W3C home > Mailing lists > Public > public-tt@w3.org > April 2010

RE: [Fwd: Timed tracks]

From: Sean Hayes <Sean.Hayes@microsoft.com>
Date: Thu, 22 Apr 2010 16:18:20 +0000
To: Philippe Le Hegaret <plh@w3.org>
CC: "public-tt@w3.org" <public-tt@w3.org>
Message-ID: <8DEFC0D8B72E054E97DC307774FE4B911A4F7EFA@DB3EX14MBXC303.europe.corp.microsoft.com>
The CSS of the host page in my opinion should not affect the caption text. The captions are authored in order to go with the video, and generally are owned by the video owner, thus the style of the captions should be delivered with those captions. In some use cases, for example selecting larger fonts or specific colour schemes for low vision users, there is some argument for allowing the UA to allow the user to control some styles in the captions; but I'm not of the opinion that mixing the page style with the caption style is an appropriate way to go about that.

I'm not expecting HTML renderers to render the content, since HTML has no concept of timing. I'm expecting a TTML renderer to render the content. Now, converting TTML to HTML and creating a timing model inside HTML is one implementation technique for creating a TTML renderer, and as you say relatively straightforward, but not necessarily the optimum one.

I will be highly disappointed if the HTML WG ends up essentially recreating SAMI.

-----Original Message-----
From: public-tt-request@w3.org [mailto:public-tt-request@w3.org] On Behalf Of Philippe Le Hegaret
Sent: Thursday, April 22, 2010 5:03 PM
To: Sean Hayes
Cc: public-tt@w3.org
Subject: RE: [Fwd: Timed tracks]

On Thu, 2010-04-22 at 15:42 +0000, Sean Hayes wrote:
> If timed text were being integrated into HTML, then I think you'd have a point.

That's actually one of the use cases being considered but I don't consider it an important one myself.

>  In which case I'd argue that HTML+TIME is probably the most sensible starting point.

The CSS inside the caption shouldn't address the entire page but only the video portion, so the positioning is different.

> However the model seems to be having it be associated externally, like a video, audio or image format. Which, since the captions and HTML are likely authored by different people, and come from different servers, seems like the most sensible model, in such a case how the timed text is structured seems somewhat irrelevant to me.

except that we're expecting HTML rendering engines to display the content and they would rather receive HTML content with timing information. Converting TTML into HTML isn't an issue even if it has its limitation (not all style properties from TTML can be done in CSS yet).
I guess the issue is who is going to do it, the video providers or the web navigators.

Received on Thursday, 22 April 2010 16:19:01 UTC

This archive was generated by hypermail 2.3.1 : Thursday, 5 October 2017 18:24:05 UTC