We only do this once for 1024 values, so I going for some accuracy here ...
12 years, 8 months ago
(2012-03-13 13:54:32 UTC)
#1
We only do this once for 1024 values, so I going for some accuracy here at the
cost of some speed. Once we're finished experimenting, some of this could be
turned into actual tables and no calculations would be needed at runtime anyway.
Also added an optional sRGB so we know what that looks like, and cleaned up
which bits of code go with which defines.
On 2012/03/13 13:54:32, bungeman wrote: > We only do this once for 1024 values, so ...
12 years, 8 months ago
(2012-03-13 15:43:33 UTC)
#2
On 2012/03/13 13:54:32, bungeman wrote:
> We only do this once for 1024 values, so I going for some accuracy here at the
> cost of some speed. Once we're finished experimenting, some of this could be
> turned into actual tables and no calculations would be needed at runtime
anyway.
>
> Also added an optional sRGB so we know what that looks like, and cleaned up
> which bits of code go with which defines.
Committed revision r3374.
Rebaselined revision r3375.
Issue 5783099: Use floats for mask table for accuracy.
(Closed)
Created 12 years, 8 months ago by bungeman
Modified 12 years, 8 months ago
Reviewers: reed1
Base URL: http://skia.googlecode.com/svn/trunk/
Comments: 0