The concept of this suggestion is a very good one; of course,
since SAOL lives within MPEG-4, we would use MPEG-2 AAC
for sample compression.
(For those of you not intimate with MPEG, MPEG-2 AAC is a new
and improved standard, just released this spring, which covers
the same territory as MP3, but gives *better* quality at *half* the
bitrate. That is, the quality of AAC is better at 64 kbps/channel
than MP3 at 128 kbps/channel. AAC is all-but-CD-quality at
64 kbps/channel.)
The only question (it's a difficult one) is how to describe this process
in the "standard decoder model" of MPEG-4. That is, MPEG-4 right
now views AAC data as streaming sound that goes right out to
the listener, possibly after some transformation by SAOL or
3-D spatialization. But in Juhana's suggestion, we have to
"capture" or "cache" the decoded sound and use it in a "separate"
part of the decoding process.
Note that this is only for the *standard* transmission of sounds;
implementors are free to include support for any sound formats
they like in an authoring or decoding terminal. But to have the
maximum interchange possibility, this sort of thing has to be
standardized in the bitstream format.
Best regards to all,
-- Eric
----------
> From: Juhana K Kouhia <kouhia@nic.funet.fi>
> To: saol-dev@media.mit.edu
> Subject: Saol suggestion
> Date: Wednesday, November 26, 1997 8:14 AM
>
> Hello. It looks like the saol samples in the mpeg bitstream are all
either
> float or 16 bits pcm audio. What if I want distribute all samples as mpeg
> layer 3 compressed? Takes less space to distribute samples and the
quality
> is good enough. I would like to have your comments please.
>
> Juhana
This archive was generated by hypermail 2b29 : Wed May 10 2000 - 12:11:10 EDT