[MPlayer-dev-eng] [RFC] pre8: regressions, critical bugs and activities

[Giacomo Comes]

On Mon, Mar 20, 2006 at 06:52:11PM +0100, Diego Biurrun wrote:
> On Mon, Mar 20, 2006 at 11:19:59AM -0400, Giacomo Comes wrote:
> > On Mon, Mar 13, 2006 at 10:42:37PM +0100, Roberto Togni wrote:
> > [...]
> > > 
> > > Critical bugs:
> > > - none that i know (remaining Coverity reports are mainly small memleaks
> > > and some use of uninitialized variables in error cases, i don't consider
> > > them critical for release)
> > 
> > There is an annoing bug still there, but probably not critical:
> > if you encode 6 channels ac3 audio with libavcodec the order of the encoded 
> > channels is wrong. In order to have a correct channels order it is necessary to
> > add -af channels=6:6:0:0:4:1:1:2:2:3:3:4:5:5 (the problem exist only with 
> > 6 and 5 channels encoding, it works correctly with 4 channels).
> > I have tryed to find where the problem is but got lost.
> > Does anybody have a clue were to look? Or, even better, how to fix it?
> 
> Maybe related to the following bug?
> 
> http://bugzilla.mplayerhq.hu/show_bug.cgi?id=99

No. The bug reported there is about dts audio always downmixed to two channels.
More than a bug it can be considered a missing features (dts 6 channels 
support) and it concerns MPlayer.

What I'm talking about is wrong channels order with ac3 encoding (MEncoder 
issue).

Giacomo