ppapi: VideoEncoder: improve accuracy of encoding tick

ppapi: VideoEncoder: improve accuracy of encoding tick

The current VideoEncoder example's scheduling algorithm for encoding frames at
30fps drifts as it schedules new callbacks and as a consequence
records fewer frames than what the frame rate requires.
This change improves the algorithm by storing the last time at which it
encoded a frame and uses this information to predict when the next tick should
happen.

BUG=none
TEST=record 30s of video using the video encoder example, check that the produced stream's length is identical (1 or 2 frames difference)

Committed: https://crrev.com/bfa767842bd01af263ae0681a6748f5458fdfc8d
Cr-Commit-Position: refs/heads/master@{#335147}

[llandwerlin-old, 2015-06-17 15:08:53]

lionel.g.landwerlin@intel.com changed reviewers:
+ bbudge@chromium.org, sbc@chromium.org

[llandwerlin-old, 2015-06-17 15:08:53]

bbudge@chromium.org: Could you review the changes in ppapi/examples ? Thanks.

sbc@chromium.org: Please LGTM the natice_client_sdk/ changes if you agree with
them, this is pretty much a copy of what's in ppapi/examples. Thanks.

[bbudge, 2015-06-17 20:45:56]

lgtm

[binji, 2015-06-17 22:04:55]

On 2015/06/17 at 20:45:56, bbudge wrote:
> lgtm

Can you describe in the CL description what problem this is solving?

[llandwerlin-old, 2015-06-18 15:00:38]

On 2015/06/17 22:04:55, binji wrote:
> On 2015/06/17 at 20:45:56, bbudge wrote:
> > lgtm
> 
> Can you describe in the CL description what problem this is solving?

Done. The problem was reported on the CL adding the example :
https://codereview.chromium.org/937643006/

[binji, 2015-06-18 18:53:43]

On 2015/06/18 at 15:00:38, lionel.g.landwerlin wrote:
> On 2015/06/17 22:04:55, binji wrote:
> > On 2015/06/17 at 20:45:56, bbudge wrote:
> > > lgtm
> > 
> > Can you describe in the CL description what problem this is solving?
> 
> Done. The problem was reported on the CL adding the example :
https://codereview.chromium.org/937643006/

Thanks, lgtm

[Stevefar, 2015-06-18 19:06:12]

On 2015/06/18 15:00:38, llandwerlin wrote:
> On 2015/06/17 22:04:55, binji wrote:
> > On 2015/06/17 at 20:45:56, bbudge wrote:
> > > lgtm
> > 
> > Can you describe in the CL description what problem this is solving?
> 
> Done. The problem was reported on the CL adding the example :
> https://codereview.chromium.org/937643006/

thanks Lionel for looking into this.

Unfortunately the problem I reported before persists for me (encoding VP8 on
Chrome Canary 45.0.2435.0 on 2015 macbook). Although now I get 84 "Cannot get
video frame from video encoder" errors thrown whereas no errors were thrown
before. Any idea why so many frames are being dropped? When frames are dropped
like this, can the previous frame be duplicated rather than the video speeding
up?

Just in case it is helpful, heres the downloaded video:
http://storage.googleapis.com/quinncia-videos/5983083b-3711-469d-b993-e80ffb8...

[llandwerlin-old, 2015-06-18 22:23:34]

The CQ bit was checked by lionel.g.landwerlin@intel.com

[commit-bot: I haz the power, 2015-06-18 22:25:24]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1187193006/1

[commit-bot: I haz the power, 2015-06-18 22:25:27]

The author lionel.g.landwerlin@intel.com has not signed Google Contributor
License Agreement. Please visit https://cla.developers.google.com to sign and
manage CLA.

[llandwerlin-old, 2015-06-18 22:34:34]

On 2015/06/18 19:06:12, Stevefar wrote:
> On 2015/06/18 15:00:38, llandwerlin wrote:
> > On 2015/06/17 22:04:55, binji wrote:
> > > On 2015/06/17 at 20:45:56, bbudge wrote:
> > > > lgtm
> > > 
> > > Can you describe in the CL description what problem this is solving?
> > 
> > Done. The problem was reported on the CL adding the example :
> > https://codereview.chromium.org/937643006/
> 
> thanks Lionel for looking into this.
> 
> Unfortunately the problem I reported before persists for me (encoding VP8 on
> Chrome Canary 45.0.2435.0 on 2015 macbook). Although now I get 84 "Cannot get
> video frame from video encoder" errors thrown whereas no errors were thrown
> before. Any idea why so many frames are being dropped? When frames are dropped
> like this, can the previous frame be duplicated rather than the video speeding
> up?
> 
> Just in case it is helpful, heres the downloaded video:
>
http://storage.googleapis.com/quinncia-videos/5983083b-3711-469d-b993-e80ffb8...

Hi Steve,

It looks like the encoder can't keep up with the framerate. This is something
the example doesn't deal with at the moment.
Could you open a bug on https://crbug.com/new ?
I'll look into improving this scenario too (it will require a different
container than the currently used IVF).

[commit-bot: I haz the power, 2015-06-18 22:35:49]

The author lionel.g.landwerlin@intel.com has not signed Google Contributor
License Agreement. Please visit https://cla.developers.google.com to sign and
manage CLA.

[commit-bot: I haz the power, 2015-06-18 22:38:19]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2015-06-18 22:39:21]

Patchset 1 (id:??) landed as
https://crrev.com/bfa767842bd01af263ae0681a6748f5458fdfc8d
Cr-Commit-Position: refs/heads/master@{#335147}