[Ffmpeg-cvslog] CVS: ffmpeg/libavcodec dsputil.c, 1.123, 1.124 dsputil.h, 1.117, 1.118 h264.c, 1.136, 1.137

[Måns Rullgård]

Loren Merritt <lorenm at u.washington.edu> writes:

> On Sat, 18 Jun 2005, [iso-8859-1] M?ns Rullg?rd wrote:
>>
>> I've made some progress, but it still doesn't look quite right.
>> Could you take a look at the pictures in
>> http://inprovide.com/~mru/h264_scaling/, and see if you have any idea
>> what the problem might be?
>
> No clue.

Well, I have one.  I think the DC coefficients are being scaled
incorrectly.  Notice the too light/dark 16x16 blocks that appear.  I
found a few uses of dequant_coeffs that I missed the first time
around, and also read the standard a bit more carefully.  It's a bit
of a pain sometimes to figure out which part of the code maps to
various parts of the standard, the latter not exactly being optimized
the way the code is.

-- 
M?ns Rullg?rd
mru at inprovide.com