[M4IF Technotes] MPEG-4 Questions on Scalability

[Michael Ditze]

Hello!
I have some questions with regard to scalability in MPEG-4. Can you 
anyone answer some of these questions?
- Is there a difference between temporal and spatial scalability when it 
comes to interdependencies bewteen the base stream and the enhancement 
stream. I assume that the enhancement stream for temporal scalability 
depends on the base stream because of the applied Motion Compensation. 
Is that the same for spatial scalability?
-  In MPEG-2 you normally have fixed spacing in a Group of Pictures and 
between successive I-frames and successive reference frames (I- and 
P-frames) that may be represented by the M and N value. Would this be 
the same for MPEG-4?
- Even if MPEG-4 is scalable, does that mean that all streams are sent 
in any case through the network where the decodes neglects the 
enhancement streams if it cannot decode them in time? Or is there a 
mechanism that allows to send Enhancement Streams only when the server 
knows that the client may decode the stream in time (e.g. though an 
Admission Control)
- As the scene description (BIFS) and video streams are sent in 
different Elementry streams, is there a mechanism in MPEG-4 that ensures 
that the scene description is decoded before the respective streams. 
Hoes does that work?
I hope you can answer some of my question as this really important stuff 
I need for my thesis. If you have sources where I can find further 
information with regard to my questions, I would be very grateful.
Thanks for your efforts anf Kind Regards,
Michael Ditze
-- 
Michael Ditze
C-LAB, 33094 Paderborn, Germany
Fürstenallee 11, Office FU 304
Voice: +49-5251-60-6124, Fax: +49-5251-60-6065
Email: mailto:michael.ditze   c-lab.de