New issue
Advanced search Search tips
Note: Color blocks (like █ or █) mean that a user may not be available. Tooltip shows the reason.

Issue 597073 link

Starred by 1 user
Status: Available
Owner: ----
Cc:
█ klh@google.com
tmathmeyer@chromium.org
dalecur...@chromium.org
wolenetz@chromium.org
foolip@chromium.org
█ renganat...@chromium.org
Components:
EstimatedDays: ----
NextAction: ----
OS: ----
Pri: 3
Type: Feature



Sign in to add a comment

Find a way, if possible, to support ISO-BMFF CEA 608/708 embedded text track parsing for MSE

Project Member Reported by wolenetz@chromium.org, Mar 22 2016 
This bug tracks getting some better understanding and potentially a solution for the problem of how to understand if ISO-BMFF stream contains CEA 608/708 embedded text track information, in the MSE structure.

On first look, it appears the embedding is within the AVC stream, as part of a video track, and the embedding is not apparent from just an MSE ISO BMFF initialization segment. Therefore, supporting such embedded CEA 608/708 text information in a compliant MSE implementation seems in conflict with current MSE ISO-BMFF bytestream spec.

I've filed an MSE spec bug to get some other UA implementor help as well as help from @silviapfeiffer at https://github.com/w3c/media-source/issues/58

Comment 1 by wolenetz@chromium.org, Mar 24 2016

Labels: MSEscrubbed

Comment 2 by yini...@chromium.org, Mar 28 2016

Owner: phil...@opera.com
Status: Assigned (was: Untriaged)
text track related.

Comment 3 by phil...@opera.com, Mar 29 2016

Cc: phil...@opera.com
Owner: wolenetz@chromium.org
Assigning to wolenetz@ as the reporter. I think the hard bit here isn't extracting the 608/708 data, but actually representing it using the TextTrackCue interface and rendering it.

Comment 4 by wolenetz@chromium.org, Mar 29 2016 

Philip, that's a great point. I'd appreciate any further elaboration you can provide on the difficulties buffering and rendering 608/708 data.

Comment 5 by phil...@opera.com, Apr 1 2016

Cc: klh@google.com
I don't know every detail of the format, but got some basic command codes implemented in https://github.com/foolip/608.js to have at least an overview. At least for 608 (708 is more advanced) I think that merely parsing and rendering it isn't an insurmountable difficulty, although you would need a spec that describes parsing in enough detail to get interoperable implementations. If memory serves the official existing spec doesn't describe error handling, it's just not written with the web platform in mind, as it didn't exist.

CC Ken Harrenstien, who is an expert in this area, and who recommended The Closed Captioning Handbook by Gary D. Robson to me. That book and the spec is all I've read, I don't have any *real* hands-on experience with 608/708.

Comment 6 by foolip@chromium.org, Nov 11 2016

Cc: foolip@chromium.org

Comment 7 by foolip@chromium.org, Nov 11 2016

Cc: -phil...@opera.com

Comment 8 by wolenetz@chromium.org, Mar 27 2018

Cc: tmathmeyer@chromium.org dalecur...@chromium.org wolenetz@chromium.org
Owner: ----
Status: Available (was: Assigned)
Impl side, this may become moot soon. See bug 826419.
►

Sign in to add a comment