Holcombe:LumColorPsychopy

SUPA Sydney University Perception and Action Lab Primer editing help refcard

Topic: existing monitor calibration Jonathan Peirce <jon.peirce@gmail.com> Jun 03 09:11AM +0100 ^

ah, right, yes.

For color spaces rgb, dkl and lms this is the case. Believe it or not that was a deliberate decision taken during the overhaul of colorspaces in 1.61. The reason is that with 256 LUT values, if you use the brightest and darkest then there isn't an entry that's half-way between them ('0.0' translates to LUT entry 127.5). That would mean that, for instance a grating has a mean luminance slightly higher (or lower, potentially) than that of the screen. For these 3 colour spaces the centre point is meaningful and important and I decided it was preferable that 0 was exactly in between -1 and 1. For this to be possible you need an odd number of LUT entries, so psychopy skips the top entry and 0.0 becomes 127)

BUT, if you want colours that include the top entry you can specify them either by name ('white') hex value ('#ffffff') or using colorspace='rgb256' ( (255,255,255) ). These colour spaces assume that the user isn't interested in precise symmetry and wants specific rgb entries.

Apologies that this isn't documented. It was discussed (actually only by me ;-) ) on the forum: http://groups.google.com/group/psychopy-users/browse_thread/thread/4d8985ef94f5d81d?fwc=1 but apparently I never got around to explaining it in the colour spaces document.

all the best, Jon