I am using streamschedule.smil to reboardcast recordings thru the Amazon Cloudfront. Wowza transrates the high quality recordings to deliver a multi bit-rate HLS feed. The feed plays fine when played in the player, but when the next video in the playlist from the streamschedule starts, the last frame of the previous video hangs and the audio continues on. If I refresh the player or start and stop, the video will comeback until the next clip starts. If I sandwich the clips (clip A, clip B, clip A) and don't refresh the player on clip B, when the following clip A plays again the video starts up again just fine. All the while the audio plays just fine.
I am curious if some data from the recorded clips is coming throw the transrate process, or if there is a slight gap when switching files? I have encoded all the recordings the same thinking that might have been the problem but it doesn't seem to have fixed the issue. I have also tried turning the pass metadata thru off and on.
I have experienced this with the JWplayer and the Videojs player (with a custom flash fallback for HLS). The weird part is if I host the players locally the problem doesn't happen. The remote Videojs player spits this out for what seems like each chunck of the HLS.
If I play the HLS stream in a native player I don't get the issue.
Level 2 Live Playlist parsing finished: reload in 11204 ms
Loading 385 of [383,385],level 2
Loaded 385 of [383,385],level 2 m/M PTS:3772621/3784824
Level 2 Live Playlist parsing finished: reload in 11112 ms
Loading 386 of [384,386],level 2
Loaded 386 of [384,386],level 2 m/M PTS:3784845/3794808
Level 2 Live Playlist parsing finished: reload in 10488 ms
Loading 387 of [385,387],level 2
Loaded 387 of [385,387],level 2 m/M PTS:3794829/3804856
I realize this might be more of a player issue, but I wanted to make sure I understood what was happening on the Wowza side of things first.