[Ogg a11y] OggText SRT format (Re: <video> + srt)

[Silvia Pfeiffer]

On Mon, Dec 8, 2008 at 1:30 PM, Henri Sivonen <hsivonen at iki.fi> wrote:
> On Dec 7, 2008, at 16:38, Silvia Pfeiffer wrote:
>
>> I would suggest to support the little formatting tags that the doom9
>> page describes, even though most srt files don't use them.
>
> Supporting HTML-like formatting tags is a can of worms: In addition to the
> tags, you need a mechanism for escaping '<'. Even with an escaping
> mechanism, there'd be the problem that a given string would mean a different
> sequence of rendered characters in normal SRT and in the extended flavor.
>
> RSS <title> is a case study of extending a datum that originally contained
> plain text to contain markup. Not pretty. The result is that there are
> strings that cannot be represented reliably.
>
> Further, with HTML-like tags, you'd need to define a parsing algorithm for
> them. The little tags would open up a lot of complexity in the direction of
> DFXP.
>
> I would keep the text payload of SRT as plain text (with Unicode bidi
> controls explicitly allowed).

A good argument - so we'll just do plain text srt inside Ogg and thus
also inside HTML5.
srt with markup will have to be stripped then by a good encoder.

BTW: sub (the VobSub) format is also very similar, e.g.

{0}{80}[Buena Vista International presents]
{110}{195}[A Claussen + Wˆbke Production]
{212}{295}[A Hans-Christian Schmid Film]|[Off-screen:] -"...as bound by duty
{305}{420}"of office, I ask you to acknowledge this vote."
{440}{500}Madam President, I accept the vote.
{505}{594}[TV:] "After long anticipation in Bonn
{599}{655}Richard von Weiz‰cker won the election.
{660}{744}He won more votes than his prior elections.
{749}{844}It's his greatest success since Theodor Heuss.

A potential oggsrt encoder could also accept the VobSub format and
convert it to srt (SubRip).

Not trying to make it difficult to implement an encoder - just trying
to explain the possibilities!

Cheers,
Silvia.