[Ffmpeg-devel] ffh264 & coded_frame
Mon Jan 29 14:36:58 CET 2007
On Thu, 2007-01-25 at 16:28 +0100, Luca Abeni wrote:
> I finally found some time to play with ffh264 again, and I found out
> that it sets avctx->coded_frame to NULL.
> Since ffmpeg.c contains code for supporting this situation, I assume
> that avctx->coded_frame == NULL is valid. But this makes impossible to
> know the frame quality, the frame type, and to recognize key frames.
> So, I tried to add support for coded_frame... I came up with the
> attached patch, but I do not know if it is correct (in particular, is it
> ok to set coded_frame = &t->po?).
> After applying this patch ffmpeg shows a reasonable "q=" value while
> encoding (I hope it is correct).
It seems to be correct, as I compared the shown q with the QP set in the
encoder by the bitrate controller, and they are identical.
> I am not setting coded_frame->pts yet,
> but I expect that something like avctx->coded_frame->pts = ((AVFrame
> *)data)->pts should work...
> Finally, when preparing this patch I had some doubts about t->po: if I
> understand it correctly, t->po contains the encoded frame... Is this
> right? If yes, why is it allocated with
> "avpicture_alloc(&t->po, avctx->pix_fmt, width, height);", implying that
> it contains non-compressed video? Also, why not directly allocating
> t->po_data0 instead of allocating t->po and doing t->po_data0 =
> (uint8_t *)t->po.data;?
> Sorry for the dumb question; I'm just trying to get acquited with the
Thanks for the patch!!!
With friendly regards,
Public key: http://www.issaris.org/pi.key
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Size: 189 bytes
Desc: This is a digitally signed message part
More information about the ffmpeg-devel