On 18 Apr, Keith Packard wrote:
>
> 
> I think the concensus is that gamma correction would be useful, but there 
> are several unanswered questions:
> 
> 1)    Are there other modifications to the compositing computation that
>       would also be useful?
> 
> 2)    Where are the intensity<->value conversions stored?
> 
> 3)    What precision is required in this computation?
> 
> 4)    How does this relate to XDCCC?
> 

To which I would add,

5) Why quit with gamma? Why not do a uniform perceptual response space,
 e.g. L* or GSDF ?  Once you've decided to deal with the issue it is not
 that hard to use a function that corresponds more accurately to the eye
 response than the simple power function of the gamma approximation.

6) What do we do with inadequate hardware support?

Since roughly 99% of the commercial display hardware for PC's has a very
limited accurate bit depth this is a significant issue.  If you measure
and adjust good consumer grade hardware into a uniform contrast
perception space you find that you only have about 7 bits of greyspace
to work with.  Adding color into the mix makes things harder still.
Since the answer "Use adequate hardware" is not a reasonable response,
we need to think through a reasonable approach for dealing with
inadequate hardware.

(And how many people actually measure and calibrate their display
systems?  And how many people actually measure and adjust for variations
in ambient lighting?  I have seen enough real world measurements to know
that these steps are needed.)

_______________________________________________
Render mailing list
[EMAIL PROTECTED]
http://XFree86.Org/mailman/listinfo/render

Reply via email to