Before reading more about ASF, I was very enthusiastic about the possibility of using it. As I read more and more, I was turned off!

As admitted by Microsoft, ASF is to be a replacement for the RIFF (Resource Interchange File Format) format designed by IBM & Microsoft. AVI, and WAV, conform to the RIFF format, which interleaves “chunks” of data of different types.

Their claim is that ASF is to improve on RIFF by making three “fundamental” changes. These are:

1. “Bona-fide” objects replace “Chunks” of data.
2. Replace the FOURCC(four-character code) identifier used in RIFF by the UUID (Universally Unique Identifier).
3. Increase the length field from 32 bits to 64 bits allowing for streams larger than 4 Gigs.

As we know, not all AVI files are suitable for streaming although they do interleave streams. This is because putting all the audio before all the video in a single AVI file is also considered to be interleaving! I do not know much about RIFF, but if AVI conforms to the RIFF format and the three items above are the “fundamental” changes, then this doesn’t say anything better about ASF then RIFF. From my understanding, these “bona-fide objects” are not so bona-fide, and the other changes are trivial.

According to Bart, Microsoft is not willing to share the specs of ASF easily. From what I have read, they were supposed to have made the specs public at the spring of 97 for public-review. This has not happened.

I believe that ASF if too general to be well suited for streaming. Using RIFF as a base for ASF might be a bad idea. RIFF might have been built with emphasis on static media types, while ASF should be built with emphasis on streaming media types.

One last thing. Libraries for ASF v1 are being written as C++ classes, whereas ASF v2 libraries will be offered as Active X (OLE/COM) objects. Ridiculous!!!

I think we could learn a lot if we saw the specs for ASF but I think we would be better off implementing our own multiplexing stream protocol that is more strictly defined. I will explain what I have in mind in a later message.