Re: Re: TTML and aspect ratio

On Wed, Jan 30, 2013 at 9:20 PM, David Ronca <dronca@netflix.com> wrote:

> >we have issue 179 related to this namely that :
>
> >"we need to clarify the pixel behaviour; in particular we should explain
> that where extent is used on the tt element this effectively *defines*
> >the size of a pixel in the sense that the root area extent is still
> mapped to a video overlay and divided into 'logical pixels'. If such an
> extent is
> >not defined on the tt element, then the px measure should probably not be
> used (or even expressly forbidden)"
>
> > In this interpretation it does not imply that the extent covers some
> specific hardware pixels, but that the region is to be counted as having x
> by y
> > pixels; so that the px measure when used within the TTML document has
> some meaning independent of the hardware. If this is coupled with square
> > aspect ratio pixels (the default), then by definition the root would
> have the specified aspect ratio.
>
> Let see if I am understanding this.  If I have a 1080p video display, and
> 1080p content, <tt tts:extent="640px,480px"> does not mean to display
> captions in a 640x480 window, it means to divide the 1920x1080 pixels into
> 640x480 logical pixels for mapping captions?


No, TTML specifies no requirement that a presentation processor align the
root container region with the related media, does not require that pixels
in the TTML document coordinate space align with pixels in the related
media, does not specify where the origin of the root container region is,
etc.

All of these things are determined by the "authoring or presentation
processor context". In other words, this was declared out of scope when
TTML was designed.

Further, by using <tt ttp:pixelAspectRatio="1,1", the client would not
> stretch the caption window horizontally, but would work within a 4:3 region
> of the video window and the logical pixels would map to 1440x1080 physical
> pixels?
>

If tts:pixelAspectRatio="1 1", all that says is the author has declared
that their model of pixels corresponds to a 1:1 PAR. It says nothing about
where these pixels are displayed or whether they are to be scaled w.r.t.
the related media object (if there is one -- there doesn't have to be).