> MP3 - has some (I believe) serious design flaws which the LAME
> developers and others have been trying to eliminate
>
MP3 has design flaws incorporated in standard definition, and nobody could
do anything about it while trying to be ISO compatible:
1. Frame-based M/S switching (this almost kills the idea of M/S coding)
2. Bad filterbank (hybrid filterbank has shown inferior when compared to
pure-MDCT approach)
3. Bad filterbank length (reasons for that are pure politcics, because some
companies in MPEG wanted "compatibility" with older layers)
> MP3Pro - Only really aimed at low bit rates
>
It is technology called 'SBR' - good thing to say is that SBR and other
similar technologies are being evaluated by MPEG right now, and in about a
year we will have official bandwidth enhancing technology, which will have
to be open (SBR is proprietary). This new technology will be part of MPEG-4
Audio standard, but it could be applied to any audio coding algorithm.
> MP3plus - Looks like a good format... the developer *seems* to have
> taken the good stuff out of MP3 and added enhancements
> of his own, but it is not well known (or accepted)
>
It is MP+, or MPC - it is based on MPEG-1 Layer II (MP2) with added
improvements, like huffman coding, variable bit rate, etc... And much better
implementation of psychoacoustic model.
> WMA - Microsoft proprietry (a bad thing to start with) also aimed
> at low bit rates, but just happens to allow the use of higher
> rates. Can't convert back to WAV (*REALLY* bad limitation)
>
One word - sucks :)
> OGG - Seems like a good format, eats MP3 at low bit rates, but then
> falls badly as the bit rate is decreased further. There was some
> talk of OGG being in patent infringement (although that is *very*
> dubious, and is probably a rumor started by Thompson or FIIS
> to scare people away - and then they can get their hooks
> into you for MP3...)
>
Ogg is a very good format, no doubt about it - only drawback is the lack of
well-defined standard, i.e. - not ISO
> AAC - I have not been able to find any encoders that can do anything
> like real-time. Decoders are also rare and usually linked to a
> particular encoder. (ie: the AAC streams produced by any given
> encoder are not necessarily compatible)
>
Wrong - All licensed AAC implementations are compatible with each other on
so-called "elementary" bitstream format. However, for end-user applications
it must be encrypted, as demanded by essential patent holders (however,
during my negotiations with them nobody really raised that question)
http://www.inf.ufpr.br/~rja00/
You will find several AAC encoders there, all compatible with ISO decoders
and two of them are really high quality :) One of them is also very fast,
working 4 times faster real-time on PIII 850 MHz
> I think the LAME developers have done an excellent job of minimising
> the flaws inherent in the MP3 specification, and further, I don't believe
> they are developing MP3 to write a better encoder for *you* - after all,
> what are they getting out of it?
>
> Only thing I see is personal understanding and intellectual enhancement.
>
_______________________________________________
mp3encoder mailing list
[EMAIL PROTECTED]
http://minnie.tuhs.org/mailman/listinfo/mp3encoder
