[foms] WebM Manifest
watsonm at netflix.com
Thu Mar 17 14:19:15 PDT 2011
On Mar 17, 2011, at 2:01 PM, Chris Double wrote:
> On Fri, Mar 18, 2011 at 4:59 AM, Mark Watson <watsonm at netflix.com> wrote:
>> To construct byte range requests you need to read and parse this index. I'm
>> data yet.
> workarounds and there is work going on to make it easier. For a
> Firefox example see:
> This reads binary data from Ogg files and parses it. The demo page is:
> I do like Philip's approach of defining a low level API with manifest
> trying different formats easier. It would also make it easier for
> services to experiment and innovate.
We should make a distinction between supporting multiple manifest formats and providing scope for experimentation with rate adaptation algorithms. They are independent things.
The former is easy, but IMO not terribly useful as whether a manifest is coded in one format or another is not terribly important and it's easy to convert if the data models match.
The latter is difficult, as discussed in other mails. It's a great goal but some real research is needed to get something both simple enough to be credible as a standard API and rich enough to actually enable non-trivial experimentation.
More information about the foms