Add QP statistics logging to Android HW encoder.

Add QP statistics logging to Android HW encoder.

- Similar to HW decoder log timing for first 5 frames
of HW encoder input and output to track correct encoder
start.
- Increment input timestamp when frame is dropped.
- Better calculate average encoding time.
- Use different tags for encoder and decoder modules.

BUG=b/26504665

Committed: https://crrev.com/f4decb5469a6aebe8d755b81c04d1fc044f290e3
Cr-Commit-Position: refs/heads/master@{#11277}

[AlexG, 2016-01-15 01:12:41]

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

[AlexG, 2016-01-15 01:12:41]

PTAL

[perkj_webrtc, 2016-01-15 07:33:08]

perkj@webrtc.org changed reviewers:
+ pbos@webrtc.org

[perkj_webrtc, 2016-01-15 07:33:09]

pbos - can you take a look at my comment in this cl?

If pbos is ok, lgtm.

https://codereview.webrtc.org/1587943004/diff/1/talk/app/webrtc/java/jni/andr...
File talk/app/webrtc/java/jni/androidmediaencoder_jni.cc (right):

https://codereview.webrtc.org/1587943004/diff/1/talk/app/webrtc/java/jni/andr...
talk/app/webrtc/java/jni/androidmediaencoder_jni.cc:623: OnDroppedFrame();
OnDropped frame now happens before quality_scaler_.OnEncodeFrame and thus, if
the frame is dropped quality_scaler_.OnEncodeFrame is not called. Is the scaler
logic ok with this ? Was it wrong before?

https://codereview.webrtc.org/1587943004/diff/1/talk/app/webrtc/java/jni/andr...
talk/app/webrtc/java/jni/androidmediaencoder_jni.cc:915: key_frame << ". Size: "
<< payload_size << ". TS: " <<
Do you want to change key_frame ? "true": "false" ;

[pbos-webrtc, 2016-01-15 10:54:04]

https://codereview.webrtc.org/1587943004/diff/1/talk/app/webrtc/java/jni/andr...
File talk/app/webrtc/java/jni/androidmediaencoder_jni.cc (right):

https://codereview.webrtc.org/1587943004/diff/1/talk/app/webrtc/java/jni/andr...
talk/app/webrtc/java/jni/androidmediaencoder_jni.cc:621: frames_dropped_++;
Can we move frames_dropped_++ into OnDroppedFrame(), and move the comment about
reporting it to quality scaler in there?

https://codereview.webrtc.org/1587943004/diff/1/talk/app/webrtc/java/jni/andr...
talk/app/webrtc/java/jni/androidmediaencoder_jni.cc:623: OnDroppedFrame();
On 2016/01/15 07:33:08, perkj_webrtc wrote:
> OnDropped frame now happens before quality_scaler_.OnEncodeFrame and thus, if
> the frame is dropped quality_scaler_.OnEncodeFrame is not called. Is the
scaler
> logic ok with this ? Was it wrong before?

I think this is fine from glancing at it. OnEncodeFrame computes which
resolution should be used for the frame, so it doesn't matter if it's not called
for a frame that's dropped. Maybe this should be named better.

[AlexG, 2016-01-15 20:48:11]

https://codereview.webrtc.org/1587943004/diff/1/talk/app/webrtc/java/jni/andr...
File talk/app/webrtc/java/jni/androidmediaencoder_jni.cc (right):

https://codereview.webrtc.org/1587943004/diff/1/talk/app/webrtc/java/jni/andr...
talk/app/webrtc/java/jni/androidmediaencoder_jni.cc:621: frames_dropped_++;
On 2016/01/15 10:54:04, pbos-webrtc wrote:
> Can we move frames_dropped_++ into OnDroppedFrame(), and move the comment
about
> reporting it to quality scaler in there?

Done.

https://codereview.webrtc.org/1587943004/diff/1/talk/app/webrtc/java/jni/andr...
talk/app/webrtc/java/jni/androidmediaencoder_jni.cc:623: OnDroppedFrame();
On 2016/01/15 07:33:08, perkj_webrtc wrote:
> OnDropped frame now happens before quality_scaler_.OnEncodeFrame and thus, if
> the frame is dropped quality_scaler_.OnEncodeFrame is not called. Is the
scaler
> logic ok with this ? Was it wrong before?

Acknowledged.

https://codereview.webrtc.org/1587943004/diff/1/talk/app/webrtc/java/jni/andr...
talk/app/webrtc/java/jni/androidmediaencoder_jni.cc:915: key_frame << ". Size: "
<< payload_size << ". TS: " <<
On 2016/01/15 07:33:08, perkj_webrtc wrote:
> Do you want to change key_frame ? "true": "false" ;

I think it will be a bit longer in the log : "false" instead of "0". This log
message is already too long.

[AlexG, 2016-01-15 20:53:26]

The CQ bit was checked by glaznev@webrtc.org

[AlexG, 2016-01-15 20:53:26]

The patchset sent to the CQ was uploaded after l-g-t-m from perkj@webrtc.org
Link to the patchset: https://codereview.webrtc.org/1587943004/#ps20001 (title:
"Address comments")

[commit-bot: I haz the power, 2016-01-15 20:53:32]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1587943004/20001
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1587943004/20001

[commit-bot: I haz the power, 2016-01-15 21:49:25]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2016-01-15 21:49:33]

Description was changed from

==========
Add QP statistics logging to Android HW encoder.

- Similar to HW decoder log timing for first 5 frames
of HW encoder input and output to track correct encoder
start.
- Increment input timestamp when frame is dropped.
- Better calculate average encoding time.
- Use different tags for encoder and decoder modules.


BUG=b/26504665
==========

to

==========
Add QP statistics logging to Android HW encoder.

- Similar to HW decoder log timing for first 5 frames
of HW encoder input and output to track correct encoder
start.
- Increment input timestamp when frame is dropped.
- Better calculate average encoding time.
- Use different tags for encoder and decoder modules.

BUG=b/26504665

Committed: https://crrev.com/f4decb5469a6aebe8d755b81c04d1fc044f290e3
Cr-Commit-Position: refs/heads/master@{#11277}
==========

[commit-bot: I haz the power, 2016-01-15 21:49:34]

Patchset 2 (id:??) landed as
https://crrev.com/f4decb5469a6aebe8d755b81c04d1fc044f290e3
Cr-Commit-Position: refs/heads/master@{#11277}