Back to the main page.
Bug 1290 - zlim colorbar weird on ft_singleplotTFR
Status | CLOSED FIXED |
Reported | 2012-01-25 17:42:00 +0100 |
Modified | 2012-02-29 17:31:09 +0100 |
Product: | FieldTrip |
Component: | documentation |
Version: | unspecified |
Hardware: | All |
Operating System: | All |
Importance: | P3 minor |
Assigned to: | Roemer van der Meij |
URL: | |
Tags: | |
Depends on: | |
Blocks: | |
See also: |
Johanna - 2012-01-25 17:42:03 +0100
It appears in the http://fieldtrip.fcdonders.nl/tutorial/timefrequencyanalysis, in Figure 4. Why does the colorbar show dark blue at zero, green at the highest value, and then blank for the lower values? (Instead, shouldn't it be the normal scale from dark blue to green to red, right?) I replicated this just now using the current FT version with the data from the wiki.
Roemer van der Meij - 2012-01-25 18:10:04 +0100
This is a matlab opengl-transparency bug if I'm correct (also so visible but the 'invisible' axes-box). It was the reason I ended up implementing cfg.maskstyle = 'saturation' for single and multiplot. If the figure is remade with this option, it should look proper. I've never uploaded a figure to the wiki though. If we were at the bug-binge right now, I'd scream out loud for someone who has :)
Roemer van der Meij - 2012-01-25 18:11:06 +0100
The same holds for figure 6 as well
Jan-Mathijs Schoffelen - 2012-01-26 08:55:46 +0100
Uhm, isn't this a manifestation of a well-known rendering issue. You need to specify the proper renderer in cfg.renderer (I keep forgetting which). I don't know whether there is already a FAQ devoted to this, but it would be good to make one if there isn't. Also, please update the figures on the wiki with the fitting colorbar
Roemer van der Meij - 2012-01-26 10:33:46 +0100
I don't think a different renderer will help, as the mentioned figures use transparency (alpha data) masking. OpenGL is the only renderer that can do this IIRC, zbuffer en software are unable to. However, a similar thing can be achieved by using zbuffer and using a saturation mask, instead of a transparency mask.
Jan-Mathijs Schoffelen - 2012-01-27 08:41:00 +0100
assigned bug to get all bugs lined up before the grand bug binge
Roemer van der Meij - 2012-02-01 14:53:02 +0100
Johanna, do you often change figures on the wiki and have easy access to the tutorial data? If so, could I ask you to do the fix by remaking the figures with cfg.maskstyle = 'saturation'? I don't really work with the wiki, and would have to spend some time on how to do it properly. Let me know if you are equally inexperienced, than I'll find out how to do it myself ;)
Roemer van der Meij - 2012-02-01 15:34:36 +0100
Done!
Johanna - 2012-02-01 15:50:20 +0100
reopened, as reminder to add link on this tutorial to the FAQ on how to make the figures look nice, in case they run into the same problem.
Roemer van der Meij - 2012-02-01 16:24:31 +0100
aaaaaand, refixed! :) (there is now a link to the relevant faq in the fig descriptions)