RE: MPEG-2 TS metadata cue guidelines

>>In that case, I would suggest startTime to be 0 and endTime to be PTS1. This means, the cue represents the segment before the ad and can have a pause-on-exit flag set to true.

The problem with startTime=0 is that during a backward seek, where the UA deletes cues generated inband because they will be regenerated again during forward playback, the UA will not know which cues to delete. If the startTime is set to UA current time then the UA will know to delete any cue where the startTime is greater than the seek-to time. As to consistency, maybe for streams other than MPEG-2 TS  there might be consistency issues (and for those we can adjust the mapping of DataCues appropriately) but for MPEG-2 TS the demux + decoding time is set by the stream encoding so this should give a consistent result across implementations. This is in contrast to the network buffering time which could be large and variable, but the textTracks won't be seen until they are demuxed from the stream in my understanding.  

>><Bob> pause-on-exit would only be set to true for this use case but the UA doesn't know the use case. The UA should set the flag to false and the app would have to set it to true. </Bob>

That is my expectation too as the UA will not know what the payload comprises of. Therefore the application will decode the payload, determine it is an ad, set the pause_on_exit flag to true, start loading a resource (set src=...), register an exit event, tie playing the resource to the firing of exit event.

Regards

Graham


-----Original Message-----
From: Bob Lund [mailto:B.Lund@CableLabs.com] 
Sent: Tuesday, March 25, 2014 6:27 AM
To: Silvia Pfeiffer; Clift, Graham
Cc: Brendan Long; public-inbandtracks@w3.org
Subject: RE: MPEG-2 TS metadata cue guidelines


________________________________________
From: Silvia Pfeiffer [silviapfeiffer1@gmail.com]
Sent: Tuesday, March 25, 2014 2:29 AM
To: Clift, Graham
Cc: Bob Lund; Brendan Long; public-inbandtracks@w3.org
Subject: Re: MPEG-2 TS metadata cue guidelines

On Tue, Mar 25, 2014 at 11:48 AM, Clift, Graham <Graham.Clift@am.sony.com> wrote:
> One other thing to consider on the approach I am suggesting. The 
> pause_on_exit flag can be useful to stop a video stream and start an 
> inserted ad. This suggests that the timing for inserting the ad is 
> tied to the exit event, or endTime not startTime. By using the exit 
> event/endTime tightly synchronized to the video frame clock there is 
> an easy way to implement accurate ad insertion.

In that case, I would suggest startTime to be 0 and endTime to be PTS1. This means, the cue represents the segment before the ad and can have a pause-on-exit flag set to true.

HTH,
Silvia.

<Bob> pause-on-exit would only be set to true for this use case but the UA doesn't know the use case. The UA should set the flag to false and the app would have to set it to true. </Bob>

Received on Tuesday, 25 March 2014 17:40:04 UTC