[Cast] Track audio queued in encoder; account for it in ShouldDropNextFrame().

[Cast] Track audio queued in encoder; account for it in ShouldDropNextFrame().

This change is a prerequisite for transitioning the "max outstanding
frames" decision logic over to a time-based heuristic.  It adds tracking
of the number of audio samples enqueued in the encoder, from which the
needed backlog stats can be determined.

Also, this introduces an abstract method in FrameSender,
GetNumberOfFramesInEncoder(), since this is computed differently for
audio versus video.

BUG=404813
Committed: https://crrev.com/182520b878bee21aad3500391cd51d486e37bf1e
Cr-Commit-Position: refs/heads/master@{#293987}

[miu, 2014-09-04 19:25:13]

miu@chromium.org changed reviewers:
+ hclam@chromium.org

[miu, 2014-09-04 19:25:14]

hclam: PTAL.

[Alpha Left Google, 2014-09-05 20:02:07]

https://codereview.chromium.org/545593002/diff/1/media/cast/sender/audio_send...
File media/cast/sender/audio_sender.cc (right):

https://codereview.chromium.org/545593002/diff/1/media/cast/sender/audio_send...
media/cast/sender/audio_sender.cc:96:
UpdateEncoderBacklogStats(+audio_bus->frames());
What's with the '+'?

https://codereview.chromium.org/545593002/diff/1/media/cast/sender/frame_send...
File media/cast/sender/frame_sender.cc (right):

https://codereview.chromium.org/545593002/diff/1/media/cast/sender/frame_send...
media/cast/sender/frame_sender.cc:186: duration_in_flight >=
target_playout_delay_;
What's the reason of choosing "duration_in_flight >= target_playout_delay_" over
"duration_in_flight + frame_duration >= target_playout_delay_"?

[miu, 2014-09-05 20:38:50]

miu@chromium.org changed reviewers:
+ hubbe@chromium.org

[miu, 2014-09-05 20:40:27]

I will wait until hubbe@'s change (https://codereview.chromium.org/542883004/)
lands, then re-merge, then we can continue reviewing.  In the meantime,
responses to your comments:

https://codereview.chromium.org/545593002/diff/1/media/cast/sender/audio_send...
File media/cast/sender/audio_sender.cc (right):

https://codereview.chromium.org/545593002/diff/1/media/cast/sender/audio_send...
media/cast/sender/audio_sender.cc:96:
UpdateEncoderBacklogStats(+audio_bus->frames());
On 2014/09/05 20:02:06, Alpha wrote:
> What's with the '+'?

Readability thing.  Just a helper to see that this is where the backlog gets
incremented.  The complier treats it as a no-op.

https://codereview.chromium.org/545593002/diff/1/media/cast/sender/frame_send...
File media/cast/sender/frame_sender.cc (right):

https://codereview.chromium.org/545593002/diff/1/media/cast/sender/frame_send...
media/cast/sender/frame_sender.cc:186: duration_in_flight >=
target_playout_delay_;
On 2014/09/05 20:02:07, Alpha wrote:
> What's the reason of choosing "duration_in_flight >= target_playout_delay_"
over
> "duration_in_flight + frame_duration >= target_playout_delay_"?

You can ignore this:

1. hubbe@ has already moved this method in his change, so I will undo these LOC
and merge after his change commits.

2. There's no code change here.  This code method was moved from VideoSender
unchanged.

3. In a later change, I've rewritten this method with better logic.  We should
spend our time discussing that logic, and not this old logic.  ;-)

[Alpha Left Google, 2014-09-05 20:50:42]

Okay. LGTM.

[miu, 2014-09-09 00:38:43]

Patchset #2 (id:20001) has been deleted

[miu, 2014-09-09 17:50:54]

The CQ bit was checked by miu@chromium.org

[commit-bot: I haz the power, 2014-09-09 17:54:02]

CQ is trying da patch. Follow status at
 https://chromium-status.appspot.com/cq/miu@chromium.org/545593002/40001

[commit-bot: I haz the power, 2014-09-09 19:33:49]

Committed patchset #2 (id:40001) as 8cd9bc9551edaf7ef7615a4c8dc4aa102e9d5bbc

[commit-bot: I haz the power, 2014-09-10 03:54:57]

Patchset 2 (id:??) landed as
https://crrev.com/182520b878bee21aad3500391cd51d486e37bf1e
Cr-Commit-Position: refs/heads/master@{#293987}