[FFmpeg-devel] [PATCH] G.729 and G.729D decoders

[Vladimir Voroshilov]

On Sun, Apr 20, 2008 at 10:00 PM, Michael Niedermayer <michaelni at gmx.at> wrote:
> On Sun, Apr 20, 2008 at 08:18:20PM +0700, Vladimir Voroshilov wrote:
>  > Hi, Michael
>  >
>  > On Sun, Apr 20, 2008 at 6:22 PM, Michael Niedermayer <michaelni at gmx.at> wrote:
>  > > On Sun, Apr 20, 2008 at 12:31:44PM +0700, Vladimir Voroshilov wrote:
>  > >  > Hi, Michael.
>  >
>  > [...]
>  >
>  > >  > I hope, lookup tables and build api can be separated.
>  > >  > Otherwise i can't create uncompressed <100k patch.
>  > >
>  > >  Well this patch started out at 77k and then grew to 98k during review
>  > >  now its even with compressed tables over 110k
>  > >  77k is already very large and hard to review
>  > >  Iam simply drawing a line and rejecting it until it returns to <100k
>  > >  It was possible in previous iterations so it should still be. If not
>  > >  which of my change requests caused such size increase?
>  > >  You have added support for g729d thats nice but it has to be
>  > >  split into a seperate patch. I dont know if it also makes sense to
>  > >  split it into several files.
>  >
>  > 1. G.729A -> G.729 added 12k of source code (due to huge long-term
>  > filter, large tilt-compensation filter and additional lookup tables
>  > for them)
>  > 2. G.729D support added additionally about 8k
>  >
>  > Here is 93k size G.729-only patch.
>  > I don't know what else can be removed from it.
>
>  your first patch was 77k base64 encoded this is 122k base64 encoded
>
>  I will try to review it but considering the size this will take time
>  This could easily take a year until it reaches svn ...

:(

>  The bigger the patch the more iterations it needs and the longer
>  each review-fix-resend iteration will take. Also the bigger the more
>  exhausted one gets during the review and thus the more things are missed.
>  Which will then need more iterations.
>
>
>  So i strongly suggest that you find a way to split it.

Here is smaller patch (but now it depends on acelp_*, see separate thread)

-- 
Regards,
Vladimir Voroshilov mailto:voroshil at gmail.com
JID: voroshil at gmail.com, voroshil at jabber.ru
ICQ: 95587719
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: g729dec_18.diff
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20080422/3cc07dfa/attachment.txt>