RE: [info-performer] Framebuffer and RGB Color Values - Another problem

Date view Thread view Subject view Author view

From: faculaganymede (faculaganymede++at++yahoo.com)
Date: 06/10/2004 14:20:50

pfExperts,

Installing nogamma seems to have fixed the problem of
getting DVP output different from framebuffer
contents, but there is still one more problem.

In one of our tests, we increase the values of the RGB
colors across the image frame lines from 0 to 511,
generating an image ramp (a gradient of min to max
intensities across an image frame, a grayscale image
of black to white across). When we read the values of
the R and G color components from the DVP output, we
were getting the correct values, but there's a problem
with the B component.

For the B color component, we got some unexpected
outputs. The output patterns we were seeing for a
sequence of consecutive numbers as input is as follows
(values are show in binary, where x can be 0 or 1):

glWritePixels input DVP output

xxxx 1111 1110 ----> xxxx 1111 1110
xxxx 1111 1111 ----> xxx1 1111 1111
xxx1 0000 0000 ----> xxx1 0000 0000

So for example, if we write the following sequence of
values to B across the frame lines "..., 254, 255,
256,..." which is "..., 11111110, 11111111,
100000000,..." in binary, but in the DVP output we get
"..., 11111110, 111111111, 100000000,...".

We noted that this is a dynamic problem caused by the
bit transition from one value to another. If the
value is consistently the same across the frame lines,
e.g. set the value to "255" across the entire image
frame, then we do get "255" or "11111111" as DVP
output.

Did anyone else ever run into this problem before?
Any fixes?

=====
Thanks,
faculaganymede

        
                
__________________________________
Do you Yahoo!?
Friends. Fun. Try the all-new Yahoo! Messenger.
http://messenger.yahoo.com/

Date view Thread view Subject view Author view

This archive was generated by hypermail 2b29 : Thu Jun 10 2004 - 14:21:44 PDT