[FFmpeg-devel] [PATCH]HE-AACv1 try 3 (all missing functionality added)
Tue Feb 16 17:07:12 CET 2010
On Tue, Feb 16, 2010 at 5:44 AM, Loren Merritt <lorenm at u.washington.edu> wrote:
> On Mon, 15 Feb 2010, Vitor Sessak wrote:
>> Alex Converse wrote:
>>> Why is ff_cos_tabs positive in quadrant II?
>> Good question. I think I use only the first quadrant. Any one of the FFT
>> authors could clarify it?
> FFT uses only the first quadrant. What comes after that in the table is not
> a continuation, but a reversed copy of the first quadrant to save on SIMD
> shuffle instructions.
Maybe this should be documented in the code somewhere.
More information about the ffmpeg-devel