- From: Silvia Pfeiffer <silviapfeiffer1@gmail.com>
- Date: Wed, 1 Apr 2009 23:36:15 +1100
- To: Jack Jansen <Jack.Jansen@cwi.nl>
- Cc: Media Fragment <public-media-fragment@w3.org>
As mentioned in the meeting today, I think that "taking care of the user" is up to the application that the user is using. The server, user agent and intermediate caches can only do best effort. Best effort needs to include the interval requested by the user so that the application can deliver the requested subpart to the user. What "best effort" is will depend on the state of support of implementations of our specifications. The best support is where the encapsulation format can support the specification of the start/end points even if the encoded data includes more before and after it. That will give any application the possibility to deal properly even with a wget downloaded file. But there will be stages of support for media fragments and some combinations will work better than others. The cleanest solution will also be without any transcoding, but to get a working solution for many encoding formats we may need to deal with an interim solution where we have transcoded files. For some file formats it will not even work without transcoding, e.g. addressing spatial fragments. But in future there may be encoding formats developed that will be capable of all these things. While we want to be purist, I think we also have to be realists. :) (sounds stupid, but I think it has to be said) Cheers, Silvia. On Wed, Apr 1, 2009 at 7:59 PM, Jack Jansen <Jack.Jansen@cwi.nl> wrote: > Another angle on the issues I found this week: maybe part of the problem is > that the only entities we've considered so far are (1) the client-side > software, (2) the server and (3) a caching server. > > If we also specify what the user can expect things may get easier. For > example, if the user does > > wget http://www.example.com/movie.ogg#t=1,2&xywy=9,9,101,101 > > there are a number of questions: > 1. Does this return the exact 1s-2s time fragment? This is probably > impossible, the next best thing is the smallest interval that contains this > interval, i.e. from the last frame before or at 1s to the first frame at or > after 2s. > 2. Alternatively, we could say the user gets a "reasonable" interval around > 1s-2s, so implementations can cut at I frames. This would forestall > transcoding. But: now we need to define "reasonable". > 3. What about audio/video sync? If the user gets synced a/v we need to do > recoding. > 4. What about timestamps in the media? Are these the originals? > 5. What about spatial crops? Same questions as for (1) and (2) > -- > Jack Jansen, <Jack.Jansen@cwi.nl>, http://www.cwi.nl/~jack > If I can't dance I don't want to be part of your revolution -- Emma Goldman > > > > >
Received on Wednesday, 1 April 2009 12:37:08 UTC