Back to the main page.

Bug 3090 - ft_volumerealign: intensity sliders fall outside the figure window

Reported 2016-03-07 21:20:00 +0100
Modified 2016-05-05 20:08:50 +0200
Product: FieldTrip
Component: core
Version: unspecified
Hardware: PC
Operating System: Mac OS
Importance: P5 normal
Assigned to: Roemer van der Meij
Depends on:
See also:

Jan-Mathijs Schoffelen - 2016-03-07 21:20:13 +0100

amici! which of your brilliant minds has implemented the intensity sliders in ft_volumerealign. I'm using Matlab2014b and only the left one is partially visible and hardly operable. Please look into whether the positioning within the figure can be done a bit more robustly.

Arjen Stolk - 2016-03-07 21:59:37 +0100

(In reply to Jan-Mathijs Schoffelen from comment #0) It wasn't me, Shaggy. Do you have the same issue with ft_electrodeplacement? Try ft_electrodeplacement([], mri). If I remember correctly, Roemer based the volumerealign sliders on the electrodeplacement implementation.

Arjen Stolk - 2016-03-08 06:00:59 +0100

(In reply to Arjen Stolk from comment #1) North California to South; can you confirm this?

Robert Oostenveld - 2016-03-08 15:38:17 +0100

git blame ft_volumerealign.m | grep slider suggests Roemer had something to do with it. Pfew, it is not me this time.

Roemer van der Meij - 2016-03-08 22:02:33 +0100

Ha, yeah, that should be me. Could you provide a bit more info? I see now the h123size are computed a bit more complicated than I assumed. What cfg.axisratio are you using?

Jan-Mathijs Schoffelen - 2016-03-08 22:11:58 +0100

I did: mri = ft_volumerealign([], mri); dus, geen user-specified cfg arguments.

Roemer van der Meij - 2016-03-09 00:20:10 +0100

Alright, it should be more robust now (pull requested). It's difficult to be sure, because I can't replicate it. JM, can you test and see if it solves the issue? If not, can you attach the MRI? The anatomy can be replaced with rand and the header partly removed to anonymize. And Arjen, don't get cocky, I copied the positioning over from your work in ft_electrodeplacement ;)

Arjen Stolk - 2016-03-09 00:31:19 +0100

Cocky?! I just never make mistakes (or no longer realize I make 'm :P). But maybe the captain was using an oddly sized MR, busting the sizes of the entire figure?

Roemer van der Meij - 2016-03-09 00:34:21 +0100

Oh, so now it's suddenly the fault of the MRI? Tssk tssk ;) ;)

Arjen Stolk - 2016-03-09 05:40:31 +0100

If it were an MR of the captain's brain himself, that would explain the odd dimensions, right? :P

Roemer van der Meij - 2016-03-09 18:09:43 +0100

(In reply to Arjen Stolk from comment #9) For sure, its dimensions would exceed anything we can plot. In normalized units, it will even exceed 1.

Arjen Stolk - 2016-03-09 18:38:34 +0100


Robert Oostenveld - 2016-03-10 10:13:38 +0100

(In reply to Roemer van der Meij from comment #10) LOL!

Jan-Mathijs Schoffelen - 2016-05-05 09:44:43 +0200

Roemer, can you look into this? At the Nijmegen toolkit we ran into problems for this function (leading to the function crashing), because we used MATLAB2012b. The functionality you wrote is not supported by older versions of MATLAB. We aim to be backward compatible for ~5 years, so please look into this. It has to do with the sorting of some handles in lines 1604-1606.

Roemer van der Meij - 2016-05-05 19:08:10 +0200

Ouch! That's not supposed to happen of course. Sorry. Ha, funny. I always thought my coding was immune to this because of using 2012 for all those years. I thought wrong! I'll look into it.

Roemer van der Meij - 2016-05-05 19:54:26 +0200

Pull request placed. Error was due to a figure property not being present prior to the new graphics engine introduced in 2014b. Luckily, neither was the problem I fixed with this figure property ;). BTW, wasn't our usual practice to test all tutorials, using our testing-framework, prior to picking a fieldtrip version for the toolkit? The test scripts should have been failing for a month or two.

Jan-Mathijs Schoffelen - 2016-05-05 20:03:39 +0200

(In reply to Roemer van der Meij from comment #15) Yes, the tutorials were tested with the FT version we used. Yet, ft_volumerealign does not feature in the tutorials, but the problem showed up when people started doing their own stuff... Can this one be closed?

Jan-Mathijs Schoffelen - 2016-05-05 20:08:50 +0200

Hi Roemer, Thanks for fixing. I placed a comment in the pull request, because it could be that the problem you originally stumbled into, also affects ft_sourceplot in matlab2014 and up. It would be great if you could have a quick peek.