Update: ffmpeg patch

[Philip Langdale]

Hi all,

I've successfully ported my mplayer patch to ffmpeg and posted it here:

http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/2010-December/102202.html

In the last few days, I've been able to improve some aspects of the
implementation,
but problems still exist:

1) I've now got mpeg2 and h.264 mbaff handled properly, but I'm not
happy with the
frame flags I'm seeing.

mbaff shows up as field pairs with the UNKNOWN_SRC and TOP_FIELD flags set.
I'd hope to have seen the FIELDPAIR flag for obvious reasons.

I've got a PAFF sample that shows up as separate fields but also with
UNKNOWN_SRC
set. I have no way to distinguish the two cases :-/

2) Can't get the DIVX311 mode to work with any of the samples I've found.

3) I've made big strides in controlling the initial feeding of the
pipeline by forcing a 50ms
gap between calls to ProcInput. No idea how optimal this really is,
but it seems right and
now my pipeline length is under control, varying from 7 to 20 frames.

4) When I seek the stream - and flush it, the pipeline refill is much
less controlled and I
end up with a lot more frames going in before I get any out. I don't
know how much of
this ffmpeg/mplayer's fault or something else.

Any advice on these would be much appreciated!

--phil

[Naren (Narendra) Sankar]

I would like to look at the interlacing issues from the FW perspective to make sure we are doing the right thing.

Can you point me to the MBAFF and PAFF samples you are using for your testing?

Thanks

Naren Sankar (+1 408 218 6327)
Architect/PLM
Broadcom Corp.

Hi all,

http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/2010-December/102202.html

--phil

--
To post to this group, send email to
crystalhd-...@googlegroups.com
For more options, visit this group at
http://groups.google.com/group/crystalhd-development?hl=en

[Philip Langdale]

On Mon, Jan 3, 2011 at 3:53 PM, Naren (Narendra) Sankar
<nsa...@broadcom.com> wrote:
> I would like to look at the interlacing issues from the FW perspective to make sure we are doing the right thing.
>
> Can you point me to the MBAFF and PAFF samples you are using for your testing?

Here's a PAFF sample:

http://samples.mplayerhq.hu/V-codecs/h264/PAFF/tv_cut.mkv

and I made my own MBAFF sample from a DVD:

http://intr.overt.org/misc/mbaff.mkv

--phil