Preliminary support of VP9 HW encoder on Android.

Preliminary support of VP9 HW encoder on Android.

Not fully tested yet. Verified in test loopback application
with fake VP9 codec factory.
Assume that encoder generates bitstream in non flexible mode with
one temporal and one spatial layers.

R=magjed@webrtc.org

Committed: https://crrev.com/ad948c42a1fd29bf22205ded2a175a967748abe4
Cr-Commit-Position: refs/heads/master@{#10695}

[AlexG, 2015-11-16 22:30:17]

glaznev@webrtc.org changed reviewers:
+ magjed@webrtc.org

[AlexG, 2015-11-16 22:30:17]

PTAL

[AlexG, 2015-11-16 22:30:23]

The CQ bit was checked by glaznev@webrtc.org to run a CQ dry run

[commit-bot: I haz the power, 2015-11-16 22:30:26]

Dry run: CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1451953002/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1451953002/20001

[commit-bot: I haz the power, 2015-11-16 23:20:26]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-11-16 23:20:26]

Dry run: This issue passed the CQ dry run.

[magjed_webrtc, 2015-11-18 13:28:24]

lgtm

https://codereview.webrtc.org/1451953002/diff/20001/talk/app/webrtc/java/jni/...
File talk/app/webrtc/java/jni/androidmediaencoder_jni.cc (right):

https://codereview.webrtc.org/1451953002/diff/20001/talk/app/webrtc/java/jni/...
talk/app/webrtc/java/jni/androidmediaencoder_jni.cc:320: scale_ =
webrtc::field_trial::FindFullName(
I would prefer 
scale_ = codecType_ != kVideoCodecVP9 && webrtc::field_trial::FindFullName(...)
== "Enabled";

https://codereview.webrtc.org/1451953002/diff/20001/talk/app/webrtc/java/jni/...
talk/app/webrtc/java/jni/androidmediaencoder_jni.cc:473: tl0_pic_idx_ =
static_cast<uint8_t>(rand());
I'm curious, do you know where srand() is called?

https://codereview.webrtc.org/1451953002/diff/20001/talk/app/webrtc/java/jni/...
talk/app/webrtc/java/jni/androidmediaencoder_jni.cc:888: if (codecType_ ==
kVideoCodecVP8 && scale_) {
Revert this change, and only check |scale_|.

[AlexG, 2015-11-18 19:03:51]

https://codereview.webrtc.org/1451953002/diff/20001/talk/app/webrtc/java/jni/...
File talk/app/webrtc/java/jni/androidmediaencoder_jni.cc (right):

https://codereview.webrtc.org/1451953002/diff/20001/talk/app/webrtc/java/jni/...
talk/app/webrtc/java/jni/androidmediaencoder_jni.cc:320: scale_ =
webrtc::field_trial::FindFullName(
On 2015/11/18 13:28:24, magjed_webrtc wrote:
> I would prefer 
> scale_ = codecType_ != kVideoCodecVP9 &&
webrtc::field_trial::FindFullName(...)
> == "Enabled";

Done.

https://codereview.webrtc.org/1451953002/diff/20001/talk/app/webrtc/java/jni/...
talk/app/webrtc/java/jni/androidmediaencoder_jni.cc:473: tl0_pic_idx_ =
static_cast<uint8_t>(rand());
On 2015/11/18 13:28:24, magjed_webrtc wrote:
> I'm curious, do you know where srand() is called?

Done. Added srand() to ctor

https://codereview.webrtc.org/1451953002/diff/20001/talk/app/webrtc/java/jni/...
talk/app/webrtc/java/jni/androidmediaencoder_jni.cc:888: if (codecType_ ==
kVideoCodecVP8 && scale_) {
On 2015/11/18 13:28:24, magjed_webrtc wrote:
> Revert this change, and only check |scale_|.

I think it is a little cleaner to keep this check - when we call vp8 specific QP
parsing we should also check that VP8 is selected. This part of code should not
assume that scale_ is always false for VP9

[AlexG, 2015-11-18 19:05:08]

The CQ bit was checked by glaznev@webrtc.org

[AlexG, 2015-11-18 19:05:09]

The patchset sent to the CQ was uploaded after l-g-t-m from magjed@webrtc.org
Link to the patchset: https://codereview.webrtc.org/1451953002/#ps40001 (title:
"Address comments")

[commit-bot: I haz the power, 2015-11-18 19:05:17]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1451953002/40001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1451953002/40001

[commit-bot: I haz the power, 2015-11-18 21:05:28]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2015-11-18 21:05:29]

Try jobs failed on following builders:
  android_arm64_rel on tryserver.webrtc (JOB_TIMED_OUT, no build URL)

[AlexG, 2015-11-18 21:06:57]

Committed patchset #3 (id:40001) manually as
ad948c42a1fd29bf22205ded2a175a967748abe4 (presubmit successful).

[commit-bot: I haz the power, 2015-11-18 21:06:59]

Patchset 3 (id:??) landed as
https://crrev.com/ad948c42a1fd29bf22205ded2a175a967748abe4
Cr-Commit-Position: refs/heads/master@{#10695}

[magjed_webrtc, 2015-11-19 08:10:03]

https://codereview.webrtc.org/1451953002/diff/20001/talk/app/webrtc/java/jni/...
File talk/app/webrtc/java/jni/androidmediaencoder_jni.cc (right):

https://codereview.webrtc.org/1451953002/diff/20001/talk/app/webrtc/java/jni/...
talk/app/webrtc/java/jni/androidmediaencoder_jni.cc:888: if (codecType_ ==
kVideoCodecVP8 && scale_) {
On 2015/11/18 19:03:51, AlexG wrote:
> On 2015/11/18 13:28:24, magjed_webrtc wrote:
> > Revert this change, and only check |scale_|.
> 
> I think it is a little cleaner to keep this check - when we call vp8 specific
QP
> parsing we should also check that VP8 is selected. This part of code should
not
> assume that scale_ is always false for VP9

I see, I didn't realize it was vp8 specific scaling.