Back to the main page.
Bug 127 - error using flag for improved freqanalysis
| Status | CLOSED FIXED |
| Reported | 2010-08-16 11:59:00 +0200 |
| Modified | 2010-08-19 14:55:40 +0200 |
| Product: | FieldTrip |
| Component: | core |
| Version: | unspecified |
| Hardware: | PC |
| Operating System: | Linux |
| Importance: | P1 normal |
| Assigned to: | Roemer van der Meij |
| URL: | |
| Tags: | |
| Depends on: | |
| Blocks: | |
| See also: |
Stephen Whitmarsh - 2010-08-16 11:59:42 +0200
entering a third input in freqanalysis: ft_freqanalysis(cfg, DataPlanarM,1) gives the following confusing error. And yes, it does not give a problem when i remove the flag, and indeed i did not specify any padding. best, S ??? Error using ==> specest_mtmconvol at 67 the padding that you specified is shorter than the data Error in ==> ft_freqanalysis at 277 [spectrum,ntaper,foi,toi] = specest_mtmconvol(dat, time, 'timeoi', cfg.toi, 'timwin', cfg.t_ftimwin, options{:}); Error in ==> do_TFRs_MM_back at 65 TFR_FPain = ft_freqanalysis(cfg, DataPlanarM,1);
Roemer van der Meij - 2010-08-18 10:05:34 +0200
Hi Stephen, Thanks for reporting! I changed a call to size(x,2) to numel when determining ntrials, that should fix it if you had data.trail cell-array as a trialx1 array (don't know why it wasn't like this in the first place). Does that fix it for you? BTW, the dpss part using variable time-windows is still not perfect. I haven't looked into it yet, but there is quite some difference to the old implementation in some cases. For Hanning windows it seems fine atm, but I am hoping for confirmation from other. Cheers, Roemer
Roemer van der Meij - 2010-08-18 10:06:27 +0200
(In reply to comment #0) > entering a third input in freqanalysis: ft_freqanalysis(cfg, DataPlanarM,1) > > gives the following confusing error. And yes, it does not give a problem when i > remove the flag, and indeed i did not specify any padding. > > best, > S > > ??? Error using ==> specest_mtmconvol at 67 > the padding that you specified is shorter than the data > > Error in ==> ft_freqanalysis at 277 > [spectrum,ntaper,foi,toi] = specest_mtmconvol(dat, time, 'timeoi', > cfg.toi, 'timwin', cfg.t_ftimwin, options{:}); > > Error in ==> do_TFRs_MM_back at 65 > TFR_FPain = ft_freqanalysis(cfg, DataPlanarM,1); Hi Stephen, Thanks for reporting! I changed a call to size(x,2) to numel when determining ntrials, that should fix it if you had data.trail cell-array as a trialx1 array (don't know why it wasn't like this in the first place). Does that fix it for you? BTW, the dpss part using variable time-windows is still not perfect. I haven't looked into it yet, but there is quite some difference to the old implementation in some cases. For Hanning windows it seems fine atm, but I am hoping for confirmation from other. Cheers, Roemer
Robert Oostenveld - 2010-08-18 12:25:40 +0200
test comment
Roemer van der Meij - 2010-08-18 12:27:29 +0200
(In reply to comment #3) > test comment comment received
Roemer van der Meij - 2010-08-18 13:22:00 +0200
(In reply to comment #2) > (In reply to comment #0) > > entering a third input in freqanalysis: ft_freqanalysis(cfg, DataPlanarM,1) > > > > gives the following confusing error. And yes, it does not give a problem when i > > remove the flag, and indeed i did not specify any padding. > > > > best, > > S > > > > ??? Error using ==> specest_mtmconvol at 67 > > the padding that you specified is shorter than the data > > > > Error in ==> ft_freqanalysis at 277 > > [spectrum,ntaper,foi,toi] = specest_mtmconvol(dat, time, 'timeoi', > > cfg.toi, 'timwin', cfg.t_ftimwin, options{:}); > > > > Error in ==> do_TFRs_MM_back at 65 > > TFR_FPain = ft_freqanalysis(cfg, DataPlanarM,1); > > > Hi Stephen, > > Thanks for reporting! I changed a call to size(x,2) to numel when determining > ntrials, that should fix it if you had data.trail cell-array as a trialx1 array > (don't know why it wasn't like this in the first place). Does that fix it for > you? > > BTW, the dpss part using variable time-windows is still not perfect. I haven't > looked into it yet, but there is quite some difference to the old > implementation in some cases. For Hanning windows it seems fine atm, but I am > hoping for confirmation from other. > > Cheers, > Roemer Could you quickly run your script again to see if that caused the problem? I'm not really sure, could be somewhere in deeper code... Cheers, Roemer
Stephen Whitmarsh - 2010-08-18 13:47:33 +0200
Hi Roemer, I'm afraid the problem persists... You're welcome to come by if you want to look for the problem here. (In reply to comment #5) > (In reply to comment #2) > > (In reply to comment #0) > > > entering a third input in freqanalysis: ft_freqanalysis(cfg, DataPlanarM,1) > > > > > > gives the following confusing error. And yes, it does not give a problem when i > > > remove the flag, and indeed i did not specify any padding. > > > > > > best, > > > S > > > > > > ??? Error using ==> specest_mtmconvol at 67 > > > the padding that you specified is shorter than the data > > > > > > Error in ==> ft_freqanalysis at 277 > > > [spectrum,ntaper,foi,toi] = specest_mtmconvol(dat, time, 'timeoi', > > > cfg.toi, 'timwin', cfg.t_ftimwin, options{:}); > > > > > > Error in ==> do_TFRs_MM_back at 65 > > > TFR_FPain = ft_freqanalysis(cfg, DataPlanarM,1); > > > > > > Hi Stephen, > > > > Thanks for reporting! I changed a call to size(x,2) to numel when determining > > ntrials, that should fix it if you had data.trail cell-array as a trialx1 array > > (don't know why it wasn't like this in the first place). Does that fix it for > > you? > > > > BTW, the dpss part using variable time-windows is still not perfect. I haven't > > looked into it yet, but there is quite some difference to the old > > implementation in some cases. For Hanning windows it seems fine atm, but I am > > hoping for confirmation from other. > > > > Cheers, > > Roemer > > Could you quickly run your script again to see if that caused the problem? I'm > not really sure, could be somewhere in deeper code... > > Cheers, > Roemer
Roemer van der Meij - 2010-08-18 14:03:37 +0200
Hmmm, that's odd, I can't replicate it here. Are you sure you are using the home/common version? I did find a bug related to padding, but that code wouldn't be reached if you didn't specify padding (you sure your cfg.pad is not an empty matrix?). I'll be at the DCCN next Wednesday :) (In reply to comment #6) > Hi Roemer, I'm afraid the problem persists... > You're welcome to come by if you want to look for the problem here. > > > (In reply to comment #5) > > (In reply to comment #2) > > > (In reply to comment #0) > > > > entering a third input in freqanalysis: ft_freqanalysis(cfg, DataPlanarM,1) > > > > > > > > gives the following confusing error. And yes, it does not give a problem when i > > > > remove the flag, and indeed i did not specify any padding. > > > > > > > > best, > > > > S > > > > > > > > ??? Error using ==> specest_mtmconvol at 67 > > > > the padding that you specified is shorter than the data > > > > > > > > Error in ==> ft_freqanalysis at 277 > > > > [spectrum,ntaper,foi,toi] = specest_mtmconvol(dat, time, 'timeoi', > > > > cfg.toi, 'timwin', cfg.t_ftimwin, options{:}); > > > > > > > > Error in ==> do_TFRs_MM_back at 65 > > > > TFR_FPain = ft_freqanalysis(cfg, DataPlanarM,1); > > > > > > > > > Hi Stephen, > > > > > > Thanks for reporting! I changed a call to size(x,2) to numel when determining > > > ntrials, that should fix it if you had data.trail cell-array as a trialx1 array > > > (don't know why it wasn't like this in the first place). Does that fix it for > > > you? > > > > > > BTW, the dpss part using variable time-windows is still not perfect. I haven't > > > looked into it yet, but there is quite some difference to the old > > > implementation in some cases. For Hanning windows it seems fine atm, but I am > > > hoping for confirmation from other. > > > > > > Cheers, > > > Roemer > > > > Could you quickly run your script again to see if that caused the problem? I'm > > not really sure, could be somewhere in deeper code... > > > > Cheers, > > Roemer
Stephen Whitmarsh - 2010-08-18 14:23:30 +0200
I just tried again with an updated (revision 1526) Fieldtrip in my own version. Oh, and i did not specify any padding. Cheers, S
Roemer van der Meij - 2010-08-19 14:23:21 +0200
It should work now, could you check again? :) (In reply to comment #8) > I just tried again with an updated (revision 1526) Fieldtrip in my own version. > Oh, and i did not specify any padding. > > Cheers, > S
Stephen Whitmarsh - 2010-08-19 14:48:58 +0200
Its running! (In reply to comment #9) > It should work now, could you check again? :) > > > (In reply to comment #8) > > I just tried again with an updated (revision 1526) Fieldtrip in my own version. > > Oh, and i did not specify any padding. > > > > Cheers, > > S