Re: RE: RE: Re: Re: TTML and aspect ratio

On Mon, Feb 4, 2013 at 1:31 PM, David Ronca <dronca@netflix.com> wrote:

>  > Presumably though, the captions were authored for 4:3****
>
> ** **
>
> This is an incorrect assumption.  We get CEA-608 captions that were
> produced using a 16:9 asset.  The person who did the caption work was
> working in a 4:3 grid within the 16:9 video window.  When we receive and
> process the incoming asset, we will know how to present it.  However, in
> W3C-TT, there is no way to convey that presentation information to a
> client.  This is the root of our issue.
>

I don't follow. If the caption author is working with a 16:9 video asset
and is designing TTML for presentation with this 16:9 asset and is
constraining the (non-root) caption regions to a 4:3 grid within the 16:9
asset, then there is no problem. The author just chooses percentages that
produce the desired results.

Example #1 - Caption Region 1 @ top-left of 4:3 grid centered in 16:9
asset, size of 16 horiz cells by 2 vertical cells (with 32x15 grid on 4:3
area)

<region xml:id="topLeftOf43Area" tts:origin="12.5% 0%" tts:extent="37.5%
13.33%"/>

Example #2 - Caption Region 2 @ bottom-right of 4:3 area centered in 16:9
asset, size of 16 horiz cells by 2 vertical cells (with 32x15 grid on 4:3
area)

<region xml:id="bottomRightOf43Area" tts:origin="50% 86.67%"
tts:extent="37.5% 13.33%"/>

Given the above specifications, no matter what 16:9 video asset is used,
you end up with the exact same centered 4:3 area with two 16 by 2 cell
caption regions, one in the top left and the other in the bottom right of
that 4:3 area.

Verification #1 - 1920 x 1080

Region 1 is origin: 240, 0; extent: 720, 144
Region 2 is origin: 960, 936; extent 720, 144

Verification #2 - 1280 x 720

Region 1 is origin: 160, 0; extent: 480, 96
Region 2 is origin: 640, 624; extent 480, 96

Given this ability, I am still failing to see any problem.