Fix pending Metadata timestamp relative to subsample offset

[phhusson]

Since MetadataRenderer queues events until event times are current
position, this leads to metadata event like SCTE-35 not being delivered.

To compare to player timestamp, we need to include subsample offset.

[googlebot]

We found a Contributor License Agreement for you (the sender of this pull request), but were unable to find agreements for all the commit author(s) or Co-authors. If you authored these, maybe you used a different email address in the git commits than was used to sign the CLA (login here to double check)? If these were authored by someone else, then they will need to sign a CLA as well, and confirm that they're okay with these being contributed to Google.
In order to pass this check, please resolve this problem and have the pull request author add another comment and the bot will run again. If the bot doesn't comment, it means it doesn't think anything has changed.

[googlebot]

CLAs look good, thanks!

[ojw28]

I don't think this fix is correct. Could you please provide some reproduction steps and sample content?

[phhusson]

I'm debugging and reproducing this by adding:

• if(pendingMetadataCount > 0 && pendingMetadataTimestamps[pendingMetadataIndex] > positionUs) {

•  android.util.Log.d("SAH", "Delaying metadata sample " + positionUs + " : " + pendingMetadataTimestamps[pendingMetadataIndex]);
  

• }
  in MetadataRender's render function
  And by adding a MetadataOutput to the player, and listening for events.

On my internal test streams, without this fix, pendingMetadataTimestamps[pendingMetadataIndex] is in the order of 600 000s, and is never received.
With it, few seconds. It is de-queued few seconds after reception, and is received by the MetadataOutput.

I'll search for a public scte-35 stream, or see if I can publish mine.

[ojw28]

You can also send your test stream to [email protected]. It's somewhat hard to tell whether the issue is with the media itself, or with ExoPlayer (and if so, where), without being able to take a look. I'm pretty sure we have some SCTE-35 test streams internally that work correctly with the current implementation.

[phhusson]

There is a sample:
https://drive.google.com/open?id=1My5PYYLOHQkUPTKDjZbEqN3wdtWsqA4b

[ojw28]

Thanks for the sample! I'm pretty sure the issue is elsewhere, so I'll go ahead and close this pull request. I've filed an issue here.