3ddelay throws unexpected error

Hi all,

An unexpected error:

3ddelay -nodtrnd -maskMSK.nii.gz -nodsamp -fs 2 -T 0 -ideal_file REF.1D -input DATASET.nii.gz -prefix DELAYMAP.nii.gz

Throws an error (or, rather, throws this error as many times as I there are voxels):


Error: Initialization problem!

Error origin: hilbertdelay_V2


If I use a polort (i.e., delete “nodtrnd”), the error does not occur. If I use a different chunk of that time series (with an appropriately sized chunk of “REF.1D”), and keep “-nodtrnd”, the error doesn’t occur.

Any idea what might be going on? The relevant part of the “.log” is copied below.

Thanks,
-Daniel

Logfile output by Hilbert Delay98 plugin

User Data Values
Input Data Set: DATASET.nii.gz
Mask Data Set: MSK.nii.gz
Ascii output file name: (NULL)
Reference File Name: REF.1D
Number of voxels in X direction: 64
Number of voxels in Y direction: 64
Number of voxels in Z direction: 20
Sampling Frequency = 2.000000
Stimulus Period = 0.000000
Delay units = 0
Delay wrap = 0
Delay reverse = 0
Delay scale = 1.000000
Number of segments = 1
Length of reference time series = 501
Number of fit parameters = 2
Number of nuisance parameters (orts)= 0
Percent overlap = 0
Plugin detrending = (Always 0, mandatory detrending is performed)
Bias correction = 1
Acquisition time correction = 0
Correlation Coefficient Threshold= 0.500000
Flag for Ascii output file = 0
Flag for Ascii time series file output = 0

option_data->errcode (debugging only)= 0

The format for the output file is the following:
VI X Y Z Duff Del Cov xCorCoef VTS

Error Log

Could not find zero crossing before maxdel limit 1235 19 19 0
De Fault, De Fault (21), the two sweetest words in the english langage ! 1236 20 19 0
De Fault, De Fault (21), the two sweetest words in the english langage ! 1249 33 19 0
De Fault, De Fault (21), the two sweetest words in the english langage ! 1250 34 19 0
De Fault, De Fault (21), the two sweetest words in the english langage ! 1251 35 19 0

Hi, any updates (or clues I could use to chase this down) here?

Howdy-

Sorry, I didn’t catch this before. And offhand, I am not sure what the problem might be. I will share a drive to get the dsets in question from you, in order to investigate, if that works.

–pt

After (too) much delay (by me), this appears to be resolved now in AFNI v22.2.08.

The issue was that a voxel where a “short enough” shift value was not found could stop the rest of the progression of voxel calculations. Now, that shouldn’t happen.

–pt