- From: David Ronca <dronca@netflix.com>
- Date: Wed, 30 Jan 2013 17:31:53 +0000
- To: "public-tt@w3.org" <public-tt@w3.org>
> That's not at all obvious to me. Firstly, why would this be different that <tt tts:extent>? <tt tts:extent> specifies the display area. A client may ignore the pixel address and only use the computed DAR, but that would be a client-specific behavior that is undefined in the spec. > You say you contract out captioning and receive back TTML >that contain <tt tts:extent='320px 240px'>, is that correct? Yes, this would be typical. > So this is 4:3, yes? It also means that the dimensions of the root container are 320x240 window, does it not? > What would it mean if you had <tt tts:extent='320px 240px' > and ttp:displayAspectRatio='16 9'>? Would not happen. Also, we would strip out the <tt tts:extent='320px 240px' > before we deploy the CDN, because our display size varies by stream. > If you are dynamically generating different media streams to satisfy > different clients, why don't you dynamically transform a master TTML > document into a display specific document that is targeted to work with the > media stream format? We do not dynamically generate streams. We generate a set of streams and various clients will pull streams based on their dynamic bandwidth profile, which changes frequently. We would like to generate one caption asset that can be presented correctly on any device from mobile to a 4K TV.
Received on Wednesday, 30 January 2013 17:32:21 UTC