[FFmpeg-devel] [RFC] lavc/ffmpeg sample_fmt implementation

[Michael Niedermayer]

On Fri, Aug 01, 2008 at 12:02:12AM +1000, pross at xvid.org wrote:
> On Wed, Jul 30, 2008 at 10:09:02PM +0200, Michael Niedermayer wrote:
> > On Mon, Jul 28, 2008 at 10:23:29PM +1000, pross at xvid.org wrote:
> > > On Sun, Jul 27, 2008 at 11:00:05PM +0200, Michael Niedermayer wrote:
> > > > On Sun, Jul 27, 2008 at 06:17:48PM +1000, pross at xvid.org wrote:
> > > > > On Sat, Jul 26, 2008 at 01:08:09AM +1000, pross at xvid.org wrote:
> > > > > > Hi.
> > > > > > 
> > > > > > This patch adds sample_fmt conversion support to lavc (and ffmpeg).
> > > > > 
> > > > > Round two patches enclosed.
> > > > > 
> > > 
> > > Round three.
> > > 
> 
> > > Index: libavcodec/Makefile
> > > ===================================================================
> > > --- libavcodec/Makefile	(revision 14454)
> > > +++ libavcodec/Makefile	(working copy)
> > > @@ -21,7 +21,7 @@
> > >         utils.o \
> > >  
> > >  
> > > -HEADERS = avcodec.h opt.h
> > > +HEADERS = avcodec.h opt.h audioconvert.h
> > 
> > this is not ok, yet
> > Id like to wait with making the API public as much as possible to avoid
> > ending up with API breakage issues if we decide that we have to change
> > something. [that is it should all be working and have been tested before]
> 
> Rgr that.
> 
> Related matter: When codecs output/input sample formats other than
> SAMPLE_FMT_S16, existing apps that expect the avcodec_decode_audio2() and
> avcodec_encode_audio() functions to handle shorts ints will break.

Yes, but they will need to be fixed eventually anyway and sample_fmt is not
a new field, it was there since a long time. IMHO applicaions that ignore its
value are already broken.

I really do not want to spend too much time thinking about what happens with
a few applications with some codecs and latest lavc for the period of a week.
Any maintained application will be fixed within that time, users dont need
to use latest lavc with applications ignoreing sample_fmt for that week.

Also its no API or ABI breakage, its rather applications ignoring it because
it happened to be always NONE/S16

[...]

-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

Everything should be made as simple as possible, but not simpler.
-- Albert Einstein
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL: <http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/attachments/20080731/57fe0a7b/attachment.pgp>