I am sending in stream captions in a live stream using TTML.
I think the TimedTree.ComputeTimeIntervals code is truncating the duration that my caption appears based on the end of the chunk time.
For example, if my smooth chunks are 3 seconds long and I have a TTML caption that starts 2 seconds into the chunk with a 4 second duration, the caption only appears for one second. I think the parsing code is limiting the duration by the remaining
time of the chunk the caption appears in.
This seems to me a bug (or enhancement request): a caption would not be able to start near the end of a chunk with a reasonable duration to read it.
Can anyone confirm this is the expected behavior?
And if so, if there a way to format the ttml data in the stream such that the caption has a duration longer than the end of the chunk?