[FFmpeg-devel] [PATCH] aac: Clip large (legal) PNS/IS values

[Michael Niedermayer]

On Fri, Oct 09, 2009 at 12:38:17PM -0400, Alex Converse wrote:
> On Fri, Oct 9, 2009 at 9:37 AM, Robert Swain <robert.swain at gmail.com> wrote:
> > 2009/10/6 Alex Converse <alex.converse at gmail.com>:
> >> Large intensity stereo and PNS indices are legal. Clip them instead of
> >> erroring out. A magnitude of 100 corresponds to 2^25 so the will most
> >> likely result in clipped output anyway.
> >>
> >> None of the conformance streams fall in the range that need to be clipped.
> >
> > Looks OK to me as long as it isn't arbitrary clipping but rather
> > something needed/allowed within and specified by the specification.
> >
> 
> The clipping is semi-arbitrary. It was designed such that after the
> third patch the ranges are the same as regular scalefactors. However
> these large vales (even when clipped) should cause cause output
> clipping on any streams that aren't purely academic exercises.

IMHO, either
A. it can be in a valid stream -> we should support it
B. it can not be in a valid stream -> error concealment (silence /repeat
   last block / something better) should be applied
cliping seems always wrong (to me) unless mandated by a spec.
switching between A and B, in the sense of "it could be valid but most
likely is not, lets ask the user" can be done by using
AVCodecContext.error_recognition

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

There will always be a question for which you do not know the correct awnser.
-------------- 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/20091010/8f61c2d0/attachment.pgp>