On Fri, Feb 28, 2003 at 03:04:08PM +0000, Ian Molton wrote: | On Thu, 27 Feb 2003 18:17:33 -0800 | Allen Akin <[EMAIL PROTECTED]> wrote: | | > | > Then there are the arguments for deeper color channels based on the | > need for higher-precision intermediate results -- for transparency, | ^^^^^^^^^^^^ | > antialiasing, multipass rendering, etc. These are compelling. | | no argument there.
Apologies if I've misunderstood your comment, but it looks like you're saying that just because intermediate results might require higher precision than the final displayable colors, you don't want high-precision displayable color buffers. On the contrary, making the displayable color buffer high-precision saves time (no need to copy from a high-precision intermediate buffer to a low-precision displayable buffer) and space (you only need one high-precision buffer rather than one high-precision buffer plus one low-precision buffer). Similar arguments apply in other situations, like dynamically generating textures. Oh, and I should have mentioned high-dynamic-range color representations in the previous note. Once you've seen them, there's no going back to the old way. :-) Allen ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
