[Libav-user] Problem with h264_videotoolbox and AV_PIX_FMT_VIDEOTOOLBOX

Richard Kern kernrj at gmail.com
Fri Feb 17 23:23:50 EET 2017

On February 17, 2017 at 4:18:38 PM, Steve Green (steve.green at flosports.tv)

While trying to debug a problem using YUV420P with h264_videotoolbox, I
came across AV_PIX_FMT_VIDEOTOOLBOX in the encoder code. Depending on what
I discover here, I might be back in another thread about a YUV issue.

It appears that one can specify AV_PIX_FMT_VIDEOTOOLBOX and then provide
CVPixelBuffers in data[3]. The issue Im having is that avcodec_open2()
finds its way to vtenc_send_frame(). I don’t understand what is going on at
this level but the internally constructed AVFrame will not have anything in
data[3] and the following code in create_cv_pixel_buffer() fails on the
second av_assert0().

> if (avctx->pix_fmt == AV_PIX_FMT_VIDEOTOOLBOX) {
> av_assert0(frame->format == AV_PIX_FMT_VIDEOTOOLBOX);
> *cv_img = (CVPixelBufferRef)frame->data[3];
> av_assert0(*cv_img);
> CFRetain(*cv_img);
> return 0;
> }

Is there something that I need to do to make this work?

That’s a bug. I’ll push a fix shortly.


Libav-user mailing list
Libav-user at ffmpeg.org
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://ffmpeg.org/pipermail/libav-user/attachments/20170217/4dab15f7/attachment.html>

More information about the Libav-user mailing list