On Wed, Sep 21, 2022 at 05:12:52AM +0000, Lynne wrote:
> ffmpeg | branch: master | Lynne <dev@lynne.ee> | Mon Sep 19 23:48:53 2022 +0200| [3ade6a8644ab519fcd7caa7ef457dd406162bc14] | committer: Lynne
> 
> x86/lpc: implement a new Welch windowing function
> 
> Old one was written with the assumption only even inputs would be given.
> This very messy replacement supports even and odd inputs, and supports
> AVX2 for extra speed. The buffers given are usually quite big (4k samples),
> so the speedup is worth it.
> The new SSE version is still faster than the old inline asm version by 33%.
> 
> Also checkasm is provided to make sure this monstrosity works.
> 
> This fixes some FATE tests.
> 
> > http://git.videolan.org/gitweb.cgi/ffmpeg.git/?a=commit;h=3ade6a8644ab519fcd7caa7ef457dd406162bc14
> ---

This seems to "break" one flac case

./ffmpeg -i asynth-44100-2-2.wav -bitexact -frame_size 4095  -max_prediction_order 16 test-old.flac

The file size becomes worse by alot:

-rw-r----- 1 michael michael 379149 Okt  4 16:45 test-new.flac
-rw-r----- 1 michael michael 235482 Okt  4 16:46 test-old.flac


[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

While the State exists there can be no freedom; when there is freedom there
will be no State. -- Vladimir Lenin