[Accessibility] Multimedia framework requirements

[Hynek Hanke]

>  But, I think from an accessibility point of view, what we really need is
> quick start and shut up times, not necessarily low latency, although low
> latency is better.  

Good point. We still need synchronization, not just quick start and stop times,
because in long term, we will want to provide cursor movement when the user
reads a continous text. For example, when the user requests reading a long
Emacs buffer, it would be nice if at least the page turned for the next one
each time the speech gets to the end of the page (e.g. for braille displays),
but many times even a word by word cursor tracking would be useful.

What the TTS engine and it's clients know is the times in the generated
soundfile when certain spots in the original text are reached. From the audio
system we need to know when that part of the soundfile is being played in the
speakers.

I agree with two points: (a) this is not a MUST HAVE but a SHOULD HAVE or even
NICE TO HAVE, (b) it's better not to use the low-latency terminology.

With Regards,
Hynek