[FFmpeg-devel] [PATCH] s302m: Add Channle Layout for 6 channels

[Michael Niedermayer]

On Tue, Apr 10, 2012 at 08:00:30PM -0700, Thierry Foucu wrote:
> On Fri, Apr 6, 2012 at 7:45 AM, Michael Niedermayer <michaelni at gmx.at>wrote:
> 
> > On Fri, Apr 06, 2012 at 01:08:33PM +0100, Kieran Kunhya wrote:
> > > FWIW all of these channel maps are wrong imo. There is no implicit
> > > channel map in 302M.
> >
> > after a bit of dicussion on IRC it appears the consensus is that these
> > set layouts in s302m.c are neither correct nor wrong but may be
> > convenient to the end user.
> >
> > I did then research this issue further and found that there is possibly
> > a optional channel layout stored in the AES3 stream:
> > (this byte 3 is spread into the channel status bits of AES3 frames
> >  IIUC)
> > Note, i have not checked at all what actual streams store in these
> > bits, so i do not know if this would allow us to set the channel
> > layout more accurately or not
> > do we have a volunteer to implement this ?
> >
> >
> > Byte 3
> >                   bit           7        Multichannel mode
> >                                 0        Undefined multichannel mode
> > (default).
> >                  state          1        Defined multichannel modes.
> > The definition of the remaining bit states depends on the state of bit 7.
> >                  bits      0 to 6 Channel number, when byte 3 bit 7 is 0.
> >                            The channel number is the numeric value of the
> > byte, with bit 0 as the least
> >                 value
> >                            significant bit, plus one.
> > OR,
> >                  bits       4 5 6        Multichannel mode, when byte 3
> > bit 7 is 1.
> >                              0 0 0       Multichannel mode 0. The channel
> > number is defined by bits 0
> >                                          to 3 of this byte.
> >                              1 0 0       Multichannel mode 1. The channel
> > number is defined by bits 0
> >                                          to 3 of this byte.
> >                 states       0 1 0       Multichannel mode 2. The channel
> > number is defined by bits 0
> >                 note:                    to 3 of this byte.
> >                  LSB         1 1 0       Multichannel mode 3. The channel
> > number is defined by bits 0
> >                  first                   to 3 of this byte.
> >                              1 1 1       User-defined multichannel mode.
> > The channel number is
> >                                          defined by bits 0 to 3 of this
> > byte.
> >                            All other states of bits 4 to 6 are reserved
> > and are not to be used until further
> >                            defined.
> >                  bits        0 to 3      Channel number, when byte 3 bit 7
> > is 1
> >                            The channel number is the numeric value of
> > these four bits, with bit 0 as the
> >                 value
> >                            least significant bit, plus one.
> > NOTE 1 The defined multichannel modes identify mappings between channel
> > numbers and function. The
> > standard mappings are under consideration. Some mappings may involve
> > groupings of up to 32 channels by
> > combining two modes.
> > NOTE 2 For compatibility with equipment that is only sensitive to the
> > channel status data in one subframe the
> > channel carried by subframe 2 may indicate the same channel number as
> > channel 1. In that case it is implicit
> > that the second channel has a number one higher than the channel of
> > subframe 1 except in single channel double
> > sampling frequency mode.
> > NOTE 3 When bit 7 is 1 the 4 bit channel number can be mapped to the
> > channel numbering in bits 20 to 23 of
> > the consumer mode channel status defined in IEC 60958-3. In this case
> > channel A of consumer mode maps to
> > channel 2, channel B maps to channel 3 and so on.
> >
> >
> We did check some files containing S302M and we could not find out the
> channel layout using the C bit.

then i guess your original patch is the best that can be done.

thanks

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

I do not agree with what you have to say, but I'll defend to the death your
right to say it. -- Voltaire
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20120411/cf925b1f/attachment.asc>