There are often times where:
- Speech and sound play simultaneously on separate channels.
- The app queues up another utterance and sound on those channels to be played together.
Today, as soon as one channel finishes its output, it moves on to the next queued command. If the sound is short, for instance, the next queued sound will begin playing while the first utterance is still being spoken. The second utterance will possibly play much later.
It's possible to sync output across channels by listening for the end of output deferreds for all audio in a group and only queueing the next commands when they occur. But this approach gets really ugly for app writers.
JSonic needs a barrier command that accepts an array of channels as parameters. The command prevents any channel from continuing to process its command queue until all of the channels have reached the barrier. A channel stop should trigger all queued barriers as if the channel reached them normally.
There are often times where:
Today, as soon as one channel finishes its output, it moves on to the next queued command. If the sound is short, for instance, the next queued sound will begin playing while the first utterance is still being spoken. The second utterance will possibly play much later.
It's possible to sync output across channels by listening for the end of output deferreds for all audio in a group and only queueing the next commands when they occur. But this approach gets really ugly for app writers.
JSonic needs a barrier command that accepts an array of channels as parameters. The command prevents any channel from continuing to process its command queue until all of the channels have reached the barrier. A channel stop should trigger all queued barriers as if the channel reached them normally.