Switch to using avcodec_decode_audio4, avcodec_alloc_context3.

Switch to using avcodec_decode_audio4, avcodec_alloc_context3.

Allows us to remove another patch from FFmpeg relating to using
deprecated features!  We're now using the latest and greatest!

FFmpeg side changes here,

https://gerrit.chromium.org/gerrit/24823

BUG=112673
TEST=ffmpeg_regression_tests, webaudio tests.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=141524

[DaleCurtis, 2012-06-08 00:40:29]

rbultje: please review my ffmpeg usage.
scherkus: please review my chrome code!
crogers: fyi.

ffmpeg sigs updated here, https://gerrit.chromium.org/gerrit/24823

will include deps roll for ffmpeg changes in this cl once I've merged in
rbultje's sigs updated from:
https://chromiumcodereview.appspot.com/10542056/

[scherkus (not reviewing), 2012-06-08 00:51:22]

https://chromiumcodereview.appspot.com/10540067/diff/1/media/filters/audio_fi...
File media/filters/audio_file_reader.cc (right):

https://chromiumcodereview.appspot.com/10540067/diff/1/media/filters/audio_fi...
media/filters/audio_file_reader.cc:128: scoped_ptr_malloc<AVFrame,
ScopedPtrAVFree> avframe(avcodec_alloc_frame());
nit: FFVD uses av_frame

https://chromiumcodereview.appspot.com/10540067/diff/1/media/filters/ffmpeg_a...
File media/filters/ffmpeg_audio_decoder.cc (right):

https://chromiumcodereview.appspot.com/10540067/diff/1/media/filters/ffmpeg_a...
media/filters/ffmpeg_audio_decoder.cc:216: decoded_audio_size =
av_samples_get_buffer_size(
the old code looked for decoded_audio_size < 0 -- is that still possible w/ the
new code?

https://chromiumcodereview.appspot.com/10540067/diff/1/media/filters/ffmpeg_a...
media/filters/ffmpeg_audio_decoder.cc:218: codec_context_->sample_fmt, 1);
OOC using the default value "0" doesn't work here?

[DaleCurtis, 2012-06-08 01:00:01]

https://chromiumcodereview.appspot.com/10540067/diff/1/media/filters/audio_fi...
File media/filters/audio_file_reader.cc (right):

https://chromiumcodereview.appspot.com/10540067/diff/1/media/filters/audio_fi...
media/filters/audio_file_reader.cc:128: scoped_ptr_malloc<AVFrame,
ScopedPtrAVFree> avframe(avcodec_alloc_frame());
On 2012/06/08 00:51:22, scherkus wrote:
> nit: FFVD uses av_frame

Done.

https://chromiumcodereview.appspot.com/10540067/diff/1/media/filters/ffmpeg_a...
File media/filters/ffmpeg_audio_decoder.cc (right):

https://chromiumcodereview.appspot.com/10540067/diff/1/media/filters/ffmpeg_a...
media/filters/ffmpeg_audio_decoder.cc:216: decoded_audio_size =
av_samples_get_buffer_size(
On 2012/06/08 00:51:22, scherkus wrote:
> the old code looked for decoded_audio_size < 0 -- is that still possible w/
the
> new code?

Technically that's still happening, it's just that the check is reversed.
frame_decoded => decoded_audio_size > 0, which is what the check on line #223 is
now.

https://chromiumcodereview.appspot.com/10540067/diff/1/media/filters/ffmpeg_a...
media/filters/ffmpeg_audio_decoder.cc:218: codec_context_->sample_fmt, 1);
On 2012/06/08 00:51:22, scherkus wrote:
> OOC using the default value "0" doesn't work here?

Per the docs:
"align: buffer size alignment (0 = default, 1 = no alignment)"

The rest of our code expects size to be accurate w/o any kind of alignment.

[scherkus (not reviewing), 2012-06-08 01:04:17]

LGTM!

[rbultje1, 2012-06-08 14:29:01]

FFmpeg part lgtm also.

https://chromiumcodereview.appspot.com/10540067/diff/7001/media/filters/audio...
File media/filters/audio_file_reader.cc (right):

https://chromiumcodereview.appspot.com/10540067/diff/7001/media/filters/audio...
media/filters/audio_file_reader.cc:174: DLOG(WARNING)
You know when I see this, I cry, because FFmpeg (or Libav) does the exact
inverse of this just before outputting the audio back to Chrome (i.e internally
in pretty much every audio codec on the planet, samples are handled and stored
per-channel).

What a waste of cycles... Fortunately someone is working on fixing that
(upstream), please encourage him. ;-).

[commit-bot: I haz the power, 2012-06-11 20:52:32]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/dalecurtis@chromium.org/10540067/9005

[commit-bot: I haz the power, 2012-06-11 22:07:37]

Change committed as 141524