Re: [MSE] Ad Insertion (was: [MSE] Timestamp offset mechanism)

On Wed, Jul 11, 2012 at 8:04 AM, Duncan Rowden <Duncan.Rowden@bbc.co.uk> wrote:
> Kevin S and Steven R mentioned on this point that you could get a
> non-deterministic delay if you use this method, but I don’t think this is a
> problem as long as the content is paused frame-accurately.

I think a frame-accurate stream switch proposal as a separate
specification is a useful idea. However, I think it should not be part
of Media Source, nor should it influence the presence of a timestamp
offset mechanism in MS. YouTube will require timestamp offsets, either
by it being in the API or by rewriting media data from JavaScript, and
we think many other applications will use it too. We need this even if
a frame-accurate switch were part of a standard, because a lot of
platforms which can do Media Source couldn't do frame-accurate
pausing.

(To give an example: on one platform, encrypted media samples
disappear down a trusted path, and the browser gets a single,
session-long opaque texture reference which it can use to composite
the latest video frame into a write-only framebuffer (rather than raw
buffers or a texref per frame). Playback time is retrieved by polling
the crypto processor, which carries an unknown latency. Pause
instructions would also carry an unknown latency. No guarantees of
frame-accurate seeking are available.)

Steve

Received on Thursday, 12 July 2012 20:17:20 UTC