Gamma correction for blending 2 projector-images

May 16, 2006 at 1:33pm

Gamma correction for blending 2 projector-images

Hi there,

for a rather simple 2-projector video installation, I wrote a patch to blend the edges. Up to now just by using a grayscale matrix for the alpha channel that I send to an jit.gl.videoplane.

Now, what’s the actual problem: I am worried, that I will need some gamma correction in the blending area as up to now I am just adding pixel values in this area. Does anyone have some experiences with edge blending (because I don’t) in Jitter and how to do the gamma correction best? I know about the jit.displays abilities and was thinking about using the “coords”-message to define the area of gamma adjustment. I won’t be able to try around much with the two projectors running, so I have to prepare everything beforehand and want to get sure that I am on the right track – so thanks a lot for any advice.

best

Paul

#25996
May 18, 2006 at 10:43am

come om guys,

there should be somebody with some experiences in edge-blending.

please!

best

paul

#77185
May 18, 2006 at 2:37pm

never done the edge blend, but don’t think you can selectively adjust the gamma for only a portion of the display – it’s all or nothin.
You can adjust the gradient that you are using for the alpha channel using jit.charmap….

P.

#77186
May 18, 2006 at 2:48pm

hi tom

don’t know about your budget but there are projectors wich are able to
do this for you

jurgen

#77187
May 18, 2006 at 6:42pm

Note that in addition to modifying the gamma table of the display
device, you can also use jit.charmap to apply any kind of transfer
curve to either your color data or your alpha data necessary. Simply
fill your transfer curve look up table with the appropriate entries.
For gamma adjustment it would be something like output_value = pow
(input_value,gamma), where input/output values are (0.-1.). You’ll of
course need to convert to/from the range (0-255) for your lookup
table. That formula is a simpliication of the standard gamma equation
which might be satisfactory for your purposes, but here’s some more
info with more detailed description of how gamma is typically
calculated.

http://www.sjbrown.co.uk/?article=gamma

Hope this helps.

-Joshua

#77188

You must be logged in to reply to this topic.