[MPlayer-dev-eng] [RFC] Support for 8 channel audio

[Jason Tackaberry]

On Fri, 2009-08-14 at 15:58 -0400, Jason Tackaberry wrote:
> I assumed the original author unrolled the loop for performance.  I
> haven't done any benchmarking, again I just basically copied the
> existing 6ch code and extended it for 8ch.
> 
> Is there a preferences to either approach?  Should I update the existing
> 5ch and 6ch functions as well?

Just to recap, with respect to the reorder functions in reorder_ch.c, is
it preferred that:

     1. I update only reorder_copy_8ch() to use a series of 3-byte
        memcpys
     2. #1, but I also modify the existing reorder_copy_5ch() and
        reorder_copy_6ch() to follow suit
     3. Neither, leaving reorder_copy_8ch() as it is (which follows
        precedence in existing code).

My preference is #3, and then after the patch as it is now is reviewed
and approved, I can go back and do some more thorough analysis of any
performance implications to the approaches Reimar suggested.  I can then
submit another patch to clean up reorder_copy_[568]ch().


> > > +    if (ao_data.channels == 6 || ao_data.channels == 5 || ao_data.channels == 8) {
> > 
> > Having this in some kind of order would be nice.
> 
> The conditions /are/ in order: from most likely to least likely to occur. :)
> 
> I'll update to numerical order.

Updated, and attached.  Again, this patch assumes my previous "Fix
incorrect channel ordering" patch is already applied in order to merge
cleanly.

Jason.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: 8-channel-audio.patch
Type: text/x-patch
Size: 18821 bytes
Desc: not available
URL: <http://lists.mplayerhq.hu/pipermail/mplayer-dev-eng/attachments/20090817/80defd5d/attachment.bin>