- From: Davy Van Deursen <davy.vandeursen@ugent.be>
- Date: Tue, 28 Oct 2008 11:45:31 +0100
- To: "'Silvia Pfeiffer'" <silviapfeiffer1@gmail.com>
- Cc: "'Media Fragment'" <public-media-fragment@w3.org>
Hi Silvia, all, >-----Original Message----- >From: Silvia Pfeiffer [mailto:silviapfeiffer1@gmail.com] >Sent: Monday, October 27, 2008 10:55 PM >To: Davy Van Deursen >Cc: Media Fragment >Subject: Re: SMIL section of state-of-the-art document done > >Hi Davy, all, > >On Tue, Oct 28, 2008 at 1:32 AM, Davy Van Deursen ><davy.vandeursen@ugent.be> wrote: >>>> On 27 okt 2008, at 12:11, Silvia Pfeiffer wrote: >>>BTW: Davy - I'd be curious if your meta-specification format of the >>>structure of audio & video could be mapped into ROE somehow... >> >> The model for audio and video resources that I have developed is not >> designed from a fragment addressing point of view. It is made from an >> adaptation point of view and is closely related with the structure of >the >> media resources and addresses a resource in terms of bytes. > >I suppose what I meant was that your structure will be much deeper and >more detailed down to they byte level. However, I assumed you would >also need to cover the more high-level structure, such as different >tracks. Correct, for the moment, we only support elementary bitstreams. An extended version of the model could cover different tracks in a container format. We are working on that ;-). > >I agree that the aim of ROE is very different. I just thought that a >comparison may be instructional. > >> A trivial >> version of such a model could be that a video resource consists of a >list of >> frames, and that for each video frame the start offset and length in >terms >> of bytes are included. The CMML/ROE solution does not provide direct >links >> to the bytes (I guess this is left over to the application) > >ROE is indeed not meant to operate on that level. The byte mapping is >encoding format dependent and therefore left to the application. > >What is interesting about your format (btw: does it have a name?) is >that it could be used as information to hand off to Web proxies in >parallel with the media byte stream and provides it with information >on how to do byte ranges and time ranges. Our format does not have a real name, I call it "model for multimedia bitstreams" :-). I fully agree that this could be meaningful information for Web proxies to perform the necessary adaptations. Note that there is already some work done regarding generic network adaptation nodes in [1] and [2]. [1] M. Ransburg, C. Timmerer, H. Hellwagner, and S. Devillers. Design and evaluation of a metadata-driven adaptation node. In Proceedings 8th International Workshop on Image Analysis for Multimedia Interactive Services (WIAMIS), pages 83-86, Santorini, Greece, June 2007. [2] R. Kuschnig, I. Kofler, M. Ransburg, H. Hellwagner. Design options and comparison of in-network H.264/SVC adaptation, Journal of Visual Communication and Image Representation. In Press, Corrected Proof, Available online 5 August 2008. Best regards, Davy -- Davy Van Deursen Ghent University - IBBT Department of Electronics and Information Systems Multimedia Lab URL: http://multimedialab.elis.ugent.be
Received on Tuesday, 28 October 2008 10:46:15 UTC