[Ogg a11y] another <video> tag srt, cmml example

Conrad Parker conrad at metadecks.org
Wed Dec 10 17:05:45 PST 2008


2008/12/11 Chris Double <chris.double at double.co.nz>:
> On Thu, Dec 11, 2008 at 12:40 PM, Michael Dale <mdale at wikimedia.org> wrote:
>>> Seeking works fine in firefox as long as the web server supports byte
>>> range requests.
>>>
>>
>> hmm... but what if your viewing an oggz chopped segment?
>
> If oggz-chop supports byte ranges then you will be able to seek within
> that segment. You could serve the entire file and catch the metadata
> loaded event to seek to the time offset you want. As long as your
> server supports byte ranges this should 'just work'. But as you say
> later in your email, this is problematic with a time bar that seems to
> never move.
>
> If you want to go the oggz-chop route then you'll need to implement
> seeking, etc yourself by changing changing the video src with the
> appropriate query parameters, as I imagine you are currently doing. Or
> document the protocol you want to use (X-Content-Duration, etc) so it
> can be implemented in the browser. If someone can knock up a document
> for 'protocol for servers serving partial ogg content', containing the
> headers, URL format, etc expected I can do something about that.

aye, I can do that.

>> If we have that info it lets us update the user interface so users know if
>> they should expect delays in the http seek vs the local (fast seek)
>> ..(youtube does this for example).
>
> Are you talking about the 'buffered' attribute in the HTML 5 video
> spec? That's not currently implemented. See bug 462957. As 'buffered'
> returns TimeRange objects we need to be able to have not just buffered
> the data, but also examined it to compute the time value. In
> liboggplay we only get the time value of data when it is decoded so
> can't really return this information without fulling decoding the
> data.

would it be better to be able to get the timestamps before decoding?
oggz can do that, we'd just need to expose it in oggplay.

K.


More information about the Accessibility mailing list