoxels beyond 1e-3 of max
clean room (fix) − fMRIPrep’s output0.0020.01212.17e−71.94e−110.0001081.53e−50

Required: mean |Δ| ≤ L_med·max + q and NCC ≥ 1 − 50·L_med², L_med the largest L_v of the run’s volumes and q the stored median’s step (its scale factor: fMRIPrep stores the reference as int16); the largest difference and the share of voxels beyond 1e-3 of the maximum reported.

Provenance of this example
  • Data: OpenNeuro ds000005 (Tom et al., 2007), snapshot 00001, sub-01, PDDL-1.0
  • Inputs: RobustAverage’s sliced volumes (sha256 2cce88990bcfbee8…); the raw BOLD run (sha256 609ae0a21822318f…); fMRIPrep’s BOLD reference (sha256 ab49738164116b9c…); the sliced volumes stored LPS (sha256 71708c457976d2d1…)
  • Runs: figures cache entry afni-volreg/fmriprep-ds000005-run-01, computed 2026-11-07T01:27:24+00:00 with larmorx c0fa9881ce74; AFNI_26.3.05, gcc (Ubuntu 11.4.0-1ubuntu1~22.04.3) 11.4.0; AFNI_25.2.09, gcc (Ubuntu 11.4.0-1ubuntu1~22.04.3) 11.4.0; AFNI_25.2.09 (Aug 24 2025) [64-bit]; larmorx 0.0.1 (crate larmorx-afni-replica, built in); every command, its environment, threads and exit status are in its RUN_INFO.json
  • Runs of the clean room: figures cache entry afni-volreg/fmriprep-ds000005-run-01-original, computed 2026-11-08T08:38:25+00:00 with larmorx ae74dcb46642 (python -m larmorx with LARMORX_IMPLEMENTATION=original: fix, the default at 14 threads and at one, reproduce at 14; on the volumes stored LPS both modes), on the replica’s entry’s inputs; every command, its environment, threads and exit status are in its RUN_INFO.json

Every number above is in afni-volreg.json, exported by python -m larmorx_validation evidence afni-volreg.

The clean room on fMRIPrep’s calls

The replica’s eight calls and two more of the data panel’s runs (ds000256’s child at the least and at the most motion), on the inputs made again from each fMRIPrep run’s node in fMRIPrep’s image. AFNI’s median is fMRIPrep’s own output byte for byte in every one, so the clean room’s median is compared with fMRIPrep’s output.

fMRIPrep’s calls, the clean room against the original: AFNI 3dvolreg 26.3.05 (its record’s standard comparison)

Call-1Dfile lines identicalparametersmaximum displacementsregistered volumesfMRIPrep's medianmedian: largest |Δ| ÷ maxchecks
ds000005-run-01single-band, int16 stored, 64 x 64 x 34, 20 volumes17/200.00500.00030.00011.53e−549/49
ds000005-run-03two volumes2/2000.000101.53e−549/49
ds005454multiband, 160 x 160 x 96, oblique10/200.007100.00480.00050.00010749/49
ds003592-echo-1multi-echo (echo 1), oblique19/200.004800.00050.00011.53e−549/49
ds004958multiband 8, uint16 stored, three volumes, 24 degrees oblique1/30.0100.00450.00160.00013749/49
ds002603-mb4Philips multiband 4, float32 stored20/20000.00010.00012.55e−649/49
pepolar-ds005040sdcflows’ PEPOLAR reference average, two topup-corrected volumes1/20.009900.00040.00021.69e−549/49
pepolar-ds001600the same, 80 x 80 x 601/20.003300.00020.00011.19e−549/49
ds000256-baselinea child’s run (the motion levels’ baseline)19/200.002500.000101.53e−549/49
ds000256-nofeedbacka child’s run, the most motion (rotations to 4.7 degrees)7/200.1230.4630.6810.1150.019849/49

Each number is the largest share of its limit over the call’s volumes (1 = at the limit; specs/3dvolreg.md §14.3): the parameters’ thresholds, the maximum displacements’ derived bounds (reproduce, AFNI’s convention), each registered volume’s limit L_v, and fMRIPrep’s median against fMRIPrep’s own reference (the node’s output) by the limit its volumes imply; the median’s largest single difference is reported. The inputs are re-made from each fMRIPrep run’s node in fMRIPrep’s image, so that AFNI’s median is fMRIPrep’s own output byte for byte (checked in every case). From the record docs/validation/afni-volreg-original.md.

A child’s high-motion run

ds000256’s rest run without feedback moves the most of the data panel’s runs (rotations to 4.7°), and the ties weigh most on the volumes that rotate most. Here the clean room comes closest to its limits, 0.68 of one volume’s, as close as AFNI’s own results come when its ties are decided the other way.

Worked example

A child's high-motion run: where the clean room comes closest to its limits

What this shows: On the data panel’s run with the most motion the replica gives AFNI’s registered volumes bit for bit. The clean room’s differ within specs/3dvolreg.md §14.3’s limits and come closest to them here: sub-brick 11 at 0.68 of its limit, as close as AFNI’s own results come when its ties are decided the other way.

fMRIPrep 25.2.6’s RobustAverage call on OpenNeuro ds000256 sub-CTS203’s rest run without feedback (a 7.5-year-old, the most motion of three runs: rotations to 4.7°), given to AFNI 26.3.05, to the replica and to LARMORX’s clean room. The replica gives AFNI’s volumes bit for bit; the clean room’s differ within specs/3dvolreg.md §14.3’s limits, closest to them on the volumes that rotate most, where the shear decomposition’s ties weigh most. fMRIPrep keeps only the median of the 20 registered volumes (the table below).

Data
OpenNeuro ds000256 (Greene et al., 2018), snapshot 00002, sub-CTS203, task restfixationnofeedback (PDDL-1.0); shown inside fMRIPrep’s BOLD brain mask, dilated by one voxel
Node
fMRIPrep 25.2.6, sub_CTS203_wf/bold_task_restfixationnofeedback_wf/bold_fit_wf/hmc_boldref_wf/gen_avg
Command
3dvolreg -Fourier -twopass -zpad 4 -prefix o.nii.gz -1Dfile o.1D -1Dmatrix_save o.aff12.1D -maxdisp1D o_md.1D sliced.nii.gz
Programs
AFNI 26.3.05’s 3dvolreg (the repository’s build, through the cache of the original’s runs); the replica and LARMORX’s clean room (crate larmorx-afni, experimental), from LARMORX at commit ae74dcb46642
Grid
64 × 64 × 32 × 20 voxels of 4 × 4 × 4 mm; shown: 42 × 50 × 32 around the brain
Input3dvolreg's input: RobustAverage's sliced volumesthe run's first 20 volumes after fMRIPrep's dummy scans, clipped and divided by the global signal's drift (fMRIPrep's Python in its image); int16 with a scale factor
OriginalAFNI 3dvolreg 26.3.05the repository's build, through the cache of the original's runs
Replicareplicalarmorx, Apache-2.0 with CC BY 4.0's attribution · fix (the default), 1 thread · against AFNI 3dvolreg 26.3.05
Clean roomclean roomlarmorx, Apache-2.0 · fix (the default), 14 threads · against AFNI 3dvolreg 26.3.05

Bit-identical2,621,440 of 2,621,440 voxels equal

the replica against the originalreplica (fix) − AFNI 3dvolreg 26.3.05
the clean room against the originalclean room (fix) − AFNI 3dvolreg 26.3.05scale ±26.2

Axial slice 19 of 32, volume 12 of 20; sub-brick 11 (counting from 0, as AFNI does), the one closest to its limit; the difference on its own symmetric scale. The interactive viewer needs JavaScript; the metrics below are complete without it.

The six motion parameters over the 20 volumes, the clean room against the original: AFNI 3dvolreg 26.3.05
  • AFNI 3dvolreg 26.3.05 (the replica gives these numbers)
  • clean room, fix
roll (°)4.72−0.112pitch (°)0.0619−0.52yaw (°)0.516−0.756dS (mm)1.610dL (mm)3.730dP (mm)1.07−0.0953

The two curves overlap at this scale; the next chart shows their differences against §14.3’s limits.

The text outputs compared, over every volume (the files without their comment line)

Comparisonlargest |Δ| rotation (°)largest |Δ| shift (mm)largest share of a parameter's thresholdidentical as text
replica (fix) − AFNI 3dvolreg 26.3.0500020 of 20 -1Dfile lines
clean room (fix) − AFNI 3dvolreg 26.3.050.00380.00140.1237 of 20 -1Dfile lines

§14.3’s thresholds as in the first example; the maximum displacements of this run (to about 7 mm) follow AFNI-99 in fix (the record’s case fmriprep-ds000256-nofeedback holds them, within the derived bounds in reproduce).

Differences volume by volume, the clean room against the original: AFNI 3dvolreg 26.3.05 (the six parameters and each registered volume’s share of its limit)
  • clean room (fix) − AFNI 3dvolreg 26.3.05
  • the volume's limit (specs/3dvolreg.md §14.3)
roll (°): clean room − AFNI 26.3.050.0436−0.0436pitch (°): clean room − AFNI 26.3.050.0226−0.0226yaw (°): clean room − AFNI 26.3.050.0238−0.0238dS (mm): clean room − AFNI 26.3.050.0481−0.0481dL (mm): clean room − AFNI 26.3.050.0586−0.0586dP (mm): clean room − AFNI 26.3.050.0454−0.0454registered volume: mean |Δ| ÷ (L_v · its largest value)10registered volume: (1 − NCC) ÷ (50 · L_v²)10

Each band is the limit §14.3 sets for that volume, from AFNI’s own values. The volumes with the largest rotations come closest: the limit L_v grows with the rotation as AFNI’s own spread under the ties does (the spec writer measured AFNI’s tie variants on this run up to 0.68 of L_v).

The registered volumes against their limits, volume by volume, the clean room against the original: AFNI 3dvolreg 26.3.05 (fix; specs/3dvolreg.md §14.3)

Sub-brickθ_v (°)L_vmean |Δ| ÷ maxshare of L_v1 − NCCshare of 50·L_v²largest |Δ| ÷ max
sub-brick 000.00200000
sub-brick 10.9790.0021.41e−77.05e−55.47e−132.74e−94.35e−6
sub-brick 20.6970.0028.58e−84.29e−51.45e−137.26e−101.92e−6
sub-brick 30.6730.0028.53e−84.26e−51.46e−137.28e−102.07e−6
sub-brick 40.7560.0028.84e−84.42e−51.56e−137.82e−102.06e−6
sub-brick 50.7140.0029.33e−84.66e−51.84e−139.19e−102.38e−6
sub-brick 60.8110.0029.32e−84.66e−51.83e−139.13e−102.3e−6
sub-brick 71.620.0027.56e−70.0003781.83e−119.14e−83.15e−5
sub-brick 85.240.005241.39e−50.002656.39e−94.66e−60.000521
sub-brick 93.170.003179.85e−70.0003112.93e−115.84e−84.19e−5
sub-brick 103.220.003220.002120.6596.94e−50.1340.028
sub-brick 113.150.003150.002150.6817.25e−50.1460.0274
sub-brick 123.10.00310.001840.5955.24e−50.1090.0255
sub-brick 133.10.00317.03e−60.002271.54e−93.21e−60.00032
sub-brick 142.990.002990.001670.5574.34e−50.09690.0222
sub-brick 152.960.002960.001490.5033.57e−50.08160.0196
sub-brick 162.650.002658.93e−83.37e−51.43e−134.08e−101.96e−6
sub-brick 174.170.004170.002310.5539.3e−50.1070.035
sub-brick 183.330.003336.39e−60.001921.11e−92e−60.0002
sub-brick 193.140.003143.06e−50.009743.06e−86.21e−50.00166

θ_v = |roll| + |pitch| + |yaw| of AFNI’s -1Dfile line (degrees); L_v = max(2e-3, 1e-3·θ_v); required: mean |Δ| ≤ L_v times the volume’s largest |AFNI| value and NCC ≥ 1 − 50·L_v² (the spec writer’s step 3 of 2026-11-08: the limit grows with the rotation, as AFNI’s own spread under the ties does); the largest single difference is reported, not required (Fourier’s wrap-around at the field of view’s edges).

Every comparison, computed in float64 over every voxel of the outputs

Comparison (A − B), verdict and reasonmax |Δ|Values that differRMS of Δmax |Δ| ÷ range1 − NCC
The replica against the original
replica (fix) − AFNI 3dvolreg 26.3.05Bit-identicalThe replica gives AFNI’s registered volumes bit for bit (its default corrects AFNI-99, which changes only the maximum displacements).identical: 0 of 2,621,440 values differ
The clean room against the original and the replica
clean room (fix) − AFNI 3dvolreg 26.3.05Within the original's own variabilityNot AFNI’s bits by design: where two axes have the same voxel size the shear decomposition’s axis orders tie in exact arithmetic, and AFNI decides each tie by its last bits at every resampling, which no specification states (specs/3dvolreg.md §5.3, §16); §14.3’s limits allow for either side of every tie, measured from AFNI’s own spread when its ties are decided the other way. On this run the ties meet larger rotations (to 5.2° summed over the three axes), and §14.3’s limit grows with each volume’s rotation: every one of the 20 volumes within §14.3’s limit (the closest, sub-brick 11: mean |Δ| 0.0021 of its largest value against L_v = 0.0032, 0.68 of it; 1 − NCC 7.2e-05, 0.15 of 50·L_v²); 7 of 20 -1Dfile lines identical, the parameters at most 0.12 of their thresholds; 2,255,213 of 2,621,440 values differ, the largest by 26.2 (0.034 of the range) at the rotated field of view’s edge. §14.326.22,255,213 of 2,621,440 (86 %)1.260.03421.83e−5
clean room (reproduce) − AFNI 3dvolreg 26.3.05Within the original's own variabilityNot AFNI’s bits by design: where two axes have the same voxel size the shear decomposition’s axis orders tie in exact arithmetic, and AFNI decides each tie by its last bits at every resampling, which no specification states (specs/3dvolreg.md §5.3, §16); §14.3’s limits allow for either side of every tie, measured from AFNI’s own spread when its ties are decided the other way. On this run the ties meet larger rotations (to 5.2° summed over the three axes), and §14.3’s limit grows with each volume’s rotation: every one of the 20 volumes within §14.3’s limit (the closest, sub-brick 11: mean |Δ| 0.0021 of its largest value against L_v = 0.0032, 0.68 of it; 1 − NCC 7.2e-05, 0.15 of 50·L_v²); 7 of 20 -1Dfile lines identical, the parameters at most 0.12 of their thresholds; 2,255,213 of 2,621,440 values differ, the largest by 26.2 (0.034 of the range) at the rotated field of view’s edge. §14.326.22,255,213 of 2,621,440 (86 %)1.260.03421.83e−5
clean room − replica, both fixWithin the original's own variabilityThe two ports’ defaults (fix), by the same limits: every one of the 20 volumes within §14.3’s limit (the closest, sub-brick 11: mean |Δ| 0.0021 of its largest value against L_v = 0.0032, 0.68 of it; 1 − NCC 7.2e-05, 0.15 of 50·L_v²).26.22,255,213 of 2,621,440 (86 %)1.260.03421.83e−5
The clean room against itself: thread counts, fix and reproduce
clean room, fix − reproduceBit-identicalThe registered volumes are the same in fix and in reproduce: AFNI-99, named by fix, changes only the maximum displacements. AFNI-99identical: 0 of 2,621,440 values differ
clean room, 1 thread − 14 threadsBit-identicalThe same bits at 1 and 14 threads.identical: 0 of 2,621,440 values differ

Δ is A − B over every value of the files (the 20 volumes, not only the brain); each volume against its limit in the charts and tables below.

fMRIPrep’s median of each program’s volumes against fMRIPrep’s own reference of this run

Median ofshare of the limitmean |Δ| ÷ maxlargest |Δ| ÷ maxvoxels beyond 1e-3 of max
AFNI 3dvolreg 26.3.05’s volumesfMRIPrep’s own reference byte for byte (its inputs re-made as fMRIPrep made them)0000
the clean room’s (fix)L_med 0.0052, q 0.01140.1150.0006030.01980.189

§14.3: mean |Δ| ≤ L_med·max + q and NCC ≥ 1 − 50·L_med² (L_med the run’s largest L_v); the largest difference and the share beyond 1e-3 reported. The median moves by at most the largest of its volumes’ differences at each voxel; on this run AFNI’s own tie variants move it by up to 0.17 of the limit (the spec writer’s measurement).

Provenance of this example
  • Data: OpenNeuro ds000256 (Greene et al., 2018), snapshot 00002, sub-CTS203, PDDL-1.0
  • Inputs: RobustAverage’s sliced volumes (sha256 7fc9eb92db4c4301…); fMRIPrep’s BOLD reference (sha256 9631b8e6ce78275c…); fMRIPrep’s BOLD brain mask (sha256 390ba2dcfcb5e919…)
  • Runs: figures cache entry afni-volreg/fmriprep-ds000256-nofeedback-original, computed 2026-11-08T08:38:30+00:00 with larmorx ae74dcb46642; AFNI_26.3.05 (oracles/afni-26.3.05, through the cache of the original’s runs); larmorx 0.0.1 (crates larmorx-afni-replica and larmorx-afni, built in); every command, its environment, threads and exit status are in its RUN_INFO.json

Every number above is in afni-volreg.json, exported by python -m larmorx_validation evidence afni-volreg.

A correction by default: AFNI-99

-maxdisp1D reports, for each volume, the largest displacement of a voxel inside the base’s brain mask, a measure of how much the head moved. AFNI converts the registration’s rotation and shift to the dataset’s axes with each axis’s sign, but takes the voxels’ positions without it. So along an axis stored from left to right, posterior to anterior or superior to inferior, it computes the displacement of the voxel mirrored about the axis’s centre (AFNI-99). The replica’s default, fix, computes the displacements of the voxels themselves, and names AFNI-99 where a value it prints or writes changes (the messages show two decimals, -maxdisp1D three); reproduce gives AFNI’s numbers.

Below, the volumes of the first example, stored as fMRIPrep gives them (their second axis posterior to anterior) and stored again with that axis reversed. AFNI’s maximum displacements depend on the storage order; the replica’s do not. The registered volumes are not affected: fMRIPrep, which uses only them, gets the same reference either way. The clean room’s fix makes the same correction and names it alike (the table under the viewer).

Worked example

AFNI-99: the maximum displacement of a mirrored voxel

What this shows: AFNI’s maximum displacements change with the storage order, by up to 0.011 mm on this run (AFNI-99). In fix, their default, both ports compute the displacements of the voxels themselves and name AFNI-99. In reproduce the replica gives AFNI’s numbers byte for byte; the clean room follows AFNI’s convention, within its derived bounds. The registered volumes do not depend on it: the replica’s are AFNI’s, the clean room’s within §14.3’s limits.

3dvolreg’s -maxdisp1D (and -savedisp) convert the registration’s rotation and shift to the dataset’s axes with each axis’s sign, but take the voxel positions without it: for an axis stored left to right, posterior to anterior or superior to inferior, the displacement computed is that of the voxel mirrored about the axis’s centre. fMRIPrep’s volumes of the first example are stored that way along their second axis. The same volumes stored with that axis reversed (LPS: every axis positive in AFNI’s order), registered by AFNI 26.3.05 and the replica in fix and in reproduce. Beside them LARMORX’s clean room, stored both ways, in fix and in reproduce: its default corrects AFNI-99 as the replica’s does (the table below).

Input
the first example’s sliced volumes; and the same stored integers, scale factor and header, the second axis reversed and the qform and sform changed so that every voxel stays where it was
Command
3dvolreg -Fourier -twopass -zpad 4 -prefix o.nii.gz -1Dfile o.1D -1Dmatrix_save o.aff12.1D -maxdisp1D o_md.1D sliced.nii.gz
Defect
AFNI-99: -maxdisp and -savedisp give the displacements of a mirrored position; corrected by default, named; reproduce gives AFNI’s numbers
Size
on this run, whose head moves little (at most 0.157 mm from the base), AFNI’s maximum displacements change by up to 0.011 mm with the storage order; the error grows with the rotation and with the voxels’ distance from the axis’s centre (on the API page’s synthetic series, 1.718 against 1.672 mm)
fMRIPrep
RobustAverage writes -maxdisp1D but uses only the registered volumes: fMRIPrep’s results do not change
OriginalAFNI 3dvolreg 26.3.05stored LAS · as fMRIPrep gives it
OriginalAFNI 3dvolreg 26.3.05stored LPS · the same volumes, the second axis reversed; shown back in place
Replicareplicalarmorx, Apache-2.0 with CC BY 4.0's attribution · stored LPS, fix (the default) · against AFNI 3dvolreg 26.3.05
Clean roomclean roomlarmorx, Apache-2.0 · stored LPS, shown back in place; fix (the default) · against AFNI 3dvolreg 26.3.05
the original against itself: the same volumes stored another wayAFNI 3dvolreg 26.3.05: stored LPS − LASscale ±0.0266

Bit-identical2,785,280 of 2,785,280 voxels equal

the replica against the originalreplica (fix) − AFNI 3dvolreg 26.3.05, LPS

Axial slice 16 of 34, volume 20 of 20; the last volume. The interactive viewer needs JavaScript; the metrics below are complete without it.

Maximum displacements over the 20 volumes in two storage orders, the replica against the original: AFNI 3dvolreg 26.3.05
  • AFNI 3dvolreg 26.3.05 (the replica in reproduce gives these numbers)
  • replica, fix
stored LAS, as fMRIPrep gives it (mm)0.1570the same volumes stored LPS (mm)0.1520LAS − LPS (mm)0.011−0.004

-maxdisp1D: for each volume, the largest displacement of a voxel of the base’s brain mask (3dAutomask’s, clip fraction 0.33), taken over the mask’s boundary voxels, where a rigid motion moves farthest. Move the pointer over a chart to read every value at one volume.

The maximum displacements compared (-maxdisp1D, over every volume)

Comparisonlargest |Δ| (mm)text
AFNI 3dvolreg 26.3.05: LAS − LPSthe same volumes; AFNI’s maximum displacement depends on the storage order0.011–
replica, fix: LAS − LPSthe voxels’ own displacements, in either storage order0AFNI-99 corrected, named
replica, fix − AFNI 3dvolreg 26.3.05, stored LASthe second axis is stored posterior to anterior: AFNI mirrors the voxels along it0.011AFNI-99 corrected, named
replica, fix − AFNI 3dvolreg 26.3.05, stored LPSevery axis positive in AFNI’s order: nothing to correct0identical as text
replica, reproduce − AFNI 3dvolreg 26.3.05, both storagesAFNI’s numbers, byte for byte0identical as text

-maxdisp1D writes 3 decimals. The registration itself does not depend on the storage order beyond the last bits of the parameters (the viewer above).

Every comparison, computed in float64 over every voxel of the outputs

Comparison (A − B), verdict and reasonmax |Δ|Values that differRMS of Δmax |Δ| ÷ range1 − NCC
The original against itself: the yardstick
AFNI 3dvolreg 26.3.05: stored LPS − LASreferenceThe same registration in the other storage order: each volume is resampled along the second axis in the other direction, so the parameters differ in their last digits and 1,685,216 of 2,785,280 values by up to 0.0266 (3.3e-05 of the range). AFNI-990.02661,685,216 of 2,785,280 (60.5 %)0.001123.34e−51.16e−11
The replica against the original
replica (fix) − AFNI 3dvolreg 26.3.05, LPSBit-identicalThe same values, every one: AFNI-99 changes no registered volume. AFNI-99identical: 0 of 2,785,280 values differ
replica (reproduce) − AFNI 3dvolreg 26.3.05, LPSBit-identicalThe same values, every one. AFNI-99identical: 0 of 2,785,280 values differ
The clean room against the original
clean room (fix) − AFNI 3dvolreg 26.3.05, LPSWithin the original's own variabilityNot AFNI’s bits by design: where two axes have the same voxel size the shear decomposition’s axis orders tie in exact arithmetic, and AFNI decides each tie by its last bits at every resampling, which no specification states (specs/3dvolreg.md §5.3, §16); §14.3’s limits allow for either side of every tie, measured from AFNI’s own spread when its ties are decided the other way. Stored LPS: every one of the 20 volumes within §14.3’s limit (the closest, sub-brick 5: mean |Δ| 6.6e-07 of its largest value against L_v = 0.002, 0.00033 of it; 1 − NCC 1.4e-11, 6.9e-08 of 50·L_v²). §14.30.01781,675,404 of 2,785,280 (60.2 %)0.0005032.24e−52.31e−12
clean room (reproduce) − AFNI 3dvolreg 26.3.05, LPSWithin the original's own variabilityNot AFNI’s bits by design: where two axes have the same voxel size the shear decomposition’s axis orders tie in exact arithmetic, and AFNI decides each tie by its last bits at every resampling, which no specification states (specs/3dvolreg.md §5.3, §16); §14.3’s limits allow for either side of every tie, measured from AFNI’s own spread when its ties are decided the other way. Stored LPS: every one of the 20 volumes within §14.3’s limit (the closest, sub-brick 5: mean |Δ| 6.6e-07 of its largest value against L_v = 0.002, 0.00033 of it; 1 − NCC 1.4e-11, 6.9e-08 of 50·L_v²). §14.30.01781,675,404 of 2,785,280 (60.2 %)0.0005032.24e−52.31e−12
The clean room against itself: the same volumes stored another way
clean room: stored LPS − LASWithin the original's own variabilityThe same volumes stored two ways: with the second axis reversed the shear decomposition meets its ties on other rows, so the clean room’s two registrations differ as each differs from AFNI’s, and §14.3’s limits hold between them: every one of the 20 volumes within §14.3’s limit (the closest, sub-brick 15: mean |Δ| 1.3e-06 of its largest value against L_v = 0.002, 0.00066 of it; 1 − NCC 5.4e-11, 2.7e-07 of 50·L_v²); 1,678,629 values differ, by at most 0.0298. AFNI’s own two storages differ by up to 0.0266 (the comparison above): the registration depends on the storage order through its last bits.0.02981,678,629 of 2,785,280 (60.3 %)0.0009023.73e−57.57e−12

Δ is A − B, computed in float64 over every value of the files (2,785,280), not only those shown; the range is B’s maximum minus its minimum. The records’ thresholds: every output bit-identical to AFNI 26.3.05’s (the registered series, -1Dfile, -1Dmatrix_save, -maxdisp1D), with its exit status and messages (the replica); AFNI’s exit status, files, headers and messages, the motion parameters within per-volume limits derived from each volume’s own parameters, and the outputs computed from them within the bounds those limits give (the clean room, specs/3dvolreg.md §14.3). The row the original against itself is AFNI’s own variability on this input: the same volumes stored another way.

The clean room’s maximum displacements (-maxdisp1D, over every volume)

Comparisonlargest |Δ| (mm)largest share of the derived boundnotices
clean room, fix: LAS − LPSthe voxels’ own displacements, in either storage order0–LAS: AFNI-99 named; LPS: nothing named
clean room, fix − AFNI 3dvolreg 26.3.05, stored LASAFNI mirrors the voxels along the axis stored posterior to anterior0.011–AFNI-99 corrected, named
clean room, fix − AFNI 3dvolreg 26.3.05, stored LPSevery axis positive in AFNI’s order: nothing to correct00within the derived bounds
clean room, reproduce − AFNI 3dvolreg 26.3.05, stored LASAFNI’s convention00within the derived bounds
clean room, reproduce − AFNI 3dvolreg 26.3.05, stored LPSAFNI’s convention00within the derived bounds
clean room − replica, both fix, stored LASboth correct AFNI-9900within the derived bounds

-maxdisp1D writes 3 decimals. Where the parameters differ (the shear decomposition’s ties, the first example), the maximum displacements are held to the bound that follows from each volume’s own parameter differences, alpha_v·r + tau_v + 0.001 (§14.3 ‘Derived quantities’); the correction of AFNI-99 in fix changes the numbers where the second axis is stored posterior to anterior, and names it.

Provenance of this example
  • Data: OpenNeuro ds000005 (Tom et al., 2007), snapshot 00001, sub-01, PDDL-1.0
  • Inputs: RobustAverage’s sliced volumes (sha256 2cce88990bcfbee8…); the raw BOLD run (sha256 609ae0a21822318f…); fMRIPrep’s BOLD reference (sha256 ab49738164116b9c…); the sliced volumes stored LPS (sha256 71708c457976d2d1…)
  • Runs: figures cache entry afni-volreg/fmriprep-ds000005-run-01, computed 2026-11-07T01:27:24+00:00 with larmorx c0fa9881ce74; AFNI_26.3.05, gcc (Ubuntu 11.4.0-1ubuntu1~22.04.3) 11.4.0; AFNI_25.2.09, gcc (Ubuntu 11.4.0-1ubuntu1~22.04.3) 11.4.0; AFNI_25.2.09 (Aug 24 2025) [64-bit]; larmorx 0.0.1 (crate larmorx-afni-replica, built in); every command, its environment, threads and exit status are in its RUN_INFO.json
  • Runs of the clean room: figures cache entry afni-volreg/fmriprep-ds000005-run-01-original, computed 2026-11-08T08:38:25+00:00 with larmorx ae74dcb46642 (python -m larmorx with LARMORX_IMPLEMENTATION=original: fix, the default at 14 threads and at one, reproduce at 14; on the volumes stored LPS both modes), on the replica’s entry’s inputs; every command, its environment, threads and exit status are in its RUN_INFO.json

Every number above is in afni-volreg.json, exported by python -m larmorx_validation evidence afni-volreg.

Random testing

The clean room’s random-testing run of half an hour (2026-11-08): random moving series with every option, each case run by AFNI 26.3.05, by the clean room in fix and in reproduce and by the replica’s fix, and compared as the record compares (random testing).

Random testing, the clean room against the original: AFNI 3dvolreg 26.3.05 (a run of half an hour, 2026-11-08)

Cases
random cases, each compared with AFNI 26.3.05 (through the cache of the original’s runs) and with the replica’s default21,710
agreeing in the run (§14.3’s fixed requirements)11,385
differing in the run10,325
fMRIPrep’s options (-Fourier, -twopass)5,208 / 8,635
the other methods: heptic, quintic, cubic, linear2,178 / 2,549 / 2,589 / 2,646
-base, -weight, -final, -tshift7,209 / 2,157 / 4,319 / 3,254
-savedisp, -SAVEDISP, -dfile, -maxdisp1D3,168 / 3,138 / 10,887 / 12,586
distinct options drawn; features covered34 options; 115 features

The random-testing run (seed volreg-original-trial, larmorx 6d4f2edb8992, 4 jobs at 1 thread, 30 minutes, every case in a memory box of 3 GiB): random moving series and every option, each case run by AFNI 26.3.05, the clean room in fix and in reproduce and the replica’s default, and compared as the record compares. The differences were judged again with the part’s final build (the next table).

Most differences come from the ties: random testing’s small phantoms with sharp edges make the fits ill-conditioned with respect to them, so AFNI’s own result moves with the side each tie takes. For such cases the specification holds the clean room to three times its own spread over twelve seeded tie draws (LARMORX_VOLREG_TIES=random:<seed>, a validation switch, not a setting), and a runaway fit (a volume beyond 15° or 20 mm) to its output’s structure. A seeded sample of the differences, judged again with the final build, and every kind of difference with what it was:

A seeded sample of 1,200 of the run’s differences, judged again with the final build

Judgement (specs/3dvolreg.md §14.3)Cases
within 3 times the clean room’s own spread under the ties (12 seeded draws, LARMORX_VOLREG_TIES=random:1 to 12; 9 times for 1 − NCC)860
agree with §14.3’s fixed requirements (fixed since the run)67
runaways (a volume beyond 15° or 20 mm, or AFNI-97): the files, names, headers, types, dimensions and masked messages agree268
explained, one by one (3 name AFNI-99 at a two-decimal rounding boundary; 2 lie at the edge of 3 times the 12-draw spread and within it with 36 draws)5

validation/investigations/afni-volreg-original/triage.py (the descriptor’s own judgement) on the sample drawn with seed 20261108: §14.3’s fixed requirements first, then ‘The ties’ spread’ (the spec writer’s step 3): random testing’s small phantoms with sharp edges make the fits ill-conditioned with respect to the shear decomposition’s ties, so AFNI’s own result moves with them (its parameters by a median of 25 thresholds over its tie variants). The counts: validation/investigations/afni-volreg-original/retriage-2026-11-08.json.

The run’s differences, by what differed, each in one category

CategoryWhat differed (the checks)
explained: The shear decomposition’s exact ties (specs/3dvolreg.md §5.3, §16): AFNI decides each resampling’s tie its own way, and on random testing’s small phantoms with sharp edges the fits are ill-conditioned with respect to them (the spec writer: AFNI’s own parameters move by a median of 25 thresholds over its tie variants).…AFNI: -NDfile parameters, the replica (fix): -NDfile parameters …
explained: The parameters’ summary (++ Min/Mean/Max) and the derivative images’ ranges with -verbose -verbose follow the parameters and the ties (the entry above): within 3 times the ties’ spread in every sampled case; the largest displacement is held to the larger of its two sub-bricks’ bounds (a check corrected in relay 4); an…AFNI: messages, the replica (fix): messages
explained: -rotcom’s 3drotate commands print the parameters: within 3 times the ties’ spread where the parameters are (the first entry).AFNI: standard output (-rotcom), the replica (fix): standard output (-rotcom)
a LARMORX mistake, fixed: -savedisp’s and -SAVEDISP’s datasets did not carry the input’s slice fields (slice_code, slice_duration) as the registered output does: fixed after the run (the suite’s tshift and outputs cases cover it).AFNI: -savedisp DX header, AFNI: -savedisp DY header …
a LARMORX mistake, fixed: §10.3’s arithmetic applied to AFNI’s own raw -savedisp values gave AFNI’s three normalised datasets bit for bit but its sixteen products differed at a few voxels: the line’s coefficients in the spec writer’s double order (§10.3 step 1, step 3 of 2026-11-08) give AFNI’s nineteen datasets bit for bit (savedisp.py --orig…-SAVEDISP step N: the normalisation, and the products of one or two factors, on AFNI’s raw displacements
explained: Fixed after the run: the option scan’s notices (AFNI-86, -90, -95) were not printed when a run stopped with a later error, AFNI-97 was never named, AFNI-99 was named from -maxdisp1D’s three decimals when no file was written, and (relay 4) AFNI-92 was named for a run that stopped before its first pass and the notices…defects named as the replica’s default
explained: AFNI-97: AFNI’s runaway Fourier first passes (rotations of hundreds of degrees) met a lone row shifted beyond the row and AFNI crashed; whether a run meets the condition is the ties’ choice (the spec writer: in 2 to 12 of AFNI’s 16 tie variants). The clean room names AFNI-97 in fix and stops as AFNI does with reprod…AFNI: exit status, AFNI: files written …
a LARMORX mistake, fixed: AFNI-99 named where no printed or written value changed (the comparison used -maxdisp1D’s three decimals also when no -maxdisp1D file was written): fixed after the run.the fix pass
the comparison’s own mistake, corrected: The part’s Python extension was replaced in place while the run went on: a run that had it mapped ended with SIGBUS. The extension is now replaced by an atomic rename.The replica (fix): exit status, the replica (fix): files written

validation/difftest/triage.toml (the entries of afni-volreg-original); the findings: docs/findings/afni-volreg-clean-room.md.

Speed

The clean room is built for speed. It does the shears by sweeps over planes and the Fourier shears by a batched FFT, runs the coarse search on its support box, and divides the work inside each volume between threads. It is measured against AFNI and the replica on fMRIPrep’s calls (the benchmark report). On one thread it runs them 2.8 to 5.2 times as fast as AFNI, and on 14 threads 9 to 15 times; the replica, which computes as AFNI does, 1.6 to 3 times.

Speed, both ports against the original: AFNI 3dvolreg 26.3.05

Seconds for the whole call (reading, both passes, writing), on a log scale. The factor is how many times faster the clean room is than AFNI, both on one thread.

  • AFNI 3dvolreg 26.3.05, one thread on this path
  • replica, one thread, as AFNI
  • clean room, 1 thread
  • clean room, 14 threads
  1. fMRIPrep's call, ds000005 run 1 (64 × 64 × 34 × 20)AFNI 3dvolreg 26.3.05, one thread on this path 9.03 sreplica, one thread, as AFNI 3.39 sclean room, 1 thread 1.75 sclean room, 14 threads 0.61 s5.2×
  2. fMRIPrep's call, ds003592 echo 1 (64 × 64 × 43 × 20)AFNI 3dvolreg 26.3.05, one thread on this path 11.1 sreplica, one thread, as AFNI 3.69 sclean room, 1 thread 2.47 sclean room, 14 threads 0.82 s4.5×
  3. fMRIPrep's call, ds004958 (104 × 104 × 72 × 3, uint16)AFNI 3dvolreg 26.3.05, one thread on this path 6.35 sreplica, one thread, as AFNI 3.39 sclean room, 1 thread 1.96 sclean room, 14 threads 0.6 s3.2×
  4. fMRIPrep's call, ds005040 PEPOLAR average (64 × 64 × 40 × 2)AFNI 3dvolreg 26.3.05, one thread on this path 1.17 sreplica, one thread, as AFNI 0.45 sclean room, 1 thread 0.25 sclean room, 14 threads 0.13 s4.7×
  5. fMRIPrep's call, ds005454 (160 × 160 × 96 × 20, multiband)AFNI 3dvolreg 26.3.05, one thread on this path 118 sreplica, one thread, as AFNI 72.8 sclean room, 1 thread 42.1 sclean room, 14 threads 10.5 s2.8×
0.1 s1 s10 s100 s1000 s

From the benchmark report docs/benchmarks/afni-volreg.md (the main machine: AMD Ryzen 9 7945HX, 16 cores / 32 threads; 2026-11-08): the median of three runs, AFNI timed afresh (ds005454: one AFNI and replica run, three of the clean room). The machine was shared with another agent’s random testing (load 23 to 26 of 32 threads), so every time is long and the many-thread times most; the ratios between programs measured together are the information. The clean room’s outputs are the same bits at every thread count.

Wall-clock seconds of the whole program, both ports against the original: AFNI 3dvolreg 26.3.05 (the benchmark report; in brackets the speed-up against AFNI)

CallAFNI 26.3.05replica (1 thread)clean room (1 thread)clean room (4 threads)clean room (14 threads)
ds000005 run 1 (64 × 64 × 34 × 20)9.03 s3.39 s (2.7×)1.75 s (5.2×)0.76 s (12×)0.61 s (15×)
ds003592 echo 1 (64 × 64 × 43 × 20)11.1 s3.69 s (3×)2.47 s (4.5×)1.1 s (10×)0.82 s (14×)
ds004958 (104 × 104 × 72 × 3, uint16)6.35 s3.39 s (1.9×)1.96 s (3.2×)0.76 s (8.4×)0.6 s (11×)
ds005040 PEPOLAR average (64 × 64 × 40 × 2)1.17 s0.45 s (2.6×)0.25 s (4.7×)0.14 s (8.4×)0.13 s (9×)
ds005454 (160 × 160 × 96 × 20, multiband)118 s72.8 s (1.6×)42.1 s (2.8×)15.4 s (7.7×)10.5 s (11×)

The benchmark report docs/benchmarks/afni-volreg.md (perf.py --afni-fresh --repeat 3, inside memory boxes). AFNI’s 3dvolreg and the replica run on one thread (3dvolreg has no parallel loop on this path); the clean room registers the volumes one after another and divides the work inside each (the shears, the coarse search, the blurs, the derivative images).

Peak resident memory, both ports against the original: AFNI 3dvolreg 26.3.05 (MiB; the rule: at most 1.1 × AFNI’s peak + 8 MiB, at one thread)

CallAFNIallowancereplica (1 thread)clean room (1 thread)clean room (4 threads)clean room (14 threads)
ds000005 run 123.133.427.725.830.738.3
ds003592 echo 128.13932.230.939.846.5
ds00495854.267.665.160.960.766.2
ds005040 PEPOLAR15.525.117.516.817.220.7
ds005454311350338332340346

The verdict is the rule at one thread (CLAUDE.md, ‘Memory similar to the original’s’; held by tests/parity/test_memory_peak.py on fMRIPrep’s call and on ds005454, both modes). With threads the extra memory is about 1 MiB per thread (each thread’s plane and batch buffers): no full-size copy per thread.

The replica’s first timings (2026-11-06, informal, before the benchmark report):

Speed, the replica against the original: AFNI 3dvolreg 26.3.05, fMRIPrep’s RobustAverage call (informal timings of 2026-11-06)
  • AFNI 3dvolreg 26.3.05, 1 thread
  • replica, larmorx afni 3dvolreg
  • replica, python -m larmorx
  1. ds000005 run 1, single-band (64 × 64 × 34 × 20)AFNI 3dvolreg 26.3.05, 1 thread 8.95 sreplica, larmorx afni 3dvolreg 2.88 sreplica, python -m larmorx 2.66 s3.1×
  2. ds003592 echo 1, multi-echo (64 × 64 × 43 × 20)AFNI 3dvolreg 26.3.05, 1 thread 9.79 sreplica, larmorx afni 3dvolreg 3.21 sreplica, python -m larmorx 3.18 s3.1×
  3. ds005454, multiband (160 × 160 × 96 × 20)AFNI 3dvolreg 26.3.05, 1 thread 93.2 sreplica, larmorx afni 3dvolreg 40.8 sreplica, python -m larmorx 41 s2.3×
  4. ds004958, multiband 8 (104 × 104 × 72 × 3)AFNI 3dvolreg 26.3.05, 1 thread 4.62 sreplica, larmorx afni 3dvolreg 2.05 sreplica, python -m larmorx 1.99 s2.2×
  5. ds005040, PEPOLAR average (64 × 64 × 40 × 2)AFNI 3dvolreg 26.3.05, 1 thread 0.831 sreplica, larmorx afni 3dvolreg 0.344 sreplica, python -m larmorx 0.351 s2.4×
0.1 s1 s10 s100 s

fMRIPrep’s call (-Fourier -twopass -zpad 4 with -1Dfile, -1Dmatrix_save, -maxdisp1D) on the comparison suite’s inputs, one thread each (3dvolreg has no parallel loop on this path), the median of three runs, reading and writing compressed NIfTI included, every output checked byte for byte; on the main machine while another agent worked: indications, not a benchmark report. The ratio is AFNI’s time over the replica’s program’s. Tool: validation/investigations/afni-volreg/speed.py.

Every case

Each case of the two records, with what it tests, its command line, its verdict and the reason. Search, filter (or select a segment of a chart above), sort; #case=<name> links to one (the clean room’s cases are named original/<case>).

The replica’s record:

190 cases

CaseCategoryWhat it testsVerdictChecks
interpolation/defaultinterpolation methodsthe defaults (heptic), every text output
command lineOMP_NUM_THREADS=1 3dvolreg -prefix o.nii -1Dfile p.1D -dfile d.txt -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1.nii.gz>
Bit-identical19/19
interpolation/Fourierinterpolation methods-Fourier
command lineOMP_NUM_THREADS=1 3dvolreg -Fourier -prefix o.nii -1Dfile p.1D -dfile d.txt <generated: s1.nii.gz>
Bit-identical13/13
interpolation/cubicinterpolation methods-cubic
command lineOMP_NUM_THREADS=1 3dvolreg -cubic -prefix o.nii -1Dfile p.1D -dfile d.txt <generated: s1.nii.gz>
Bit-identical13/13
interpolation/quinticinterpolation methods-quintic
command lineOMP_NUM_THREADS=1 3dvolreg -quintic -prefix o.nii -1Dfile p.1D -dfile d.txt <generated: s1.nii.gz>
Bit-identical12/12
interpolation/linearinterpolation methods-linear
command lineOMP_NUM_THREADS=1 3dvolreg -linear -prefix o.nii -1Dfile p.1D -dfile d.txt <generated: s1.nii.gz>
Bit-identical12/12
interpolation/hepticinterpolation methods-heptic
command lineOMP_NUM_THREADS=1 3dvolreg -heptic -prefix o.nii -1Dfile p.1D -dfile d.txt <generated: s1.nii.gz>
Bit-identical12/12
interpolation/fourier-lowercaseinterpolation methods-fourier: AFNI compares four letters
command lineOMP_NUM_THREADS=1 3dvolreg -fourier -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
interpolation/final-NNinterpolation methods-final NN
command lineOMP_NUM_THREADS=1 3dvolreg -final NN -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical11/11
interpolation/final-Fourierinterpolation methods-final Fourier
command lineOMP_NUM_THREADS=1 3dvolreg -final Fourier -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
interpolation/final-cubicinterpolation methods-final cubic
command lineOMP_NUM_THREADS=1 3dvolreg -final cubic -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
interpolation/final-linearinterpolation methods-final linear
command lineOMP_NUM_THREADS=1 3dvolreg -final linear -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
interpolation/final-quinticinterpolation methods-final quintic
command lineOMP_NUM_THREADS=1 3dvolreg -final quintic -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
interpolation/final-hepticinterpolation methods-final heptic
command lineOMP_NUM_THREADS=1 3dvolreg -final heptic -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
interpolation/final-cubic-noclipinterpolation methods-final cubic -noclip
command lineOMP_NUM_THREADS=1 3dvolreg -final cubic -noclip -prefix o.nii <generated: s1.nii.gz>
Bit-identical9/9
interpolation/Fourier-noclipinterpolation methods-Fourier -noclip
command lineOMP_NUM_THREADS=1 3dvolreg -Fourier -noclip -prefix o.nii <generated: s1.nii.gz>
Bit-identical8/8
interpolation/clipitinterpolation methods-clipit (the default, a notice), -wtrim, -coarserot
command lineOMP_NUM_THREADS=1 3dvolreg -clipit -wtrim -coarserot -prefix o.nii <generated: s1.nii.gz>
Bit-identical8/8
interpolation/zpad-0interpolation methods-zpad 0
command lineOMP_NUM_THREADS=1 3dvolreg -zpad 0 -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical11/11
interpolation/zpad-2interpolation methods-zpad 2
command lineOMP_NUM_THREADS=1 3dvolreg -zpad 2 -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
interpolation/zpad-8interpolation methods-zpad 8
command lineOMP_NUM_THREADS=1 3dvolreg -zpad 8 -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
interpolation/zpad-twiceinterpolation methods-zpad 3 -zpad 2: the last one
command lineOMP_NUM_THREADS=1 3dvolreg -zpad 3 -zpad 2 -prefix o.nii <generated: s1.nii.gz>
Bit-identical8/8
interpolation/Fourier-zpad-0interpolation methods-Fourier -zpad 0
command lineOMP_NUM_THREADS=1 3dvolreg -Fourier -zpad 0 -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
base/indexthe base volume (-base)-base 3
command lineOMP_NUM_THREADS=1 3dvolreg -base 3 -prefix o.nii -1Dmatrix_save m.aff12.1D <generated: s1.nii.gz>
Bit-identical11/11
base/lastthe base volume (-base)-base 5 (the last volume)
command lineOMP_NUM_THREADS=1 3dvolreg -base 5 -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
base/datasetthe base volume (-base)-base <3D dataset>
command lineOMP_NUM_THREADS=1 3dvolreg -base <generated: base3d.nii.gz> -prefix o.nii -1Dmatrix_save m.aff12.1D -dfile d.txt <generated: s1.nii.gz>
Bit-identical12/12
base/dataset-twopassthe base volume (-base)-twopass -base <3D dataset>
command lineOMP_NUM_THREADS=1 3dvolreg -twopass -base <generated: base3d.nii.gz> -prefix o.nii -1Dmatrix_save m.aff12.1D <generated: s1.nii.gz>
Bit-identical10/10
base/dataset-4dthe base volume (-base)-base <4D dataset>: its first volume
command lineOMP_NUM_THREADS=1 3dvolreg -base <generated: s2_short.nii.gz> -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Rejected by both

AFNI refuses it (exit 1), and so does the replica in reproduce, with AFNI’s messages; its default, fix, accepts it and names the defect (the record’s list of what fix changed).

5/5
base/3d-inputthe base volume (-base)a 3D input with an external base
command lineOMP_NUM_THREADS=1 3dvolreg -tshift 0 -base <generated: s1.nii.gz> -prefix o.nii <generated: base3d.nii.gz>
Bit-identical8/8
base/vol3d-inputthe base volume (-base)a 3D input, the base a moving series
command lineOMP_NUM_THREADS=1 3dvolreg -base <generated: s1.nii.gz> -prefix o.nii -1Dfile p.1D <generated: e_vol3d.nii.gz>
Bit-identical10/10
twopass/twopasstwo passes and the coarse search (-twopass)-twopass
command lineOMP_NUM_THREADS=1 3dvolreg -twopass -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical11/11
twopass/twoblur-3two passes and the coarse search (-twopass)-twopass -twoblur 3
command lineOMP_NUM_THREADS=1 3dvolreg -twopass -twoblur 3 -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical11/11
twopass/coarse-1-1two passes and the coarse search (-twopass)-twopass -coarse 1 1
command lineOMP_NUM_THREADS=1 3dvolreg -twopass -coarse 1 1 -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
twopass/coarse-0-0two passes and the coarse search (-twopass)-twopass -coarse 0 0
command lineOMP_NUM_THREADS=1 3dvolreg -twopass -coarse 0 0 -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical

in fix, its default, the replica computes with correctly rounded maths where glibc 2.35’s is not correctly rounded (GLIBC-07, not named): 53 values change (the record’s list of what fix changed)

10/10
twopass/sinittwo passes and the coarse search (-twopass)-twopass -sinit
command lineOMP_NUM_THREADS=1 3dvolreg -twopass -sinit -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical11/11
twopass/nomaxdisptwo passes and the coarse search (-twopass)-twopass -nomaxdisp: AFNI-92's clip fraction
command lineOMP_NUM_THREADS=1 3dvolreg -twopass -nomaxdisp -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
twopass/nocoarserottwo passes and the coarse search (-twopass)-nocoarserot -twopass: AFNI reads -noclip (AFNI-86)
command lineOMP_NUM_THREADS=1 3dvolreg -nocoarserot -twopass -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical

in fix, its default, the replica corrects AFNI-86 and names it: 123,278 values change (the record’s list of what fix changed)

10/10
twopass/twopass-weighttwo passes and the coarse search (-twopass)-twopass -weight
command lineOMP_NUM_THREADS=1 3dvolreg -twopass -weight {1}[0] -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
twopass/twopass-wtinptwo passes and the coarse search (-twopass)-twopass -wtinp
command lineOMP_NUM_THREADS=1 3dvolreg -twopass -wtinp -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
twopass/twopass-maxite-2two passes and the coarse search (-twopass)-twopass -maxite 2: raised to 66
command lineOMP_NUM_THREADS=1 3dvolreg -twopass -maxite 2 -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
twopass/fmriprep-calltwo passes and the coarse search (-twopass)fMRIPrep's call on the synthetic series
command lineOMP_NUM_THREADS=1 3dvolreg -Fourier -twopass -1Dfile o.1D -1Dmatrix_save o.aff12.1D -prefix o.nii.gz -zpad 4 -maxdisp1D o_md.1D <generated: s1.nii.gz>
Bit-identical17/17
weights/wtinpweights and edging-wtinp
command lineOMP_NUM_THREADS=1 3dvolreg -wtinp -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical11/11
weights/weightweights and edging-weight 'wset[1]'
command lineOMP_NUM_THREADS=1 3dvolreg -weight {1}[1] -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical11/11
weights/edging-3weights and edging-edging 3
command lineOMP_NUM_THREADS=1 3dvolreg -edging 3 -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical11/11
weights/edging-10pcweights and edging-edging 10%
command lineOMP_NUM_THREADS=1 3dvolreg -edging 10% -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
weights/edging-0weights and edging-edging 0
command lineOMP_NUM_THREADS=1 3dvolreg -edging 0 -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
weights/controlsweights and edging-maxite 2 -x_thresh 0.1 -rot_thresh 0.1 -delta 1
command lineOMP_NUM_THREADS=1 3dvolreg -maxite 2 -x_thresh 0.1 -rot_thresh 0.1 -delta 1 -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical11/11
weights/paramsweights and edging-params 3 0.05 0.05 0.5
command lineOMP_NUM_THREADS=1 3dvolreg -params 3 0.05 0.05 0.5 -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
weights/controls-nonpositiveweights and edging-maxite 0 -delta -1: ignored
command lineOMP_NUM_THREADS=1 3dvolreg -maxite 0 -delta -1 -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
environment/edging-envenvironment variablesAFNI_VOLREG_EDGING=10
command lineOMP_NUM_THREADS=1 AFNI_VOLREG_EDGING=10 3dvolreg -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
environment/edging-env-pcenvironment variablesAFNI_VOLREG_EDGING=12%
command lineOMP_NUM_THREADS=1 AFNI_VOLREG_EDGING=12% 3dvolreg -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
environment/blur-firfacenvironment variables-twopass with AFNI_BLUR_FIRFAC=1.5
command lineOMP_NUM_THREADS=1 AFNI_BLUR_FIRFAC=1.5 3dvolreg -twopass -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
environment/blur-firoldenvironment variables-twopass with AFNI_BLUR_FIROLD=YES
command lineOMP_NUM_THREADS=1 AFNI_BLUR_FIROLD=YES 3dvolreg -twopass -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
environment/rota-zpadenvironment variablesAFNI_ROTA_ZPAD=7: not read
command lineOMP_NUM_THREADS=1 AFNI_ROTA_ZPAD=7 3dvolreg -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
outputs/rotcomoutputs-rotcom -prefix NULL
command lineOMP_NUM_THREADS=1 3dvolreg -rotcom -prefix NULL -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical8/8
outputs/prefix-nulloutputs-prefix NULL with every text output
command lineOMP_NUM_THREADS=1 3dvolreg -prefix NULL -dfile d.txt -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1.nii.gz>
Bit-identical14/14
outputs/matrix-no-suffixoutputs-1Dmatrix_save m: .aff12.1D added
command lineOMP_NUM_THREADS=1 3dvolreg -1Dmatrix_save m -prefix o.nii <generated: s1.nii.gz>
Bit-identical10/10
outputs/maxdisp1D-stdoutoutputs-maxdisp1D -: standard output, and the file -_delt
command lineOMP_NUM_THREADS=1 3dvolreg -maxdisp1D - -prefix o.nii <generated: s1.nii.gz>
Bit-identical10/10
outputs/floatoutputs-float on int16
command lineOMP_NUM_THREADS=1 3dvolreg -float -prefix o.nii <generated: s2_short.nii.gz>
Bit-identical9/9
outputs/nii-gzoutputsa compressed output
command lineOMP_NUM_THREADS=1 3dvolreg -prefix o.nii.gz <generated: s1.nii.gz>
Bit-identical8/8
outputs/savedispoutputs-savedisp
command lineOMP_NUM_THREADS=1 3dvolreg -savedisp sd.nii.gz -prefix o.nii <generated: s1.nii.gz>
Bit-identical15/15
outputs/SAVEDISPoutputs-SAVEDISP -verbose: 19 datasets
command lineOMP_NUM_THREADS=1 3dvolreg -SAVEDISP SD.nii -prefix o.nii -verbose <generated: s1.nii.gz>
Bit-identical46/46
outputs/savedisp-obliqueoutputs-savedisp -prefix NULL on an oblique series
command lineOMP_NUM_THREADS=1 3dvolreg -savedisp sd.nii.gz -prefix NULL <generated: s3_oblique.nii.gz>
Bit-identical

in fix, its default, the replica corrects AFNI-99 and names it: 568,675 values change (the record’s list of what fix changed)

12/12
outputs/savedisp-bareoutputs-savedisp .nii
command lineOMP_NUM_THREADS=1 3dvolreg -savedisp .nii -prefix o.nii <generated: s1.nii.gz>
Rejected by both

AFNI refuses it (exit 1), and so does the replica in reproduce, with AFNI’s messages; its default, fix, accepts it and names the defect (the record’s list of what fix changed).

5/5
tshift/tshift-0slice timing first (-tshift)-tshift 0 on sequential slices
command lineOMP_NUM_THREADS=1 3dvolreg -tshift 0 -prefix o.nii -1Dfile p.1D <generated: s6_st.nii.gz>
Bit-identical11/11
tshift/tshift-1slice timing first (-tshift)-tshift 1
command lineOMP_NUM_THREADS=1 3dvolreg -tshift 1 -prefix o.nii -1Dfile p.1D <generated: s6_st.nii.gz>
Bit-identical10/10
tshift/tshift-verboseslice timing first (-tshift)-verbose -tshift 2 -base 1
command lineOMP_NUM_THREADS=1 3dvolreg -verbose -tshift 2 -base 1 -prefix o.nii -1Dfile p.1D <generated: s6_st.nii.gz>
Bit-identical10/10
tshift/tshift-twopassslice timing first (-tshift)-tshift 0 -twopass -Fourier
command lineOMP_NUM_THREADS=1 3dvolreg -tshift 0 -twopass -Fourier -prefix o.nii -dfile d.txt <generated: s6_st.nii.gz>
Bit-identical10/10
tshift/tshift-negative-slopeslice timing first (-tshift)-tshift 0 on int16 with a negative slope (AFNI-01) and alternating slices
command lineOMP_NUM_THREADS=1 3dvolreg -tshift 0 -prefix o.nii -1Dfile p.1D <generated: s6_stneg.nii.gz>
Bit-identical

in fix, its default, the replica corrects AFNI-01 and names it: 253,264 values change (the record’s list of what fix changed)

10/10
tshift/tshift-no-timingslice timing first (-tshift)-tshift 0 without slice times
command lineOMP_NUM_THREADS=1 3dvolreg -tshift 0 -prefix o.nii <generated: s1.nii.gz>
Bit-identical8/8
tshift/tshift-ignore-largeslice timing first (-tshift)-tshift 0 on 4 volumes: ignore too large
command lineOMP_NUM_THREADS=1 3dvolreg -tshift 0 -prefix o.nii <generated: s6_st4.nii.gz>
Bit-identical8/8
tshift/tshift-savedispslice timing first (-tshift)-tshift 0 -savedisp
command lineOMP_NUM_THREADS=1 3dvolreg -tshift 0 -savedisp sd.nii -prefix o.nii <generated: s6_st.nii.gz>
Bit-identical14/14
rotparent/rotparentanother dataset's motion or grid (-rotparent, -gridparent)-rotparent
command lineOMP_NUM_THREADS=1 3dvolreg -rotparent <generated: rp.nii.gz> -prefix o.nii -1Dfile p.1D -dfile d.txt <generated: s1.nii.gz>
Bit-identical13/13
rotparent/rotparent-gridparentanother dataset's motion or grid (-rotparent, -gridparent)-verbose -rotparent -gridparent (27 slices)
command lineOMP_NUM_THREADS=1 3dvolreg -verbose -rotparent <generated: rp.nii.gz> -gridparent <generated: gp27.nii.gz> -prefix o.nii <generated: s1.nii.gz>
Bit-identical8/8
rotparent/gridparent-24another dataset's motion or grid (-rotparent, -gridparent)-rotparent -gridparent (24 slices) -Fourier -final cubic
command lineOMP_NUM_THREADS=1 3dvolreg -rotparent <generated: rp.nii.gz> -gridparent <generated: gp24.nii.gz> -prefix o.nii -Fourier -final cubic <generated: s1.nii.gz>
Bit-identical8/8
rotparent/gridparent-movedanother dataset's motion or grid (-rotparent, -gridparent)-rotparent -gridparent (moved origin) -noclip -final linear
command lineOMP_NUM_THREADS=1 3dvolreg -rotparent <generated: rp.nii.gz> -gridparent <generated: gp27m.nii.gz> -prefix o.nii -noclip -final linear <generated: s1.nii.gz>
Bit-identical8/8
rotparent/rotparent-slice-timesanother dataset's motion or grid (-rotparent, -gridparent)-verbose -rotparent -gridparent on slice-timed data
command lineOMP_NUM_THREADS=1 3dvolreg -verbose -rotparent <generated: rp_st.nii.gz> -gridparent <generated: gp27m.nii.gz> -prefix o.nii <generated: s6_st.nii.gz>
Bit-identical8/8
rotparent/rotparent-tshiftanother dataset's motion or grid (-rotparent, -gridparent)-tshift 0 -rotparent
command lineOMP_NUM_THREADS=1 3dvolreg -tshift 0 -rotparent <generated: rp_st.nii.gz> -prefix o.nii <generated: s6_st.nii.gz>
Bit-identical8/8
rotparent/gridparent-gp22another dataset's motion or grid (-rotparent, -gridparent)-rotparent -gridparent with too few slices
command lineOMP_NUM_THREADS=1 3dvolreg -rotparent <generated: rp.nii.gz> -gridparent <generated: gp22.nii.gz> -prefix o.nii <generated: s1.nii.gz>
Rejected by both

AFNI refuses it (exit 1), and so does the replica in reproduce, with AFNI’s messages; its default, fix, accepts it and names the defect (the record’s list of what fix changed).

5/5
rotparent/gridparent-gpxanother dataset's motion or grid (-rotparent, -gridparent)-rotparent -gridparent with another x size
command lineOMP_NUM_THREADS=1 3dvolreg -rotparent <generated: rp.nii.gz> -gridparent <generated: gpx.nii.gz> -prefix o.nii <generated: s1.nii.gz>
Rejected by both

AFNI refuses it (exit 1), and so does the replica in reproduce, with AFNI’s messages; its default, fix, accepts it and names the defect (the record’s list of what fix changed).

5/5
rotparent/gridparent-gpspanother dataset's motion or grid (-rotparent, -gridparent)-rotparent -gridparent with another spacing
command lineOMP_NUM_THREADS=1 3dvolreg -rotparent <generated: rp.nii.gz> -gridparent <generated: gpsp.nii.gz> -prefix o.nii <generated: s1.nii.gz>
Rejected by both

AFNI refuses it (exit 1), and so does the replica in reproduce, with AFNI’s messages; its default, fix, accepts it and names the defect (the record’s list of what fix changed).

5/5
rotparent/gridparent-aloneanother dataset's motion or grid (-rotparent, -gridparent)-gridparent without -rotparent
command lineOMP_NUM_THREADS=1 3dvolreg -gridparent <generated: gp27.nii.gz> -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
rotparent/rotparent-twiceanother dataset's motion or grid (-rotparent, -gridparent)-rotparent twice
command lineOMP_NUM_THREADS=1 3dvolreg -rotparent <generated: rp.nii.gz> -rotparent <generated: rp.nii.gz> -prefix o.nii <generated: s1.nii.gz>
Rejected by both

AFNI refuses it (exit 1), and so does the replica in reproduce, with AFNI’s messages; its default, fix, accepts it and names the defect (the record’s list of what fix changed).

5/5
rotparent/gridparent-twiceanother dataset's motion or grid (-rotparent, -gridparent)-gridparent twice
command lineOMP_NUM_THREADS=1 3dvolreg -gridparent <generated: gp27.nii.gz> -gridparent <generated: gp27.nii.gz> -prefix o.nii <generated: s1.nii.gz>
Rejected by both

AFNI refuses it (exit 1), and so does the replica in reproduce, with AFNI’s messages; its default, fix, accepts it and names the defect (the record’s list of what fix changed).

5/5
rotparent/rotparent-missinganother dataset's motion or grid (-rotparent, -gridparent)-rotparent of a missing file
command lineOMP_NUM_THREADS=1 3dvolreg -rotparent nothere.nii.gz -prefix o.nii <generated: s1.nii.gz>
Rejected by both

AFNI refuses it (exit 1), and so does the replica in reproduce, with AFNI’s messages; its default, fix, accepts it and names the defect (the record’s list of what fix changed).

5/5
rotparent/rotparent-twopassanother dataset's motion or grid (-rotparent, -gridparent)-rotparent -twopass
command lineOMP_NUM_THREADS=1 3dvolreg -rotparent <generated: rp.nii.gz> -twopass -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
rotparent/rotparent-obliqueanother dataset's motion or grid (-rotparent, -gridparent)-rotparent on an oblique series
command lineOMP_NUM_THREADS=1 3dvolreg -rotparent <generated: rp_obl.nii.gz> -prefix o.nii -1Dmatrix_save m.aff12.1D <generated: s3_oblique.nii.gz>
Bit-identical10/10
rotparent/rotparent-twodup-gridanother dataset's motion or grid (-rotparent, -gridparent)-rotparent -twopass -twodup -base <shifted> -gridparent
command lineOMP_NUM_THREADS=1 3dvolreg -rotparent <generated: rp.nii.gz> -twopass -twodup -base <generated: base_shift.nii.gz> -gridparent <generated: gp27.nii.gz> -prefix o.nii <generated: s1.nii.gz>
Bit-identical8/8
rotparent/rotparent-twodupanother dataset's motion or grid (-rotparent, -gridparent)-rotparent -twopass -twodup -base <shifted>
command lineOMP_NUM_THREADS=1 3dvolreg -rotparent <generated: rp.nii.gz> -twopass -twodup -base <generated: base_shift.nii.gz> -prefix o.nii <generated: s1.nii.gz>
Bit-identical8/8
rotparent/rotparent-savedispanother dataset's motion or grid (-rotparent, -gridparent)-rotparent -savedisp -gridparent
command lineOMP_NUM_THREADS=1 3dvolreg -rotparent <generated: rp.nii.gz> -savedisp sd.nii.gz -gridparent <generated: gp27.nii.gz> -prefix o.nii <generated: s1.nii.gz>
Bit-identical14/14
rotparent/rotparent-int16another dataset's motion or grid (-rotparent, -gridparent)-rotparent on int16 (LPI)
command lineOMP_NUM_THREADS=1 3dvolreg -rotparent <generated: rp.nii.gz> -prefix o.nii <generated: s2_short.nii.gz>
Bit-identical8/8
rotparent/rotparent-without-attributesanother dataset's motion or grid (-rotparent, -gridparent)-rotparent of a dataset without VOLREG attributes: AFNI's error
command lineOMP_NUM_THREADS=1 3dvolreg -rotparent <generated: s1.nii.gz> -prefix o.nii <generated: s1.nii.gz>
Rejected by both

AFNI exit 1: ** -rotparent dataset doesn’t have VOLREG attributes!?; the replica exit 1: ** -rotparent dataset doesn’t have VOLREG attributes!? (the same messages)

5/5
twodup/twodup-twodup-twopass -twodup -base <shifted>
command lineOMP_NUM_THREADS=1 3dvolreg -twopass -twodup -base <generated: base_shift.nii.gz> -prefix o.nii -1Dmatrix_save m.aff12.1D <generated: s1.nii.gz>
Bit-identical11/11
twodup/twodup-onepass-twodup-twodup without -twopass
command lineOMP_NUM_THREADS=1 3dvolreg -twodup -base <generated: base_shift.nii.gz> -prefix o.nii <generated: s1.nii.gz>
Bit-identical8/8
twodup/twodup-internal-twodup-twopass -twodup, an internal base
command lineOMP_NUM_THREADS=1 3dvolreg -twopass -twodup -prefix o.nii -1Dfile p.1D <generated: s1.nii.gz>
Bit-identical10/10
twodup/twodup-savedisp-twodup-twopass -twodup -base <shifted> -savedisp on slice-timed data
command lineOMP_NUM_THREADS=1 3dvolreg -twopass -twodup -base <generated: base_shift.nii.gz> -savedisp sd.nii.gz -prefix o.nii <generated: s6_st.nii.gz>
Bit-identical14/14
messages/verbosemessages-verbose
command lineOMP_NUM_THREADS=1 3dvolreg -verbose -prefix o.nii <generated: s1.nii.gz>
Bit-identical8/8
messages/verbose-verbosemessages-verbose -verbose -twopass -Fourier
command lineOMP_NUM_THREADS=1 3dvolreg -verbose -verbose -twopass -Fourier -prefix o.nii <generated: s1.nii.gz>
Bit-identical8/8
messages/base-before-verbosemessages-base <dset> -verbose: messages in the scan's order
command lineOMP_NUM_THREADS=1 3dvolreg -base <generated: base3d.nii.gz> -verbose -prefix o.nii <generated: s1.nii.gz>
Bit-identical8/8
messages/verbose-base-mismatchmessages-verbose -base <another grid>
command lineOMP_NUM_THREADS=1 3dvolreg -verbose -base <generated: s2_short.nii.gz> -prefix o.nii <generated: s1.nii.gz>
Rejected by both

AFNI refuses it (exit 1), and so does the replica in reproduce, with AFNI’s messages; its default, fix, accepts it and names the defect (the record’s list of what fix changed).

5/5
messages/verbose-weightmessages-verbose -weight -base 2
command lineOMP_NUM_THREADS=1 3dvolreg -verbose -weight {1}[1] -base 2 -prefix o.nii <generated: s1.nii.gz>
Bit-identical8/8
messages/verbose-namesmessages-verbose with a base and a weight: their names
command lineOMP_NUM_THREADS=1 3dvolreg -verbose -base <generated: base3d.nii.gz> -weight {2}[0] -prefix o.nii <generated: s1.nii.gz>
Bit-identical8/8
messages/verbose-bytemessages-verbose on bytes
command lineOMP_NUM_THREADS=1 3dvolreg -verbose -prefix o.nii -1Dfile p.1D <generated: s5_byte.nii.gz>
Bit-identical

in fix, its default, the replica corrects AFNI-87 and names it: 81,911 values change (the record’s list of what fix changed)

10/10
messages/verbose-tshift-basemessages-verbose -tshift 0 -base <a 4D weight>
command lineOMP_NUM_THREADS=1 3dvolreg -verbose -tshift 0 -base <generated: weight.nii.gz> -prefix o.nii <generated: s6_st.nii.gz>
Bit-identical8/8
messages/verbose-3d-inputmessages-verbose, a 3D input and a 4D base
command lineOMP_NUM_THREADS=1 3dvolreg -verbose -base <generated: s1.nii.gz> -prefix o.nii <generated: base3d.nii.gz>
Bit-identical8/8
messages/weight-index-beyondmessages-weight 'wset[5]': beyond its volumes
command lineOMP_NUM_THREADS=1 3dvolreg -weight {1}[5] -prefix o.nii <generated: s1.nii.gz>
Rejected by both

AFNI refuses it (exit 1), and so does the replica in reproduce, with AFNI’s messages; its default, fix, accepts it and names the defect (the record’s list of what fix changed).

5/5
datatypes/s2_shortdata typesint16, LPI
command lineOMP_NUM_THREADS=1 3dvolreg -prefix o.nii -1Dfile p.1D -dfile d.txt <generated: s2_short.nii.gz>
Bit-identical12/12
datatypes/s2_scaleddata typesint16 with a slope: written as float
command lineOMP_NUM_THREADS=1 3dvolreg -prefix o.nii -1Dfile p.1D -dfile d.txt <generated: s2_scaled.nii.gz>
Bit-identical12/12
datatypes/s3_obliquedata typesoblique, odd sizes
command lineOMP_NUM_THREADS=1 3dvolreg -Fourier -twopass -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D -prefix o.nii -1Dfile p.1D -dfile d.txt <generated: s3_oblique.nii.gz>
Bit-identical

in fix, its default, the replica corrects AFNI-99 and names it: 10 values change (the record’s list of what fix changed)

18/18
datatypes/s4_sagittaldata typessagittal
command lineOMP_NUM_THREADS=1 3dvolreg -Fourier -twopass -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D -prefix o.nii -1Dfile p.1D -dfile d.txt <generated: s4_sagittal.nii.gz>
Bit-identical

in fix, its default, the replica corrects AFNI-99 and names it: 10 values change (the record’s list of what fix changed)

18/18
datatypes/e_nifti2data typesNIfTI-2
command lineOMP_NUM_THREADS=1 3dvolreg -prefix o.nii -1Dfile p.1D -dfile d.txt <generated: e_nifti2.nii>
Bit-identical12/12
datatypes/e_plaindata typesuncompressed NIfTI-1
command lineOMP_NUM_THREADS=1 3dvolreg -prefix o.nii -1Dfile p.1D -dfile d.txt <generated: e_plain.nii>
Bit-identical12/12
datatypes/e_qformdata typesa qform only
command lineOMP_NUM_THREADS=1 3dvolreg -prefix o.nii -1Dfile p.1D -dfile d.txt <generated: e_qform.nii.gz>
Bit-identical12/12
datatypes/e_bigendiandata typesint16, big-endian
command lineOMP_NUM_THREADS=1 3dvolreg -prefix o.nii -1Dfile p.1D -dfile d.txt <generated: e_bigendian.nii>
Bit-identical12/12
datatypes/bytedata typesbytes (AFNI-87)
command lineOMP_NUM_THREADS=1 3dvolreg -prefix o.nii -1Dfile p.1D <generated: s5_byte.nii.gz>
Bit-identical

in fix, its default, the replica corrects AFNI-87 and names it: 81,911 values change (the record’s list of what fix changed)

10/10
datatypes/input-optiondata types-input, and a dataset name left over
command lineOMP_NUM_THREADS=1 3dvolreg -input <generated: s1.nii.gz> -prefix o.nii extra <generated: s1.nii.gz>
Bit-identical8/8
edge cases/one-volumeedge casesa 4D file of one volume
command lineOMP_NUM_THREADS=1 3dvolreg -prefix o.nii -1Dfile p.1D -dfile d.txt <generated: e_one.nii.gz>
Rejected by both

AFNI refuses it (exit 1), and so does the replica in reproduce, with AFNI’s messages; its default, fix, accepts it and names the defect (the record’s list of what fix changed).

5/5
edge cases/one-volume-twopassedge casesone volume, -twopass
command lineOMP_NUM_THREADS=1 3dvolreg -twopass -prefix o.nii -1Dfile p.1D <generated: e_one.nii.gz>
Rejected by both

AFNI refuses it (exit 1), and so does the replica in reproduce, with AFNI’s messages; its default, fix, accepts it and names the defect (the record’s list of what fix changed).

5/5
edge cases/no-motionedge casesfour identical volumes
command lineOMP_NUM_THREADS=1 3dvolreg -prefix o.nii -1Dfile p.1D -maxdisp1D md.1D <generated: e_still.nii.gz>
Bit-identical14/14
edge cases/no-motion-fmriprepedge casesfour identical volumes, fMRIPrep's call
command lineOMP_NUM_THREADS=1 3dvolreg -Fourier -twopass -zpad 4 -prefix o.nii -1Dfile p.1D -1Dmatrix_save m.aff12.1D <generated: e_still.nii.gz>
Bit-identical12/12
edge cases/large_zedge casesrotations of 12 degrees per volume about z, -twopass -Fourier
command lineOMP_NUM_THREADS=1 3dvolreg -twopass -Fourier -prefix o.nii -1Dfile p.1D -1Dmatrix_save m.aff12.1D <generated: e_large_z.nii.gz>
Bit-identical12/12
edge cases/large_z-onepassedge casesthe same, one pass
command lineOMP_NUM_THREADS=1 3dvolreg -prefix o.nii -1Dfile p.1D <generated: e_large_z.nii.gz>
Bit-identical10/10
edge cases/large_z-nocoarserotedge casesthe same, -twopass -nocoarserot (AFNI-86)
command lineOMP_NUM_THREADS=1 3dvolreg -twopass -nocoarserot -prefix o.nii -1Dfile p.1D <generated: e_large_z.nii.gz>
Bit-identical

in fix, its default, the replica corrects AFNI-86 and names it: 121,396 values change (the record’s list of what fix changed)

10/10
edge cases/large_xedge casesrotations of 9 degrees per volume about x, -twopass -Fourier
command lineOMP_NUM_THREADS=1 3dvolreg -twopass -Fourier -prefix o.nii -1Dfile p.1D -1Dmatrix_save m.aff12.1D <generated: e_large_x.nii.gz>
Bit-identical12/12
edge cases/large_x-onepassedge casesthe same, one pass
command lineOMP_NUM_THREADS=1 3dvolreg -prefix o.nii -1Dfile p.1D <generated: e_large_x.nii.gz>
Bit-identical10/10
edge cases/large_x-nocoarserotedge casesthe same, -twopass -nocoarserot (AFNI-86)
command lineOMP_NUM_THREADS=1 3dvolreg -twopass -nocoarserot -prefix o.nii -1Dfile p.1D <generated: e_large_x.nii.gz>
Bit-identical

in fix, its default, the replica corrects AFNI-86 and names it: 120,124 values change (the record’s list of what fix changed)

10/10
edge cases/nanedge casesNaN in volume 2
command lineOMP_NUM_THREADS=1 3dvolreg -prefix o.nii -1Dfile p.1D -dfile d.txt <generated: e_nan.nii.gz>
Bit-identical12/12
edge cases/nan-twopassedge casesNaN in volume 2, fMRIPrep's call
command lineOMP_NUM_THREADS=1 3dvolreg -Fourier -twopass -zpad 4 -prefix o.nii -1Dfile p.1D <generated: e_nan.nii.gz>
Bit-identical10/10
edge cases/nan-baseedge casesNaN in the base (-base 2)
command lineOMP_NUM_THREADS=1 3dvolreg -base 2 -prefix o.nii -1Dfile p.1D <generated: e_nan.nii.gz>
Bit-identical10/10
edge cases/two-slicesedge casestwo slices
command lineOMP_NUM_THREADS=1 3dvolreg -prefix o.nii -1Dfile p.1D <generated: e_nz2.nii.gz>
Bit-identical10/10
edge cases/one-sliceedge casesone slice
command lineOMP_NUM_THREADS=1 3dvolreg -prefix o.nii <generated: s7_nz1.nii.gz>
Rejected by both

AFNI refuses it (exit 1), and so does the replica in reproduce, with AFNI’s messages; its default, fix, accepts it and names the defect (the record’s list of what fix changed).

5/5
edge cases/one-slice-twopassedge casesone slice, -twopass
command lineOMP_NUM_THREADS=1 3dvolreg -twopass -prefix o.nii <generated: s7_nz1.nii.gz>
Rejected by both

AFNI refuses it (exit 1), and so does the replica in reproduce, with AFNI’s messages; its default, fix, accepts it and names the defect (the record’s list of what fix changed).

5/5
edge cases/subsamplededge cases120 x 120 x 120 voxels, -twopass: the coarse search blurred and subsampled
command lineOMP_NUM_THREADS=1 3dvolreg -twopass -prefix o.nii -1Dfile p.1D -1Dmatrix_save m.aff12.1D <generated: e_big120.nii.gz>
Bit-identical12/12
edge cases/subsampled-fmriprepedge cases120 x 120 x 120 voxels, fMRIPrep's call
command lineOMP_NUM_THREADS=1 3dvolreg -Fourier -twopass -zpad 4 -prefix o.nii -1Dfile p.1D <generated: e_big120.nii.gz>
Bit-identical10/10
edge cases/subsampled-coarse-2edge cases120 x 120 x 120 voxels, -twopass -coarse 2 1: the subsampled step 1
command lineOMP_NUM_THREADS=1 3dvolreg -twopass -coarse 2 1 -prefix o.nii -1Dfile p.1D <generated: e_big120.nii.gz>
Bit-identical10/10
edge cases/four-slices-twopassedge casesfour slices, -twopass: an empty coarse region (AFNI-93)
command lineOMP_NUM_THREADS=1 3dvolreg -twopass -prefix o.nii -1Dfile p.1D <generated: s7_nz4.nii.gz>
Bit-identical

in fix, its default, the replica corrects AFNI-93 and names it: 20,448 values change (the record’s list of what fix changed)

10/10
edge cases/four-slices-nocoarserotedge casesfour slices, -twopass -nocoarserot
command lineOMP_NUM_THREADS=1 3dvolreg -twopass -nocoarserot -prefix o.nii -1Dfile p.1D <generated: s7_nz4.nii.gz>
Bit-identical

in fix, its default, the replica corrects AFNI-86, AFNI-93 and names it: 20,450 values change (the record’s list of what fix changed)

10/10
edge cases/constantedge casesa constant series: one pass registers it
command lineOMP_NUM_THREADS=1 3dvolreg -prefix o.nii <generated: s7_const.nii.gz>
Bit-identical8/8
edge cases/constant-twopassedge casesa constant series, -twopass (AFNI-89)
command lineOMP_NUM_THREADS=1 3dvolreg -twopass -prefix o.nii <generated: s7_const.nii.gz>
Expected divergence

AFNI crashes (exit 1, Fatal Signal 11 (SIGSEGV) received); the replica refuses (exit 1: ** FATAL ERROR: the base’s least squares could not be set up (the messages above say why: all-zero weights, or a base whose rotated and shifted copies do not change under the …

4/4
errors/unknown-optionerrors-bogus <input>
command lineOMP_NUM_THREADS=1 3dvolreg -bogus <generated: s1.nii.gz>
Rejected by both

AFNI exit 1: ** Unknown option: -bogus; the replica exit 1: ** Unknown option: -bogus (the same messages)

5/5
errors/final-boguserrors-final bogus <input>
command lineOMP_NUM_THREADS=1 3dvolreg -final bogus <generated: s1.nii.gz>
Rejected by both

AFNI exit 1: ** Illegal mode after -final; the replica exit 1: ** Illegal mode after -final (the same messages)

5/5
errors/twoblur-lowerrors-twoblur 1 <input>
command lineOMP_NUM_THREADS=1 3dvolreg -twoblur 1 <generated: s1.nii.gz>
Rejected by both

AFNI exit 1: ** ERROR: value after -twoblur is < 2.0; the replica exit 1: ** ERROR: value after -twoblur is < 2.0 (the same messages)

5/5
errors/zpad-negativeerrors-zpad -2 <input>
command lineOMP_NUM_THREADS=1 3dvolreg -zpad -2 <generated: s1.nii.gz>
Rejected by both

AFNI exit 1: ** ERROR: Can’t use -zpad -2; the replica exit 1: ** ERROR: Can’t use -zpad -2 (the same messages)

5/5
errors/base-beyonderrors-base 99 <input>
command lineOMP_NUM_THREADS=1 3dvolreg -base 99 <generated: s1.nii.gz>
Rejected by both

AFNI exit 1: ** Dataset /oracles/inputs/afni-volreg/s1.nii.gz doesn’t have base brick index = 99; the replica exit 1: ** Dataset /oracles/inputs/afni-volreg/s1.nii.gz doesn’t have base brick index = …

5/5
errors/base-negativeerrors-base -1 <input>
command lineOMP_NUM_THREADS=1 3dvolreg -base -1 <generated: s1.nii.gz>
Rejected by both

AFNI exit 1: ** FATAL ERROR: Illegal negative number after -base; the replica exit 1: ** FATAL ERROR: Illegal negative number after -base (the same messages)

5/5
errors/edging-negativeerrors-edging -1 <input>
command lineOMP_NUM_THREADS=1 3dvolreg -edging -1 <generated: s1.nii.gz>
Rejected by both

AFNI exit 1: ** Illegal value after -edging; the replica exit 1: ** Illegal value after -edging (the same messages)

5/5
errors/edging-30pcerrors-edging 30% <input>
command lineOMP_NUM_THREADS=1 3dvolreg -edging 30% <generated: s1.nii.gz>
Rejected by both

AFNI exit 1: ** Illegal percentage after -edging; the replica exit 1: ** Illegal percentage after -edging (the same messages)

5/5
errors/weight-no-selectorerrors-weight <weight> <input>
command lineOMP_NUM_THREADS=1 3dvolreg -weight <generated: weight.nii.gz> <generated: s1.nii.gz>
Rejected by both

AFNI exit 1: ** Illegal weight dataset after -weight; the replica exit 1: ** Illegal weight dataset after -weight (the same messages)

5/5
errors/weight-missingerrors-weight nothere.nii.gz[0] -prefix o.nii <input>
command lineOMP_NUM_THREADS=1 3dvolreg -weight nothere.nii.gz[0] -prefix o.nii <generated: s1.nii.gz>
Rejected by both

AFNI exit 1: ** Can’t open weight dataset nothere.nii.gz; the replica exit 1: ** Can’t open weight dataset nothere.nii.gz (the same messages)

5/5
errors/base-missingerrors-base nothere.nii.gz -prefix o.nii <input>
command lineOMP_NUM_THREADS=1 3dvolreg -base nothere.nii.gz -prefix o.nii <generated: s1.nii.gz>
Rejected by both

AFNI exit 1: ** FATAL ERROR: Couldn’t open -base dataset nothere.nii.gz; the replica exit 1: ** FATAL ERROR: Couldn’t open -base dataset nothere.nii.gz (the same messages)

5/5
errors/tshift-negativeerrors-tshift -1 <input>
command lineOMP_NUM_THREADS=1 3dvolreg -tshift -1 <generated: s1.nii.gz>
Rejected by both

AFNI exit 1: ** -tshift parameter is negative!; the replica exit 1: ** -tshift parameter is negative! (the same messages)

5/5
errors/no-dataseterrors-prefix o.nii
command lineOMP_NUM_THREADS=1 3dvolreg -prefix o.nii
Rejected by both

AFNI exit 1: ** FATAL ERROR: Too few arguments!? Last=o.nii; the replica exit 1: ** FATAL ERROR: Too few arguments!? Last=o.nii (the same messages)

5/5
errors/missing-inputerrors-prefix o.nii /nonexistent/x.nii
command lineOMP_NUM_THREADS=1 3dvolreg -prefix o.nii /nonexistent/x.nii
Rejected by both

AFNI exit 1: ** Can’t open dataset /nonexistent/x.nii; the replica exit 1: ** Can’t open dataset /nonexistent/x.nii (the same messages)

5/5
errors/base-twiceerrors-base 0 -base 3 -prefix o.nii <input>
command lineOMP_NUM_THREADS=1 3dvolreg -base 0 -base 3 -prefix o.nii <generated: s1.nii.gz>
Bit-identical8/8
errors/prefix-after-dataseterrors-prefix o.nii <input> -prefix
command lineOMP_NUM_THREADS=1 3dvolreg -prefix o.nii <generated: s1.nii.gz> -prefix
Rejected by both

AFNI exit 1: ** FATAL ERROR: Too many arguments?! Dataset=/oracles/inputs/afni-volreg/s1.nii.gz?; the replica exit 1: ** FATAL ERROR: Too many arguments?! Dataset=/oracles/inputs/afni-volreg/s1.nii.g…

5/5
errors/maxite-lasterrors-maxite
command lineOMP_NUM_THREADS=1 3dvolreg -maxite
Expected divergence

AFNI crashes (exit 1, Fatal Signal 11 (SIGSEGV) received); the replica refuses (exit 1: ** FATAL ERROR: need argument after -maxite (AFNI-88: AFNI crashes here)), after the same messages

4/4
errors/input-lasterrors-input
command lineOMP_NUM_THREADS=1 3dvolreg -input
Rejected by both

AFNI exit 1: ** FATAL ERROR: Can’t open input dataset ‘(null)’; the replica exit 1: ** FATAL ERROR: Can’t open input dataset ‘(null)’ (the same messages)

5/5
errors/rotparent-lasterrors-rotparent
command lineOMP_NUM_THREADS=1 3dvolreg -rotparent
Rejected by both

AFNI exit 1: ; the replica exit 1: (the same messages)

5/5
errors/gridparent-lasterrors-gridparent
command lineOMP_NUM_THREADS=1 3dvolreg -gridparent
Rejected by both

AFNI exit 1: ; the replica exit 1: (the same messages)

5/5
errors/matrix-save-lasterrors-1Dmatrix_save
command lineOMP_NUM_THREADS=1 3dvolreg -1Dmatrix_save
Rejected by both

AFNI exit 1: ** FATAL ERROR: no argument after ‘-1Dmatrix_save’!; the replica exit 1: ** FATAL ERROR: no argument after ‘-1Dmatrix_save’! (the same messages)

5/5
errors/one-volume-no-baseerrors-base 1 -prefix o.nii <one volume>
command lineOMP_NUM_THREADS=1 3dvolreg -base 1 -prefix o.nii <generated: e_one.nii.gz>
Rejected by both

AFNI exit 1: ** Dataset /oracles/inputs/afni-volreg/e_one.nii.gz doesn’t have base brick index = 1; the replica exit 1: ** Dataset /oracles/inputs/afni-volreg/e_one.nii.gz doesn’t have base brick ind…

5/5
errors/base-beyond-inputerrors-base 99 -prefix o.nii -input <input>
command lineOMP_NUM_THREADS=1 3dvolreg -base 99 -prefix o.nii -input <generated: s1.nii.gz>
Rejected by both

AFNI exit 1: ** Dataset (null) doesn’t have base brick index = 99; the replica exit 1: ** Dataset (null) doesn’t have base brick index = 99 (the same messages)

5/5
errors/base-beyond-input-strayerrors-base 99 -prefix o.nii -input <input> stray
command lineOMP_NUM_THREADS=1 3dvolreg -base 99 -prefix o.nii -input <generated: s1.nii.gz> stray
Rejected by both

AFNI exit 1: ** Dataset stray doesn’t have base brick index = 99; the replica exit 1: ** Dataset stray doesn’t have base brick index = 99 (the same messages)

5/5
divergences/default-prefixexpected divergencesno -prefix: AFNI writes its own format (volreg+orig)
command lineOMP_NUM_THREADS=1 3dvolreg <generated: s1.nii.gz>
Expected divergence

AFNI’s default prefix ‘volreg’ is its own dataset format, which the replica does not write: it refuses (NIfTI only).

2/3
divergences/savedisp-afni-formatexpected divergences-savedisp sd: AFNI's own format
command lineOMP_NUM_THREADS=1 3dvolreg -savedisp sd -prefix o.nii <generated: s1.nii.gz>
Expected divergence

-savedisp without .nii/.nii.gz writes AFNI’s own dataset format, which the replica does not write: it refuses (NIfTI only).

2/3
global/existsglobal optionsan existing output (AFNI-11)
command lineOMP_NUM_THREADS=1 3dvolreg -prefix o.nii <generated: s1.nii.gz> (with o.nii present)
Bit-identical

Neither writes an output; the exit status (0) and the messages are AFNI’s. The default, fix, refuses instead (exit 1) and names AFNI-11 (the record’s list of what fix changed).

6/6
global/overwriteglobal options-overwrite
command lineOMP_NUM_THREADS=1 3dvolreg -overwrite -prefix o.nii <generated: s1.nii.gz> (with o.nii present)
Bit-identical8/8
global/deconflict-yesglobal optionsAFNI_DECONFLICT=YES: renamed
command lineOMP_NUM_THREADS=1 AFNI_DECONFLICT=YES 3dvolreg -prefix o.nii <generated: s1.nii.gz> (with o.nii present)
Bit-identical8/8
global/D-write-type-2global options-DAFNI_NIFTI_WRITE_TYPE=2
command lineOMP_NUM_THREADS=1 3dvolreg -DAFNI_NIFTI_WRITE_TYPE=2 -prefix o.nii <generated: s1.nii.gz>
Bit-identical8/8
environment/quiet-startupenvironment variablesAFNI_QUIET_STARTUP=YES
command lineOMP_NUM_THREADS=1 AFNI_QUIET_STARTUP=YES 3dvolreg -prefix o.nii <generated: s1.nii.gz>
Bit-identical8/8
global/prefix-with-blankglobal optionsa -prefix with a blank: AFNI writes nothing, with exit status 0 (AFNI-12)
command lineOMP_NUM_THREADS=1 3dvolreg -prefix 'a b.nii' <generated: s1.nii.gz>
Bit-identical

Neither writes an output; the exit status (0) and the messages are AFNI’s. The default, fix, refuses instead (exit 1) and names AFNI-12 (the record’s list of what fix changed).

6/6
global/savedisp-with-blankglobal options-savedisp with a blank (AFNI-12)
command lineOMP_NUM_THREADS=1 3dvolreg -savedisp 's d.nii' -prefix o.nii <generated: s1.nii.gz>
Rejected by both

AFNI refuses it (exit 1), and so does the replica in reproduce, with AFNI’s messages; its default, fix, accepts it and names AFNI-12 (the record’s list of what fix changed).

5/5
global/input-dir-with-blankglobal optionsan input in a directory with a blank (AFNI-18)
command lineOMP_NUM_THREADS=1 3dvolreg -prefix o.nii <generated: s1.nii.gz, in a directory with a blank>
Rejected by both

AFNI refuses it (exit 1), and so does the replica in reproduce, with AFNI’s messages; its default, fix, accepts it and names AFNI-18 (the record’s list of what fix changed).

6/6
global/missing-directoryglobal optionsan output in a missing directory (AFNI-26)
command lineOMP_NUM_THREADS=1 3dvolreg -prefix nodir/o.nii <generated: s1.nii.gz>
Bit-identical

Neither writes an output; the exit status (0) and the messages are AFNI’s. The default, fix, refuses instead (exit 1) and names AFNI-26 (the record’s list of what fix changed).

6/6
defects/prefix-255defects of AFNIa -prefix of 255 bytes
command lineOMP_NUM_THREADS=1 3dvolreg -prefix DDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDD/LLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLL.nii <generated: s1.nii.gz> (with DDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDD/.keep present)
Bit-identical8/8
defects/prefix-256defects of AFNIa -prefix of 256 bytes: AFNI aborts (AFNI-90)
command lineOMP_NUM_THREADS=1 3dvolreg -prefix DDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDD/LLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLL.nii <generated: s1.nii.gz> (with DDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDDD/.keep present)
Expected divergence

AFNI crashes (exit 1, Fatal Signal 6 (SIGABRT) received); the replica refuses (exit 1: ** FATAL ERROR: the -prefix name has 256 bytes; AFNI’s buffer holds 255 and AFNI aborts here (*** buffer overflow detected ***, AFNI-90)), after the same messages

5/5
fmriprep/ds000005-run-01 worked examplefMRIPrep's callsfMRIPrep's RobustAverage call (single-band, int16, 64 x 64 x 34, 20 volumes)
command lineOMP_NUM_THREADS=1 3dvolreg -Fourier -twopass -zpad 4 -prefix o.nii.gz -1Dfile o.1D -1Dmatrix_save o.aff12.1D -maxdisp1D o_md.1D <fMRIPrep ds000005-run-01: RobustAverage's sliced volumes>
Bit-identical

fMRIPrep 25.2.6’s AFNI give the bytes of the repository’s build; average.nii.gz: fMRIPrep’s own output, the same bytes; in fix, its default, the replica corrects AFNI-99 and names it: 35 values change: the maximum displacements of -maxdisp1D, as every fMRIPrep input of the record is stored with its second axis posterior to anterior; the registered volumes, and so fMRIPrep’s reference, are the same in fix and in reproduce (the record’s list of what fix changed)

21/21
fmriprep/ds000005-run-03fMRIPrep's callsfMRIPrep's RobustAverage call (two volumes)
command lineOMP_NUM_THREADS=1 3dvolreg -Fourier -twopass -zpad 4 -prefix o.nii.gz -1Dfile o.1D -1Dmatrix_save o.aff12.1D -maxdisp1D o_md.1D <fMRIPrep ds000005-run-03: RobustAverage's sliced volumes>
Bit-identical

fMRIPrep 25.2.6’s AFNI give the bytes of the repository’s build; average.nii.gz: fMRIPrep’s own output, the same bytes; in fix, its default, the replica corrects AFNI-99 and names it: 2 values change: the maximum displacements of -maxdisp1D, as every fMRIPrep input of the record is stored with its second axis posterior to anterior; the registered volumes, and so fMRIPrep’s reference, are the same in fix and in reproduce (the record’s list of what fix changed)

20/20
fmriprep/ds005454fMRIPrep's callsfMRIPrep's RobustAverage call (multiband, 160 x 160 x 96 (the subsampled coarse search), oblique)
command lineOMP_NUM_THREADS=1 3dvolreg -Fourier -twopass -zpad 4 -prefix o.nii.gz -1Dfile o.1D -1Dmatrix_save o.aff12.1D -maxdisp1D o_md.1D <fMRIPrep ds005454: RobustAverage's sliced volumes>
Bit-identical

fMRIPrep 25.2.6’s AFNI give the bytes of the repository’s build; average.nii.gz: fMRIPrep’s own output, the same bytes; in fix, its default, the replica corrects AFNI-99 and names it: 32 values change: the maximum displacements of -maxdisp1D, as every fMRIPrep input of the record is stored with its second axis posterior to anterior; the registered volumes, and so fMRIPrep’s reference, are the same in fix and in reproduce (the record’s list of what fix changed)

20/20
fmriprep/ds003592-echo-1fMRIPrep's callsfMRIPrep's RobustAverage call (multi-echo (echo 1), oblique, 64 x 64 x 43)
command lineOMP_NUM_THREADS=1 3dvolreg -Fourier -twopass -zpad 4 -prefix o.nii.gz -1Dfile o.1D -1Dmatrix_save o.aff12.1D -maxdisp1D o_md.1D <fMRIPrep ds003592-echo-1: RobustAverage's sliced volumes>
Bit-identical

fMRIPrep 25.2.6’s AFNI give the bytes of the repository’s build; average.nii.gz: fMRIPrep’s own output, the same bytes; in fix, its default, the replica corrects AFNI-99 and names it: 35 values change: the maximum displacements of -maxdisp1D, as every fMRIPrep input of the record is stored with its second axis posterior to anterior; the registered volumes, and so fMRIPrep’s reference, are the same in fix and in reproduce (the record’s list of what fix changed)

20/20
fmriprep/ds004958fMRIPrep's callsfMRIPrep's RobustAverage call (multiband 8, uint16 stored, 104 x 104 x 72, 3 volumes, 24 degrees oblique)
command lineOMP_NUM_THREADS=1 3dvolreg -Fourier -twopass -zpad 4 -prefix o.nii.gz -1Dfile o.1D -1Dmatrix_save o.aff12.1D -maxdisp1D o_md.1D <fMRIPrep ds004958: RobustAverage's sliced volumes>
Bit-identical

fMRIPrep 25.2.6’s AFNI give the bytes of the repository’s build; average.nii.gz: fMRIPrep’s own output, the same bytes; in fix, its default, the replica corrects AFNI-99 and names it: 4 values change: the maximum displacements of -maxdisp1D, as every fMRIPrep input of the record is stored with its second axis posterior to anterior; the registered volumes, and so fMRIPrep’s reference, are the same in fix and in reproduce (the record’s list of what fix changed)

20/20
fmriprep/ds002603-mb4fMRIPrep's callsfMRIPrep's RobustAverage call (Philips multiband 4, float32 stored)
command lineOMP_NUM_THREADS=1 3dvolreg -Fourier -twopass -zpad 4 -prefix o.nii.gz -1Dfile o.1D -1Dmatrix_save o.aff12.1D -maxdisp1D o_md.1D <fMRIPrep ds002603-mb4: RobustAverage's sliced volumes>
Bit-identical

fMRIPrep 25.2.6’s AFNI give the bytes of the repository’s build; average.nii.gz: fMRIPrep’s own output, the same bytes; in fix, its default, the replica corrects AFNI-99 and names it: 22 values change: the maximum displacements of -maxdisp1D, as every fMRIPrep input of the record is stored with its second axis posterior to anterior; the registered volumes, and so fMRIPrep’s reference, are the same in fix and in reproduce (the record’s list of what fix changed)

20/20
fmriprep/pepolar-ds005040fMRIPrep's callsfMRIPrep's RobustAverage call (sdcflows' PEPOLAR reference average: 2 topup-corrected volumes, float32)
command lineOMP_NUM_THREADS=1 3dvolreg -Fourier -twopass -zpad 4 -prefix o.nii.gz -1Dfile o.1D -1Dmatrix_save o.aff12.1D -maxdisp1D o_md.1D <fMRIPrep pepolar-ds005040: RobustAverage's sliced volumes>
Bit-identical

fMRIPrep 25.2.6’s AFNI give the bytes of the repository’s build; average.nii.gz: fMRIPrep’s own output, the same bytes; in fix, its default, the replica corrects AFNI-99 and names it: 2 values change: the maximum displacements of -maxdisp1D, as every fMRIPrep input of the record is stored with its second axis posterior to anterior; the registered volumes, and so fMRIPrep’s reference, are the same in fix and in reproduce (the record’s list of what fix changed)

20/20
fmriprep/pepolar-ds001600fMRIPrep's callsfMRIPrep's RobustAverage call (sdcflows' PEPOLAR reference average: 2 volumes, 80 x 80 x 60)
command lineOMP_NUM_THREADS=1 3dvolreg -Fourier -twopass -zpad 4 -prefix o.nii.gz -1Dfile o.1D -1Dmatrix_save o.aff12.1D -maxdisp1D o_md.1D <fMRIPrep pepolar-ds001600: RobustAverage's sliced volumes>
Bit-identical

fMRIPrep 25.2.6’s AFNI give the bytes of the repository’s build; average.nii.gz: fMRIPrep’s own output, the same bytes; in fix, its default, the replica corrects AFNI-99 and names it: 2 values change: the maximum displacements of -maxdisp1D, as every fMRIPrep input of the record is stored with its second axis posterior to anterior; the registered volumes, and so fMRIPrep’s reference, are the same in fix and in reproduce (the record’s list of what fix changed)

20/20
defects/coarse-1-largedefects of AFNIBit-identical4/4
difftest/afni99-notice-unchangedrandom testingBit-identical12/12
difftest/constant-twopass-lsqfitrandom testingExpected divergence

AFNI crashes (exit 1, Fatal Signal 11 (SIGSEGV) received); the replica refuses (exit 1: ** FATAL ERROR: the base’s least squares could not be set up (the messages above say why: all-zero weights, or a base whose rotated and shifted copies do not change under the …

4/4
difftest/crash-after-progress-linerandom testingExpected divergence

AFNI crashes (exit 1, Fatal Signal 11 (SIGSEGV) received); the replica refuses (exit 1: ** FATAL ERROR: a Fourier rotation shifted the last of an odd count of rows beyond the row length together with the previous pair’s second shift (a rotation far beyond a head’…

4/4
difftest/fourier-lone-row-defaultrandom testingExpected divergence

AFNI crashes (exit 1, Fatal Signal 11 (SIGSEGV) received); the replica refuses (exit 1: ** FATAL ERROR: a Fourier rotation shifted the last of an odd count of rows beyond the row length together with the previous pair’s second shift (a rotation far beyond a head’…

5/5
difftest/fourier-lone-row-nullrandom testingExpected divergence

AFNI crashes (exit 1, Fatal Signal 11 (SIGSEGV) received); the replica refuses (exit 1: ** FATAL ERROR: a Fourier rotation shifted the last of an odd count of rows beyond the row length together with the previous pair’s second shift (a rotation far beyond a head’…

5/5
difftest/notice-after-open-linerandom testingBit-identical

in fix, its default, the replica corrects AFNI-92, AFNI-93 and names it: 1,436 values change (the record’s list of what fix changed)

10/10
difftest/uint16-base-and-inputrandom testingRejected by both

AFNI refuses it (exit 1), and so does the replica in reproduce, with AFNI’s messages; its default, fix, accepts it and names the defect (the record’s list of what fix changed).

5/5

The clean room’s record:

251 cases

CaseCategoryWhat it testsVerdictChecks
original/defaultinterpolation methodsthe default heptic method on the moving series s1
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/fmriprep-s1two passes and the coarse search (-twopass)fMRIPrep's options on s1 (motions of 3 degrees and 5 mm)
command line3dvolreg -Fourier -twopass -zpad 4 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0, volumes 0.014, matrices 0.096, maximum displacements 0; 6/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/int16-rasdata typesint16 stored L2R, P2A, I2S (AFNI-99 in -maxdisp)
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_ras_i16>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0.031, matrices 0.71, maximum displacements 0.37; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

48/48
original/savedispoutputs-savedisp's three datasets
command line3dvolreg -savedisp sd.nii -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

60/60
original/verbose-twopassmessages-verbose with -twopass: the coarse searches' lines
command line3dvolreg -verbose -twopass -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0.052, matrices 0.73, maximum displacements 0.46; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/error-unknownerrorsan unknown option
command line3dvolreg -bogus -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/method-linearinterpolation methodsestimation and output by linear
command line3dvolreg -linear -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.0086, volumes 0.0004, matrices 0.68, maximum displacements 0.37; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/method-cubicinterpolation methodsestimation and output by cubic
command line3dvolreg -cubic -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.46, volumes 0.034, matrices 0.79, maximum displacements 0.47; 3/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/method-quinticinterpolation methodsestimation and output by quintic
command line3dvolreg -quintic -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.27, volumes 0.053, matrices 0.75, maximum displacements 0.46; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/method-hepticinterpolation methodsestimation and output by heptic
command line3dvolreg -heptic -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/method-Fourierinterpolation methodsestimation and output by Fourier
command line3dvolreg -Fourier -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.0087, volumes 0.0057, matrices 0.37, maximum displacements 0.44; 5/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/final-NNinterpolation methodsheptic estimation, -final NN
command line3dvolreg -final NN -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

44/44
original/final-linearinterpolation methodsheptic estimation, -final linear
command line3dvolreg -final linear -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/final-cubicinterpolation methodsheptic estimation, -final cubic
command line3dvolreg -final cubic -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/final-quinticinterpolation methodsheptic estimation, -final quintic
command line3dvolreg -final quintic -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/final-Fourierinterpolation methodsheptic estimation, -final Fourier
command line3dvolreg -final Fourier -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/final-heptic-linearinterpolation methods-linear estimation, -final heptic
command line3dvolreg -linear -final heptic -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.0086, volumes 0.0004, matrices 0.68, maximum displacements 0.37; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/noclip-cubicinterpolation methods-noclip with cubic
command line3dvolreg -cubic -noclip -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.46, volumes 0.034, matrices 0.79, maximum displacements 0.47; 3/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/noclip-fourierinterpolation methods-noclip with Fourier
command line3dvolreg -Fourier -noclip -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.0087, volumes 0.0057, matrices 0.37, maximum displacements 0.44; 5/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/clipitinterpolation methods-clipit (a notice: the default)
command line3dvolreg -clipit -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/zpad-0interpolation methods-zpad 0
command line3dvolreg -zpad 0 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0.071, matrices 0.71, maximum displacements 0.5; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/zpad-2interpolation methods-zpad 2
command line3dvolreg -zpad 2 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/zpad-9interpolation methods-zpad 9
command line3dvolreg -zpad 9 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.047, volumes 0.029, matrices 0.58, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/fourier-zpad0-noclipinterpolation methodsFourier without padding or clipping
command line3dvolreg -Fourier -zpad 0 -noclip -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.018, volumes 0.07, matrices 0.5, maximum displacements 0.35; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/spellingsinterpolation methods-Linear, -Quintic: the methods' other spellings
command line3dvolreg -Linear -Quintic -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.27, volumes 0.053, matrices 0.75, maximum displacements 0.46; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/linear-noclip-final-nninterpolation methods-linear -noclip -final NN
command line3dvolreg -linear -noclip -final NN -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.0086, matrices 0.68, maximum displacements 0.37; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

44/44
original/iso-ras-fourierinterpolation methodsisotropic 2 mm voxels stored RAS, Fourier
command line3dvolreg -Fourier -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: iso>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.61, volumes 0.058, matrices 0.78, maximum displacements 0.58; 1/4 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/base-2the base volume (-base)-base 2
command line3dvolreg -base 2 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.01, volumes 0.0001, matrices 0.42, maximum displacements 0.46; 3/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/base-lastthe base volume (-base)-base 5 (the last volume)
command line3dvolreg -base 5 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.014, volumes 0.0001, matrices 0.51, maximum displacements 0; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/base-dsetthe base volume (-base)an external base (volume 0 of a 3D dataset)
command line3dvolreg -base <generated: base_ext> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.01, volumes 0.0001, matrices 0.42, maximum displacements 0.46; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/base-dset-verbosethe base volume (-base)an external base elsewhere: the centres' distance
command line3dvolreg -verbose -base <generated: base_moved> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.094, volumes 0.0014, matrices 0.6, maximum displacements 0.54; 0/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/base-4dthe base volume (-base)a 4D -base dataset: its warning, volume 0
command line3dvolreg -base <generated: base4d> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.094, volumes 0.0014, matrices 0.6, maximum displacements 0.54; 0/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/base-dset-twopassthe base volume (-base)an external base with -twopass
command line3dvolreg -twopass -base <generated: base_moved> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.015, volumes 0.035, matrices 0.5, maximum displacements 0.5; 0/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/base-abbrevthe base volume (-base)-bas 1 (four letters)
command line3dvolreg -bas 1 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.094, volumes 0.0014, matrices 0.6, maximum displacements 0.54; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/scaleddata typesint16 with scale factors: stored as float32
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_scaled>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0.0019, matrices 0.72, maximum displacements 0.54; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/scaled-floatdata typesint16 with scale factors, -float
command line3dvolreg -float -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_scaled>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0.0019, matrices 0.72, maximum displacements 0.54; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/int16-floatdata typesint16 with -float
command line3dvolreg -float -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_ras_i16>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0.031, matrices 0.71, maximum displacements 0.37; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/uint8data typesuint8 (AFNI-87)
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_u8>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.18, volumes 0.0018, matrices 0.72, maximum displacements 0.46; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

48/48
original/uint8-fourierdata typesuint8 with fMRIPrep's options (AFNI-87)
command line3dvolreg -Fourier -twopass -zpad 4 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_u8>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.0036, volumes 0, matrices 0.1, maximum displacements 0; 3/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

48/48
original/float64data typesfloat64 stored (converted to float32)
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_f64>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/uint16data typesuint16 stored
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_u16>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0.053, matrices 0.73, maximum displacements 0.46; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/bigendiandata typesa big-endian float32 file
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_be>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/nifti2data typesa NIfTI-2 file
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_n2>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/plain-niidata typesan uncompressed .nii, output .nii.gz
command line3dvolreg -prefix out.nii.gz -1Dfile p.1D <generated: s1_plain>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

28/28
original/int16-fmriprepdata typesint16 RAS with fMRIPrep's options
command line3dvolreg -Fourier -twopass -zpad 4 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_ras_i16>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.013, volumes 0.014, matrices 0.32, maximum displacements 0; 3/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

48/48
original/one-volume-baseedge casesone volume with an external base
command line3dvolreg -base <generated: base_ext> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: vol3>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.0085, volumes 0.0001, matrices 0.42, maximum displacements 0.46; 0/1 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/stilledge casesno motion at all
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: still>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0, volumes 0, matrices 0, maximum displacements 0; 4/4 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/big-motionedge casesrotations of 6 to 12 degrees and 8 mm without -twopass
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: big>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0, matrices 0.65, maximum displacements 0.4; 1/4 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/big-motion-twopassedge casesthe same with -twopass: the coarse searches find them
command line3dvolreg -twopass -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: big>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0, matrices 0.59, maximum displacements 0.38; 1/4 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/big-motion-fmriprepedge casesthe same with fMRIPrep's options
command line3dvolreg -Fourier -twopass -zpad 4 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: big>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.003, volumes 0, matrices 0.35, maximum displacements 0; 1/4 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/two-volumesedge casestwo volumes
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: two>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.0085, volumes 0.0001, matrices 0.3, maximum displacements 0; 1/2 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/emptyishedge casesa volume with fewer than 66 non-zero voxels: not registered
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: emptyish>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.0085, volumes 0.0001, matrices 0.3, maximum displacements 0; 3/4 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/emptyish-twopassedge casesthe same with -twopass
command line3dvolreg -twopass -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: emptyish>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.0085, volumes 0.0001, matrices 0.34, maximum displacements 0; 2/4 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/constedge casesa constant series without -twopass: registers
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: const>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0, volumes 0, matrices 0.0011, maximum displacements 0; 1/4 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/const29-twopassedge casesa constant 29^3 series with -twopass: AFNI-89 not reached
command line3dvolreg -twopass -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: const29>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0, matrices 0.65, maximum displacements 0.1; 1/3 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/one-sliceedge casesone slice
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: nz1>
Rejected by both

both refuse the input with the same messages

10/10
original/four-slicesedge casesfour slices
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: nz4>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.31, volumes 0.2, matrices 0.95, maximum displacements 0.61; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/odd-sizesedge cases37 x 41 x 19 voxels of 3.2 x 2.9 x 4 mm
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: odd>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0, volumes 0, matrices 0.0016, maximum displacements 0; 4/4 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/odd-fourieredge casesthe odd grid with Fourier (odd FFT lengths) and -twopass
command line3dvolreg -Fourier -twopass -zpad 4 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: odd>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0, volumes 0, matrices 0.016, maximum displacements 0; 4/4 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/obliqueedge casesan oblique grid (the obliquity report)
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_oblique>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/oblique-fmriprepedge casesthe oblique grid with fMRIPrep's options
command line3dvolreg -Fourier -twopass -zpad 4 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_oblique>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0, volumes 0.014, matrices 0.096, maximum displacements 0; 6/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/sagittaledge casessagittal slices (axes y, z, x)
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_sagittal>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.17, volumes 0.052, matrices 0.7, maximum displacements 0.4; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/sagittal-twopassedge casessagittal slices with -twopass
command line3dvolreg -twopass -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_sagittal>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.019, volumes 0.0002, matrices 0.56, maximum displacements 0.37; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/nanedge casesNaN voxels
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_nan>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0.029, matrices 0.71, maximum displacements 0.45; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/negativeedge casesnegative values (clipping to the input's range)
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_neg>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.36, volumes 0.33, matrices 0.8, maximum displacements 0.66; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/qform-onlyedge casesa qform only
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_qonly>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/sform-onlyedge casesan sform only
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_sonly>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/no-transformedge casesneither code: the NO spatial transform warning
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_nocode>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/code2-verboseedge casescodes 2 (./ in AFNI's names) with -verbose
command line3dvolreg -verbose -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_code2>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/msecedge casesa TR in milliseconds
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_ms>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/afni86defects of AFNI-nocoarserot read as -noclip (AFNI-86)
command line3dvolreg -twopass -nocoarserot -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0.053, matrices 0.73, maximum displacements 0.46; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/afni86-alonedefects of AFNI-nocoarserot without -twopass (AFNI-86)
command line3dvolreg -nocoarserot -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/afni88-prefixdefects of AFNI-prefix without a value (AFNI-88)
command line3dvolreg <generated: s1> -prefix
Rejected by both

both refuse the input with the same messages

10/10
original/afni88-inputdefects of AFNI-input without a value (AFNI-88)
command line3dvolreg -input
Rejected by both

both refuse the input with the same messages

10/10
original/afni88-rotparentdefects of AFNI-rotparent without a value (AFNI-88)
command line3dvolreg -rotparent
Rejected by both

both refuse the input with the same messages

10/10
original/afni88-matrixdefects of AFNI-1Dmatrix_save without a value (AFNI-88)
command line3dvolreg -1Dmatrix_save
Rejected by both

both refuse the input with the same messages

10/10
original/afni88-basedefects of AFNI-base without a value (AFNI-88)
command line3dvolreg -prefix o.nii -base
Rejected by both

AFNI crashes and the clean room stops with AFNI’s messages and an error

11/11
original/afni89-constdefects of AFNIa constant 40 x 44 x 24 series with -twopass (AFNI-89)
command line3dvolreg -twopass -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: const>
Rejected by both

AFNI crashes and the clean room stops with AFNI’s messages and an error

11/11
original/afni89-halfdefects of AFNIa half-constant series with -twopass (AFNI-89)
command line3dvolreg -twopass -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: halfconst>
Rejected by both

AFNI crashes and the clean room stops with AFNI’s messages and an error

11/11
original/afni89-weightdefects of AFNIan all-zero -weight (AFNI-89)
command line3dvolreg -weight <generated: weight_zero>[0] -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

AFNI crashes and the clean room stops with AFNI’s messages and an error

11/11
original/afni89-weight-twopassdefects of AFNIan all-zero -weight with -twopass (AFNI-89)
command line3dvolreg -twopass -weight <generated: weight_zero>[0] -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

AFNI crashes and the clean room stops with AFNI’s messages and an error

11/11
original/afni90-prefixdefects of AFNIa -prefix of 256 bytes (AFNI-90)
command line3dvolreg -prefix pppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppppp.nii -1Dfile p.1D <generated: s1>
Rejected by both

AFNI crashes and the clean room stops with AFNI’s messages and an error

11/11
original/afni90-dfiledefects of AFNIa -dfile of 300 bytes (AFNI-90)
command line3dvolreg -dfile dddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddddd -prefix out.nii <generated: s1>
Rejected by both

AFNI crashes and the clean room stops with AFNI’s messages and an error

17/17
original/afni90-1dfiledefects of AFNIa -1Dfile of 256 bytes (AFNI-90)
command line3dvolreg -1Dfile xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx -prefix out.nii <generated: s1>
Rejected by both

AFNI crashes and the clean room stops with AFNI’s messages and an error

17/17
original/afni90-weightdefects of AFNIa -weight name of 300 bytes (AFNI-90)
command line3dvolreg -weight wwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwwww.nii[0] -prefix out.nii <generated: s1>
Rejected by both

AFNI crashes and the clean room stops with AFNI’s messages and an error

11/11
original/afni90-maxdispdefects of AFNIa -maxdisp1D name of 254 bytes (written)
command line3dvolreg -maxdisp1D mmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmmm.1D -prefix out.nii <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, volumes 0.0004; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

22/22
original/afni92defects of AFNI-nomaxdisp with -twopass: the coarse automask's clip fraction (AFNI-92)
command line3dvolreg -nomaxdisp -twopass -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0.052, matrices 0.73; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

38/38
original/afni93defects of AFNIfour slices with -twopass: an empty search region (AFNI-93)
command line3dvolreg -twopass -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: nz4>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.057, volumes 0.14, matrices 0.76, maximum displacements 0.43; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/afni93-one-slicedefects of AFNIone slice with -twopass (AFNI-93)
command line3dvolreg -twopass -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: nz1>
Rejected by both

both refuse the input with the same messages

10/10
original/afni94defects of AFNI-coarse 1 2 with -twopass on 120^3 voxels (AFNI-94; against the replica: AFNI never ends)
command line3dvolreg -twopass -coarse 1 2 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: e120>
Rejected by both

AFNI never ends (AFNI-94)

10/10
original/afni95defects of AFNI-base 0 then another -base (AFNI-95)
command line3dvolreg -base 0 -base 2 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.01, volumes 0.0001, matrices 0.42, maximum displacements 0.46; 3/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

28/28
original/afni96-centredefects of AFNIa -rotparent without VOLREG_CENTER_OLD (AFNI-96)
command line3dvolreg -rotparent <generated: rp_nocentre> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

AFNI crashes and the clean room stops with AFNI’s messages and an error

11/11
original/afni96-basedefects of AFNIa -rotparent without VOLREG_CENTER_BASE (AFNI-96)
command line3dvolreg -rotparent <generated: rp_nobase> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

AFNI crashes and the clean room stops with AFNI’s messages and an error

11/11
original/afni96-shortdefects of AFNIa -rotparent centre of two numbers (AFNI-96)
command line3dvolreg -rotparent <generated: rp_short> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Expected divergence

AFNI-96 (rule B, an error in fix and in reproduce, spec §12): AFNI reads the third number past the attribute’s two and goes on with whatever it finds there; the clean room stops

10/10
original/afni98-basedefects of AFNI-input with a base index too large: (null) (AFNI-98)
command line3dvolreg -input <generated: s1> -base 9 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D
Rejected by both

both refuse the input with the same messages

10/10
original/afni98-weightdefects of AFNI-input with a -weight of another size (AFNI-98)
command line3dvolreg -input <generated: s1> -weight <generated: weight_small>[0] -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D
Rejected by both

both refuse the input with the same messages

10/10
original/afni99-lpsdefects of AFNIthe same data stored R2L, A2P, I2S: the reference of AFNI-99
command line3dvolreg -verbose -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/afni99-rasdefects of AFNIstored L2R, P2A, I2S with -verbose (AFNI-99)
command line3dvolreg -verbose -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_ras_i16>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0.031, matrices 0.71, maximum displacements 0.37; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

48/48
original/afni99-savedispdefects of AFNI-savedisp on a RAS dataset (AFNI-99)
command line3dvolreg -savedisp sd.nii -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_ras_i16>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0.031, matrices 0.71, maximum displacements 0.37; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

62/62
original/afni01-tshiftdefects of AFNI-tshift on int16 with a negative scale factor (AFNI-01)
command line3dvolreg -tshift 0 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_st_neg>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.026, volumes 0.053, matrices 0.45, maximum displacements 0.43; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/too-fewerrorsno dataset
command line3dvolreg -prefix o.nii
Rejected by both

both refuse the input with the same messages

10/10
original/too-manyerrorstwo datasets
command line3dvolreg -prefix o.nii <generated: s1> <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/final-baderrors-final spline
command line3dvolreg -final spline -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/twoblur-smallerrors-twoblur 1.5
command line3dvolreg -twoblur 1.5 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/zpad-negativeerrors-zpad -1
command line3dvolreg -zpad -1 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/edging-negativeerrors-edging -1
command line3dvolreg -edging -1 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/edging-percenterrors-edging 30%
command line3dvolreg -edging 30% -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/tshift-negativeerrors-tshift -1
command line3dvolreg -tshift -1 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/base-negativeerrors-base -1
command line3dvolreg -base -1 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/base-too-largeerrors-base 9
command line3dvolreg -base 9 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/base-missingerrors-base of a missing file
command line3dvolreg -base nothere.nii -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/base-twiceerrors-base 1 -base 2
command line3dvolreg -base 1 -base 2 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/input-missingerrors-input of a missing file
command line3dvolreg -input nothere.nii -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D
Rejected by both

both refuse the input with the same messages

10/10
original/dataset-missingerrorsa missing dataset
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: nothere.nii>
Rejected by both

both refuse the input with the same messages

10/10
original/weight-missingerrors-weight of a missing file
command line3dvolreg -weight nothere.nii[0] -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/weight-no-bracketerrors-weight without [n]
command line3dvolreg -weight <generated: weight> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/weight-no-indexerrors-weight with [x]
command line3dvolreg -weight <generated: weight>[x] -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/weight-bad-indexerrors-weight with [3]
command line3dvolreg -weight <generated: weight>[3] -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/weight-twiceerrorstwo -weight
command line3dvolreg -weight <generated: weight>[0] -weight <generated: weight>[0] -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/weight-dimserrorsa -weight of another size
command line3dvolreg -weight <generated: weight_small>[0] -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/matrix-twiceerrorstwo -1Dmatrix_save
command line3dvolreg -1Dmatrix_save a -1Dmatrix_save b <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/matrix-badnameerrors-1Dmatrix_save of a refused name
command line3dvolreg -1Dmatrix_save a;b <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/input-twiceerrorstwo -input
command line3dvolreg -input <generated: s1> -input <generated: s1> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D
Rejected by both

both refuse the input with the same messages

10/10
original/input-extraerrors-input with an argument left (a warning)
command line3dvolreg -prefix out.nii -1Dfile p.1D -input <generated: s1> extra
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

28/28
original/one-brick-itselferrorsone volume and an internal base
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: vol3>
Rejected by both

both refuse the input with the same messages

10/10
original/base-dimserrorsan external base of another size
command line3dvolreg -base <generated: base_dims> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/base-spacingerrorsan external base of other steps
command line3dvolreg -base <generated: base_spacing> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/base-orienterrorsan external base of another orientation
command line3dvolreg -base <generated: base_orient> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/base-sparseerrorsa base of fewer than 100 non-zero voxels
command line3dvolreg -base <generated: base_sparse> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/savedisp-badnameerrors-savedisp x.nii (too short)
command line3dvolreg -savedisp x.nii -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

60/60
original/prefix-afni-formaterrorsa -prefix in AFNI's own format (the replica refuses it)
command line3dvolreg -prefix volreg_out <generated: s1>
Expected divergence

The clean room writes no .HEAD/.BRIK dataset: it stops with an error where AFNI writes its own format (all the the clean room AFNI tools’ rule, spec §9.4)

10/10
original/prefix-defaulterrorsno -prefix (AFNI's default 'volreg', its own format)
command line3dvolreg -1Dfile p.1D <generated: s1>
Expected divergence

The clean room writes no .HEAD/.BRIK dataset: the default prefix ‘volreg’ is refused (spec §9.4)

10/10
original/prefix-illegalerrorsa -prefix NIfTI name AFNI refuses when writing
command line3dvolreg -prefix a*b.nii -1Dfile p.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

22/22
original/rotcomoutputs-rotcom: 3drotate's commands
command line3dvolreg -rotcom -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/rotcom-noclip-zpad0outputs-rotcom with -noclip -zpad 0 -final Fourier
command line3dvolreg -rotcom -noclip -zpad 0 -final Fourier -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0.07, matrices 0.71, maximum displacements 0.5; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/prefix-nulloutputs-prefix NULL: parameters only
command line3dvolreg -prefix NULL -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, matrices 0.53; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

26/26
original/maxdisp-stdoutoutputs-maxdisp1D - (standard output, and -_delt)
command line3dvolreg -prefix out.nii -1Dfile p.1D -maxdisp1D - <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

28/28
original/maxdisp-noextoutputs-maxdisp1D md (md_delt)
command line3dvolreg -prefix out.nii -1Dfile p.1D -maxdisp1D md <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

36/36
original/matrix-noextoutputs-1Dmatrix_save ff (ff.aff12.1D)
command line3dvolreg -prefix out.nii -1Dfile p.1D -1Dmatrix_save ff <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

34/34
original/overwrite-textoutputsexisting text outputs: their warnings
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, volumes 0.0004; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

22/22
original/nomaxdispoutputs-nomaxdisp
command line3dvolreg -nomaxdisp -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

38/38
original/SAVEDISPoutputs-SAVEDISP's nineteen datasets
command line3dvolreg -SAVEDISP sd.nii -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

93/93
original/SAVEDISP-stilloutputs-SAVEDISP without motion (zeros left alone)
command line3dvolreg -SAVEDISP sd.nii.gz -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: still>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0, volumes 0, matrices 0, maximum displacements 0; 4/4 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

93/93
original/savedisp-nulloutputs-savedisp with -prefix NULL
command line3dvolreg -savedisp sd.nii.gz -prefix NULL -1Dfile p.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

30/30
original/savedisp-obliqueoutputs-savedisp on the oblique grid
command line3dvolreg -savedisp sd.nii -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_oblique>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

60/60
original/existing-prefixoutputsan existing -prefix (AFNI-11)
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

22/22
original/verbose-outputsoutputs-verbose with every output (the order of writing)
command line3dvolreg -verbose -savedisp sd.nii -rotcom -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

60/60
original/verbosemessages-verbose
command line3dvolreg -verbose -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/verb-abbrevmessages-verb (five letters)
command line3dvolreg -verb -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/vv-hepticmessages-verbose -verbose: the basis and the fits
command line3dvolreg -verbose -verbose -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

50/50
original/vv-linearmessages-verbose -verbose -linear
command line3dvolreg -verbose -verbose -linear -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.0086, volumes 0.0004, matrices 0.68, maximum displacements 0.37; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

50/50
original/vv-fouriermessages-verbose -verbose -Fourier
command line3dvolreg -verbose -verbose -Fourier -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.0087, volumes 0.0057, matrices 0.37, maximum displacements 0.44; 5/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

50/50
original/vv-twopassmessages-verbose -verbose -twopass
command line3dvolreg -verbose -verbose -twopass -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0.052, matrices 0.73, maximum displacements 0.46; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

50/50
original/vv-wtinpmessages-verbose -verbose -wtinp
command line3dvolreg -verbose -verbose -wtinp -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.028, volumes 0.0002, matrices 0.66, maximum displacements 0.44; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

50/50
original/vv-weightmessages-verbose -verbose -weight
command line3dvolreg -verbose -verbose -weight <generated: weight>[0] -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.019, volumes 0.0002, matrices 0.62, maximum displacements 0.46; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

50/50
original/verbose-bytemessages-verbose on a byte input
command line3dvolreg -verbose -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_u8>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.18, volumes 0.0018, matrices 0.72, maximum displacements 0.46; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

48/48
original/verbose-scaledmessages-verbose on a scaled input
command line3dvolreg -verbose -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_scaled>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0.0019, matrices 0.72, maximum displacements 0.54; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/noticesmessages-clipit -coarserot -wtrim: their notices
command line3dvolreg -clipit -coarserot -wtrim -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/zpad-twicemessages-zpad 2 -zpad 3: the second -zpad's warning
command line3dvolreg -zpad 2 -zpad 3 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.047, volumes 0.029, matrices 0.58, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/rpanother dataset's motion or grid (-rotparent, -gridparent)-rotparent (AFNI's own output)
command line3dvolreg -rotparent <generated: rp> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.052, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/rp-verboseanother dataset's motion or grid (-rotparent, -gridparent)-rotparent -verbose
command line3dvolreg -verbose -rotparent <generated: rp> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.052, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/rp-fourieranother dataset's motion or grid (-rotparent, -gridparent)-rotparent -Fourier
command line3dvolreg -Fourier -rotparent <generated: rp> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.0087, volumes 0.014, matrices 0.37, maximum displacements 0.44; 5/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/rp-linear-noclipanother dataset's motion or grid (-rotparent, -gridparent)-rotparent -linear -noclip
command line3dvolreg -linear -noclip -rotparent <generated: rp> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.0086, volumes 0.076, matrices 0.68, maximum displacements 0.37; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/rp-twopassanother dataset's motion or grid (-rotparent, -gridparent)-rotparent -twopass (its warning)
command line3dvolreg -twopass -rotparent <generated: rp> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0.052, matrices 0.73, maximum displacements 0.46; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/rp-gp27another dataset's motion or grid (-rotparent, -gridparent)a -gridparent of 27 slices: padding
command line3dvolreg -verbose -rotparent <generated: rp> -gridparent <generated: gp27> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.036, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/rp-gp29another dataset's motion or grid (-rotparent, -gridparent)a -gridparent of 29 slices: padding 2 and 3
command line3dvolreg -rotparent <generated: rp> -gridparent <generated: gp29> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.034, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/rp-gp24another dataset's motion or grid (-rotparent, -gridparent)a -gridparent of as many slices: its origin
command line3dvolreg -rotparent <generated: rp> -gridparent <generated: gp24> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.039, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/rp-gp27manother dataset's motion or grid (-rotparent, -gridparent)a -gridparent with S2I slices
command line3dvolreg -rotparent <generated: rp> -gridparent <generated: gp27m> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/rp-gp22another dataset's motion or grid (-rotparent, -gridparent)a -gridparent with fewer slices (error)
command line3dvolreg -rotparent <generated: rp> -gridparent <generated: gp22> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/rp-gpspanother dataset's motion or grid (-rotparent, -gridparent)a -gridparent of other steps (error)
command line3dvolreg -rotparent <generated: rp> -gridparent <generated: gpsp> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/rp-gpxanother dataset's motion or grid (-rotparent, -gridparent)a -gridparent of another nx (error)
command line3dvolreg -rotparent <generated: rp> -gridparent <generated: gpx> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/gp-aloneanother dataset's motion or grid (-rotparent, -gridparent)-gridparent without -rotparent (a warning)
command line3dvolreg -gridparent <generated: gp27> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/rp-noattranother dataset's motion or grid (-rotparent, -gridparent)a -rotparent without VOLREG attributes
command line3dvolreg -rotparent <generated: s1> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/rp-missinganother dataset's motion or grid (-rotparent, -gridparent)a missing -rotparent
command line3dvolreg -rotparent nothere.nii -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/rp-twiceanother dataset's motion or grid (-rotparent, -gridparent)two -rotparent
command line3dvolreg -rotparent <generated: rp> -rotparent <generated: rp> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/gp-missinganother dataset's motion or grid (-rotparent, -gridparent)a missing -gridparent
command line3dvolreg -rotparent <generated: rp> -gridparent nothere.nii -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/gp-twiceanother dataset's motion or grid (-rotparent, -gridparent)two -gridparent
command line3dvolreg -rotparent <generated: rp> -gridparent <generated: gp27> -gridparent <generated: gp27> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/rp-skewanother dataset's motion or grid (-rotparent, -gridparent)a -rotparent matrix not orthogonal
command line3dvolreg -rotparent <generated: rp_skew> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/rp-stanother dataset's motion or grid (-rotparent, -gridparent)-rotparent on a series with slice offsets (moved)
command line3dvolreg -verbose -rotparent <generated: rp_st> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_st>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.052, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/rp-st-gp27another dataset's motion or grid (-rotparent, -gridparent)slice offsets padded by a -gridparent
command line3dvolreg -rotparent <generated: rp_st> -gridparent <generated: gp27> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_st>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.036, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/rp-obliqueanother dataset's motion or grid (-rotparent, -gridparent)an oblique series and -rotparent
command line3dvolreg -rotparent <generated: rp_obl> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_oblique>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.052, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/rp-twodupanother dataset's motion or grid (-rotparent, -gridparent)-gridparent overrides -twodup
command line3dvolreg -twopass -base <generated: base_moved> -twodup -rotparent <generated: rp> -gridparent <generated: gp27> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.015, volumes 0.0003, matrices 0.5, maximum displacements 0.5; 0/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/rp-savedispanother dataset's motion or grid (-rotparent, -gridparent)-rotparent with -savedisp
command line3dvolreg -rotparent <generated: rp> -savedisp sd.nii -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.052, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

60/60
original/tshiftslice timing first (-tshift)-tshift 0 (alt+z offsets in the header)
command line3dvolreg -tshift 0 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_st>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0.0019, matrices 0.73, maximum displacements 0.45; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/tshift-verbose-ignoreslice timing first (-tshift)-tshift 1 -verbose
command line3dvolreg -verbose -tshift 1 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_st>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.17, volumes 0.052, matrices 0.68, maximum displacements 0.5; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/tshift-seqslice timing first (-tshift)-tshift 0 with seq+z offsets
command line3dvolreg -tshift 0 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_st_seq>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.2, volumes 0.029, matrices 0.71, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/tshift-noneslice timing first (-tshift)-tshift on a series without offsets (a warning)
command line3dvolreg -tshift 0 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/tshift-large-ignoreslice timing first (-tshift)-tshift 3 on 6 volumes: too large (a warning)
command line3dvolreg -tshift 3 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_st>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/tshift-base-priorslice timing first (-tshift)-tshift 1 with -base 0 (a warning)
command line3dvolreg -tshift 1 -base 0 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_st>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.17, volumes 0.052, matrices 0.68, maximum displacements 0.5; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/tshift-one-brickslice timing first (-tshift)-tshift on one volume (a warning)
command line3dvolreg -tshift 0 -base <generated: base_ext> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: vol3>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.0085, volumes 0.0001, matrices 0.42, maximum displacements 0.46; 0/1 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/tshift-failsslice timing first (-tshift)-tshift with offsets beyond the TR: it fails (a warning)
command line3dvolreg -tshift 0 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_st_long>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/tshift-fmriprepslice timing first (-tshift)-tshift 0 with fMRIPrep's options
command line3dvolreg -tshift 0 -Fourier -twopass -zpad 4 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_st>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.013, volumes 0.014, matrices 0.32, maximum displacements 0; 4/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/twodup-twodup-twodup with -twopass and a moved base
command line3dvolreg -twopass -base <generated: base_moved> -twodup -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.015, volumes 0.035, matrices 0.5, maximum displacements 0.5; 0/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/twodup-onepass-twodup-twodup without -twopass (ignored)
command line3dvolreg -base <generated: base_moved> -twodup -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.094, volumes 0.0014, matrices 0.6, maximum displacements 0.54; 0/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/twodup-internal-twodup-twodup with an internal base (ignored)
command line3dvolreg -twopass -twodup -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0.052, matrices 0.73, maximum displacements 0.46; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/twodup-st-twodup-twodup on a series with slice offsets
command line3dvolreg -twopass -base <generated: base_moved> -twodup -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_st>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.015, volumes 0.035, matrices 0.5, maximum displacements 0.5; 0/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/twopasstwo passes and the coarse search (-twopass)-twopass
command line3dvolreg -twopass -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0.052, matrices 0.73, maximum displacements 0.46; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/twoblur-3two passes and the coarse search (-twopass)-twopass -twoblur 3
command line3dvolreg -twopass -twoblur 3 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.047, volumes 0.052, matrices 0.59, maximum displacements 0.46; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/twoblur-ffttwo passes and the coarse search (-twopass)-twopass -twoblur 16: the FFT blur
command line3dvolreg -twopass -twoblur 16 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0.052, matrices 0.72, maximum displacements 0.5; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/coarse-shift-onlytwo passes and the coarse search (-twopass)-coarse 2 1
command line3dvolreg -twopass -coarse 2 1 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0.052, matrices 0.73, maximum displacements 0.46; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/coarse-4-3two passes and the coarse search (-twopass)-coarse 4 3
command line3dvolreg -twopass -coarse 4 3 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: big>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0, matrices 0.59, maximum displacements 0.38; 1/4 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/coarserottwo passes and the coarse search (-twopass)-twopass -coarserot (a notice)
command line3dvolreg -twopass -coarserot -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0.052, matrices 0.73, maximum displacements 0.46; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/sinittwo passes and the coarse search (-twopass)-twopass -sinit (no scale initialisation)
command line3dvolreg -twopass -sinit -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0.052, matrices 0.73, maximum displacements 0.46; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/maxite-raisedtwo passes and the coarse search (-twopass)-twopass -maxite 5 (raised to 66)
command line3dvolreg -twopass -maxite 5 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0.052, matrices 0.73, maximum displacements 0.46; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/twopass-lineartwo passes and the coarse search (-twopass)-twopass -linear
command line3dvolreg -twopass -linear -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.0095, volumes 0.0001, matrices 0.65, maximum displacements 0.52; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/twopass-final-nntwo passes and the coarse search (-twopass)-twopass -Fourier -final NN
command line3dvolreg -twopass -Fourier -final NN -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0, matrices 0.096, maximum displacements 0; 6/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

44/44
original/twopass-deltatwo passes and the coarse search (-twopass)-twopass -delta 0.3
command line3dvolreg -twopass -delta 0.3 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0.002, matrices 0.72, maximum displacements 0.44; 3/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/weightweights and edging-weight vol[0]
command line3dvolreg -weight <generated: weight>[0] -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.019, volumes 0.0002, matrices 0.62, maximum displacements 0.46; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/weight-4dweights and edging-weight 4D[1]
command line3dvolreg -weight <generated: weight4d>[1] -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.019, volumes 0.0002, matrices 0.62, maximum displacements 0.46; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/wtinpweights and edging-wtinp
command line3dvolreg -wtinp -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.028, volumes 0.0002, matrices 0.66, maximum displacements 0.44; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/weight-wtinpweights and edging-weight with -wtinp (overridden)
command line3dvolreg -wtinp -weight <generated: weight>[0] -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.019, volumes 0.0002, matrices 0.62, maximum displacements 0.46; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/edging-3weights and edging-edging 3
command line3dvolreg -edging 3 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.21, volumes 0.052, matrices 0.76, maximum displacements 0.5; 3/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/edging-10pcweights and edging-edging 10%
command line3dvolreg -edging 10% -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.2, volumes 0.029, matrices 0.74, maximum displacements 0.52; 3/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/edging-0weights and edging-edging 0
command line3dvolreg -edging 0 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.028, volumes 0.0002, matrices 0.66, maximum displacements 0.44; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/weight-twopassweights and edging-weight with -twopass
command line3dvolreg -twopass -weight <generated: weight>[0] -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.24, volumes 0.0025, matrices 0.78, maximum displacements 0.53; 1/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/paramsthe fit's settings-params 5 0.02 0.05 0.9
command line3dvolreg -params 5 0.02 0.05 0.9 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.11, volumes 0.053, matrices 0.73, maximum displacements 0.42; 5/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/maxite-3the fit's settings-maxite 3
command line3dvolreg -maxite 3 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/maxdthe fit's settings-maxd 3 (read as -maxite)
command line3dvolreg -maxd 3 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/x-threshthe fit's settings-x_thresh 0.05
command line3dvolreg -x_thresh 0.05 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/rot-threshthe fit's settings-rot_thresh 0.1
command line3dvolreg -rot_thresh 0.1 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.048, volumes 0.0019, matrices 0.72, maximum displacements 0.45; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/delta-0.5the fit's settings-delta 0.5
command line3dvolreg -delta 0.5 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.21, volumes 0.029, matrices 0.72, maximum displacements 0.49; 4/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/env-edgingenvironment variablesAFNI_VOLREG_EDGING=10
command lineAFNI_VOLREG_EDGING=10 3dvolreg -verbose -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.28, volumes 0.054, matrices 0.55, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/env-edging-percentenvironment variablesAFNI_VOLREG_EDGING=40% (error)
command lineAFNI_VOLREG_EDGING=40% 3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/env-edging-badenvironment variablesAFNI_VOLREG_EDGING=-2 (error)
command lineAFNI_VOLREG_EDGING=-2 3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/env-quietenvironment variablesAFNI_QUIET_STARTUP=YES
command lineAFNI_QUIET_STARTUP=YES 3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/env-type-warnenvironment variablesAFNI_NIFTI_TYPE_WARN=YES on uint16 and a float64 base
command lineAFNI_NIFTI_TYPE_WARN=YES 3dvolreg -base <generated: base_f64x4> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_u16>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.013, volumes 0.0001, matrices 0.42, maximum displacements 0; 3/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/type-notice-onceenvironment variablesthe type conversion notice once (uint16 input, float64 base)
command line3dvolreg -base <generated: base_f64x4> -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1_u16>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.013, volumes 0.0001, matrices 0.42, maximum displacements 0; 3/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/env-deconflict-noenvironment variablesAFNI_DECONFLICT=NO with an existing output
command lineAFNI_DECONFLICT=NO 3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Rejected by both

both refuse the input with the same messages

10/10
original/env-firfacenvironment variablesAFNI_BLUR_FIRFAC=3 with -twopass
command lineAFNI_BLUR_FIRFAC=3 3dvolreg -twopass -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.19, volumes 0.052, matrices 0.73, maximum displacements 0.46; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/env-norealpathenvironment variablesAFNI_NOREALPATH=YES with -verbose
command lineAFNI_NOREALPATH=YES 3dvolreg -verbose -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/global-overwriteglobal options-overwrite with an existing output
command line3dvolreg -overwrite -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

34/34
original/global-D-edgingglobal options-DAFNI_VOLREG_EDGING=3
command line3dvolreg -DAFNI_VOLREG_EDGING=3 -verbose -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.21, volumes 0.052, matrices 0.76, maximum displacements 0.5; 3/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/global-deconflictglobal options-DAFNI_DECONFLICT=YES with an existing output
command line3dvolreg -DAFNI_DECONFLICT=YES -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

34/34
original/global-nifti2global options-DAFNI_NIFTI_WRITE_TYPE=2
command line3dvolreg -DAFNI_NIFTI_WRITE_TYPE=2 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/global-skip-afnircglobal options-skip_afnirc
command line3dvolreg -skip_afnirc -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.011, volumes 0.0003, matrices 0.53, maximum displacements 0.43; 2/6 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

46/46
original/global-afterglobal options-overwrite after the dataset
command line3dvolreg -prefix out.nii <generated: s1> -overwrite
Bit-identical

AFNI’s bytes: every file it writes, the messages and the exit status; the default against the replica’s default the same

10/10
original/stdin-wordsglobal options-@: the options read from standard input (against the replica only)
command line3dvolreg -@
Bit-identical

AFNI’s bytes: every file it writes, the messages and the exit status; the default against the replica’s default the same

10/10
original/stdin-emptyglobal options-@ with nothing on standard input (against the replica only)
command line3dvolreg -prefix out.nii -@
Rejected by both

The clean room and the replica refuse the input alike

10/10
original/threads-hepticthread countsthe default at 1, 2, 5, 14 threads: every byte the same
command line3dvolreg -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Bit-identical

every file, message and exit status the same at 1, 2, 5 and 14 threads (LARMORX against itself)

4/4
original/threads-fmriprepthread countsthe default at 1, 2, 5, 14 threads: every byte the same
command line3dvolreg -Fourier -twopass -zpad 4 -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: s1>
Bit-identical

every file, message and exit status the same at 1, 2, 5 and 14 threads (LARMORX against itself)

4/4
original/threads-odd-fourierthread countsthe default at 1, 2, 5, 14 threads: every byte the same
command line3dvolreg -Fourier -savedisp sd.nii -prefix out.nii -1Dfile p.1D -dfile d.1D -1Dmatrix_save m.aff12.1D -maxdisp1D md.1D <generated: odd>
Bit-identical

every file, message and exit status the same at 1, 2, 5 and 14 threads (LARMORX against itself)

4/4
original/fmriprep-ds000005-run-01 worked examplefMRIPrep's callsfMRIPrep's RobustAverage call: single-band, int16 stored, 64 x 64 x 34, 20 volumes
command line3dvolreg <fMRIPrep's sliced volumes: ds000005-run-01>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.005, volumes 0.0003, matrices 0.072, maximum displacements 0; 17/20 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

49/49
original/fmriprep-ds000005-run-03fMRIPrep's callsfMRIPrep's RobustAverage call: two volumes
command line3dvolreg <fMRIPrep's sliced volumes: ds000005-run-03>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0, volumes 0.0001, matrices 0.027, maximum displacements 0; 2/2 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

49/49
original/fmriprep-ds005454fMRIPrep's callsfMRIPrep's RobustAverage call: multiband, 160 x 160 x 96, oblique
command line3dvolreg <fMRIPrep's sliced volumes: ds005454>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.0071, volumes 0.0048, matrices 0.12, maximum displacements 0; 10/20 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

49/49
original/fmriprep-ds003592-echo-1fMRIPrep's callsfMRIPrep's RobustAverage call: multi-echo (echo 1), oblique
command line3dvolreg <fMRIPrep's sliced volumes: ds003592-echo-1>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.0048, volumes 0.0005, matrices 0.0079, maximum displacements 0; 19/20 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

49/49
original/fmriprep-ds004958fMRIPrep's callsfMRIPrep's RobustAverage call: multiband 8, uint16 stored, three volumes, 24 degrees oblique
command line3dvolreg <fMRIPrep's sliced volumes: ds004958>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.01, volumes 0.0045, matrices 0.27, maximum displacements 0; 1/3 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

49/49
original/fmriprep-ds002603-mb4fMRIPrep's callsfMRIPrep's RobustAverage call: Philips multiband 4, float32 stored
command line3dvolreg <fMRIPrep's sliced volumes: ds002603-mb4>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0, volumes 0.0001, matrices 0.011, maximum displacements 0; 20/20 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

49/49
original/fmriprep-pepolar-ds005040fMRIPrep's callsfMRIPrep's RobustAverage call: sdcflows' PEPOLAR reference average, two topup-corrected volumes
command line3dvolreg <fMRIPrep's sliced volumes: pepolar-ds005040>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.0099, volumes 0.0004, matrices 0.23, maximum displacements 0; 1/2 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

49/49
original/fmriprep-pepolar-ds001600fMRIPrep's callsfMRIPrep's RobustAverage call: the same, 80 x 80 x 60
command line3dvolreg <fMRIPrep's sliced volumes: pepolar-ds001600>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.0033, volumes 0.0002, matrices 0.2, maximum displacements 0; 1/2 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

49/49
original/fmriprep-ds000256-baselinefMRIPrep's callsfMRIPrep's RobustAverage call: a child's run (the motion levels' baseline)
command line3dvolreg <fMRIPrep's sliced volumes: ds000256-baseline>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.0025, volumes 0.0001, matrices 0.0096, maximum displacements 0; 19/20 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

49/49
original/fmriprep-ds000256-nofeedback worked examplefMRIPrep's callsfMRIPrep's RobustAverage call: a child's run, the most motion (rotations to 4.7 degrees)
command line3dvolreg <fMRIPrep's sliced volumes: ds000256-nofeedback>
Within the original's own variability

within specs/3dvolreg.md §14.3’s limits (either side of the shear decomposition’s exact ties): the largest shares of a limit, parameters 0.12, volumes 0.68, matrices 0.88, maximum displacements 0.46; 7/20 -1Dfile lines identical; exit status, files, headers and messages AFNI’s; the default within the same limits of the replica’s

49/49

Corrections by default

Where AFNI has a deterministic defect that changes a result, the replica corrects it by default, fix, and warns. upstream_defects="reproduce" (on the command line --upstream-defects reproduce from the next release, LARMORX_UPSTREAM_DEFECTS=reproduce in 0.1.0rc1) gives AFNI’s behaviour byte for byte where AFNI’s is defined (the API page’s table).

  • -nocoarserot turns the coarse rotation search off, as documented (AFNI-86: AFNI reads it as -noclip, whose four letters it tests first).
  • A byte series is written with its values rounded (AFNI-87: AFNI stretches them to the full byte range).
  • Output names of 256 bytes or more are written (AFNI-90: AFNI aborts on a buffer overflow; -maxdisp1D’s name is cut at 997 bytes).
  • The coarse rotation search’s mask always uses its own clip fraction (AFNI-92: -nomaxdisp changes it in AFNI), and an empty search region is not taken as the best (AFNI-93).
  • A second -base is refused (AFNI-95: AFNI keeps the last).
  • A Fourier rotation that would clear a row through a null pointer clears it (AFNI-97: AFNI crashes).
  • Two messages name the dataset they are about (AFNI-98: AFNI prints (null) or a stray argument).
  • The maximum displacements are those of the voxels themselves (AFNI-99, above).

Inputs that make no sense stop with a clear error in fix and in reproduce: an option given last without its value (AFNI-88), a base on which the least squares cannot be set up (AFNI-89), -coarse 1 on a volume of 120 or more voxels along every axis (AFNI-94: AFNI loops for ever), a -rotparent without its centres (AFNI-96). The clean room makes every one of these corrections by default and names them alike; its record holds each case’s fix against the replica’s.

Known limits

The replica’s standard comparison holds fMRIPrep’s RobustAverage calls on six BOLD runs of five fMRIPrep runs (single-band, multiband, multi-echo, uint16 and float32 storage) and sdcflows’ two PEPOLAR reference averages. The full comparison (seven more of the data panel’s runs) and long random-testing runs come in the final validation phase.

Sub-brick selectors (dset[n]) and AFNI’s own .HEAD/.BRIK format are refused, as by every LARMORX AFNI program, and LARMORX writes no AFNI header extension, so a LARMORX output cannot serve as a -rotparent (AFNI’s can). The cross-platform checks (how they work) hold the other five platforms to LARMORX’s own Linux output.

The benchmark report’s times were measured with other work on the machine beside them, so they are long: the ratios between the programs measured together are the information.

The clean room is held to the specification’s limits, not to AFNI’s bits, wherever the ties decide them. AFNI is deterministic, so a band of its own variability from imperceptible changes of its input (as for mcflirt) comes in the final validation phase, with the clean room’s full comparison and long random-testing runs.

History

  • 2026-11-06: the replica’s first timings, informal.
  • 2026-11-07: the replica’s record, 190 cases against AFNI 26.3.05.
  • 2026-11-08: the clean room, written from specs/3dvolreg.md; its record (251 cases), its random-testing run of half an hour and the benchmark report.

Data credits

The data this page shows or names: each dataset with its version, DOI and licence, from LARMORX's test-data catalog. Public-domain dedications (CC0, PDDL) ask nothing; the datasets are cited all the same.