feat(DASH): Parse and use target latency

[avelad]

Related to #6193
Thanks to @gkatsev

[shaka-bot]

Incremental code coverage: 64.44%

[gkatsev]

Thanks! Is there a reason for not adding the target latency as on option as well? Or just want that to be as part of the updated #6193?

[avelad]

Thanks! Is there a reason for not adding the target latency as on option as well? Or just want that to be as part of the updated #6193?

By rereading again, then min and max stop making sense, because we would always use target if we use liveSync = true.

So there are two options:

• Do this which is the base and do nothing with the other.
• Do this which is the base and deprecate min and max through config and always use target + tolerance.

Since in both options you have to do this PR, I have gone for the simple thing.

But I would like to hear your opinion.

[gkatsev]

It's a bit tricky. From the DASH spec, the only property that's required is target latency. From my understanding, the DASH spec is also a bit stricter in terms of what max and min latency mean, where playback should happen in those areas (potentially, triggering a seek to the target latency?).
I don't have concerns with this PR as is, but it would be worth ironing out how shaka wants to handle these properties and work towards that. This change is a good first step, regardless of the following direction.

[avelad]

It's a bit tricky. From the DASH spec, the only property that's required is target latency. From my understanding, the DASH spec is also a bit stricter in terms of what max and min latency mean, where playback should happen in those areas (potentially, triggering a seek to the target latency?). I don't have concerns with this PR as is, but it would be worth ironing out how shaka wants to handle these properties and work towards that. This change is a good first step, regardless of the following direction.

After this PR I will work on making the configuration simpler, so I will add liveSyncTargetLatency and remove liveSyncMaxLatency and liveSyncMinLatency. Thanks for you point of view