Show all posts by user
Dear AFNI users-
We are very pleased to announce that the new AFNI Message Board framework is up! Please join us at:
https://discuss.afni.nimh.nih.gov
Existing user accounts have been migrated, so returning users can login by requesting a password reset. New users can create accounts, as well, through a standard account creation process. Please note that these setup emails might initially go to spam folders (esp. for NIH users!), so please check those locations in the beginning.
The current Message Board discussion threads have been migrated to the new framework. The current Message Board will remain visible, but read-only, for a little while.
Sincerely,
AFNI HQ
History of AFNI updates
I just clicked the search box without looking. "FALLBACK" turned up nothing, but "dylib" (which I thought I'd tried) turns up a bunch of stuff when I search now, and not just this thread. So I'll just chalk this one up to embarrassing search-fail.
by
Isaac Schwabacher
-
AFNI Message Board
This question comes up on these boards all the time, but for some reason the search doesn't seem to find it (which might be related to afni.nimh.nih.gov dropping off of google?). It's also not in the FAQs, which is surprising. Try setting your DYLD_FALLBACK_LIBRARY_PATH to the directory where AFNI is installed:
If you are using bash (the default on OS X) and your AFNI binaries are
by
Isaac Schwabacher
-
AFNI Message Board
Hi gurus,
Sorry to bother again, but I'm trying to get 3dDetrend to read 1D input from stdin, which it doesn't want to do. The problem is that it's trying to read the input as a 3D dataset, which means that only filenames that begin with '1D:' or end with '.1D' will be treated as 1D datasets, and there is currently no universal AFNI name for stdin that match
by
Isaac Schwabacher
-
AFNI Message Board
Hi maintainers of the python libraries,
I just noticed a bug in method cols_by_label_list() of class Afni1D in module lib_afni1D.py at line 1591 (as of the Jan 15 build). A quick grep shows that this method is completely unused, so this is low priority, but I'm using this method in my code, so I've fixed it for myself and I thought I'd pass on the fix. The code as it stands is
by
Isaac Schwabacher
-
AFNI Message Board
Use -ONELINE as the first option to cat_matvec. That forces the output matrix into a single line (form 1a in the cat_matvec help).
by
Isaac Schwabacher
-
AFNI Message Board
Oh, I see now. The buffer is 16385 bytes long, the message is allowed to be 32222 bytes long "for Ziad (who else is so crazy?)" plus the admonition, and kablooie. I guess I'll use -NN 1 because undefined behavior is scary.
by
Isaac Schwabacher
-
AFNI Message Board
Hi gurus,
I just ran into some memory nastiness with 3dClustSim using a small mask. It happened after the output file was successfully written, so it probably doesn't matter in the grand scheme of things, but I figure that any time malloc() blows chunks you'd like to hear about it.
Here are some small failure cases:
$ 3dUndump -srad 2 -dimen 10 10 10 - <<<'5 5 5&
by
Isaac Schwabacher
-
AFNI Message Board
Kausar,
I don't know what other people are using, but I've had success with Erik Beall's PESTICA package, which you can find here:
Hope this helps,
Isaac
by
Isaac Schwabacher
-
AFNI Message Board
Glad to hear it. Those three little lines nerd-sniped me so badly.
by
Isaac Schwabacher
-
AFNI Message Board
Hi folks,
I just updated my binaries and noticed this little message:
Combining all command completions into all_progs.COMP
Not ready to setup completion for shell bash.
Complain to AFNI folks.
I'm a bash user and this seemed useful, so I did some quick reading on tcsh and bash command completion. Rather than complaining, I wrote a little snippet of bash code that will translate the t
by
Isaac Schwabacher
-
AFNI Message Board
I like the idea of a shading in the background. Maybe something like the visual in the technical note for @ANATICOR.
by
Isaac Schwabacher
-
AFNI Message Board
In the case I mentioned, the cardiac data was collected years ago and originally sampled at 1 kHz as standard procedure rather than for any particular purpose, so the downsampling left a signal that was still sampled at ~30 Hz, which seemed reasonable for retroicor purposes. (I have no idea why the standard procedure was to sample at such high frequency.)
by
Isaac Schwabacher
-
AFNI Message Board
How big of a cube did you have to extract in order to get the interpolation to work? I know I experimented with trying to use 3dresample to upsample some cardiac data I had (since 1dUpsample only works for integer ratios, and the collection equipment had spontaneously decided to reduce the sampling frequency by a factor of 35.7 or something for a period of half a minute) and I discovered that I
by
Isaac Schwabacher
-
AFNI Message Board
Hi gurus,
I just want to report a bit of 3dTstat behavior that I think may be unintended, if a little amusing: $ 3dTstat -prefix stdout: -mean processed/004/epis/warm.mtn.1D\'
++ 3dTstat: AFNI version=AFNI_2011_12_21_1014 (Oct 22 2012) [64-bit]
++ Authored by: KR Hammett & RW Cox
*+ WARNING: Input dataset is not 3D+time; assuming TR=1.0
0.09242
0.0596765
-0.142048
0.155061
by
Isaac Schwabacher
-
AFNI Message Board
I would recommend against simply stripping out outlier time points, since many AFNI programs assume that the volumes in a 3d+time dataset occur at regular intervals. If you throw out volumes in the middle of the timeseries you will break this guarantee and many programs will silently give you incorrect output. What you can do is create a censor file, which 3dDeconvolve and various other analysi
by
Isaac Schwabacher
-
AFNI Message Board
...in the Makefile that I modified because I have gcc in /usr/bin and forgot I'd changed. /sheepish
by
Isaac Schwabacher
-
AFNI Message Board
I ended up needing libexpat, and I wonder if I misinterpreted the OpenMotif license-- it's only valid for open source platforms and while OS X isn't open source, AFAIK Darwin is. In the end I did have to knuckle under and actually install fink. I know, such a sacrifice... ;P
EDIT: Also, is there a reason that some of compiler variables in the Makefile.macosx_10.7_Intel_64 are set t
by
Isaac Schwabacher
-
AFNI Message Board
Huh. No, <Xm/*> is not included by mrilib.h, but various X/Motif libraries appear in the general link flags in the makefile. I must have gotten confused because I was trying to test-compile 1dplot a while ago and that definitely does need Motif. Time to see how many libraries we can strip out of the makefile.
by
Isaac Schwabacher
-
AFNI Message Board
I'm trying to test build a custom program (yes, Gabe, that one... sorry it's taken so long to get to it) on OS X 10.7 but I'm trying to avoid having to install all the fixings (in particular LessTif since it hasn't been updated since 2009 and OpenMotif isn't licensed for OS X). My program doesn't use the GUI in any way and I'm intending to use the pre-built bin
by
Isaac Schwabacher
-
AFNI Message Board
Erm... well... I will admit to not being particularly knowledgable about the atlases packaged with AFNI, but my data is currently processing, so let's see what we can do.
I looked at and found that atlas information is stored in AFNI_atlas_spaces.niml, and when I looked in that file, I found <ATLAS
atlas_name="CA_N27_ML"
dset_name="TT_caez_ml_18+tlrc"
...
by
Isaac Schwabacher
-
AFNI Message Board
I think the datasets you're looking for already exist. Try looking in the same directory that the AFNI binaries are in (type "which afni" if you don't know where that is). I think all of the "Where am I?" atlases are there.
by
Isaac Schwabacher
-
AFNI Message Board
Most Merciful Emperor Zhark,
1dplot is behaving strangely. I've got several reports to make, only one of which I actually looked for in the source code:
First, the y axis is getting pushed any time there is more than one box, even if the -yaxis option has been given. The -nopush option only makes things weirder:
1dplot -yaxis 0:.15:1:1 '1D:0 0' #works
1dplot -yaxis -on
by
Isaac Schwabacher
-
AFNI Message Board
Thanks! I just looked back at the code I suggested last night and can't really figure out what I was thinking, since out_prefix() strips the path information and afni_name.new(prefix) strips the prefix, so nothing is left of the original afni_name instance. So that might not be the best way to do that...
by
Isaac Schwabacher
-
AFNI Message Board
Hi gurus,
I've been having a couple of issues with align_epi_anat.py recently. The first issue is that the script crashes when I try to run it on EPIs that are not in the current directory. The temporary and output files are getting created in the current working directory (which is where I want them, and where fresh_start() looks for them), but the script subsequently searches for them
by
Isaac Schwabacher
-
AFNI Message Board
Thanks for the advice.
I was considering just assuming that the first few frames were in register, but it occurred to me that if the subject is surprised by the sudden burst of noise when the scan starts, that's precisely when you'll see motion. Then again, in that case you'll see the motion during the first frame rather than two seconds later between frames, so that would spea
by
Isaac Schwabacher
-
AFNI Message Board