- mppenc: Some highly experimental code for better thumb encoding (may be interesting for bitrates <140 kbps)
Untuned, early alpha stage
- WinAMP plugin: several bug fixes, tagger tries to guess the entries if file has no tags
- XMMS: Seek speedup, bugfix for out-of-sync error
other unimportant changes
- mppdec / WinAMP / XMMS: Support of new thumb files (currently source only)
- Examples for thumb: fossiles_old.mpc (90 kbps) and fossiles_new.mpc (79 kbps).

Example for lowest bitrate: KMFDM-Dogma-new.mpc (29 kbps)

mppenc binary: http://www.uni-jena.de/~pfk/mpp/bin/mppenc-windows-1.02.zip
Compiled Winamp plugin: http://www.saunalahti.fi/~cse/in_mpc_0.93.zip
Compiled decoder binaries for Windows: http://www.saunalahti.fi/~cse/mppdec_1.02.zip

Well, it's not better thumb profile at least quality wise. fossiles_old thumb (although the link is not working) sounds much better than fossiles_new.

You did use the new decoder to play it? To me the new version sounds better, old is too muffled.

With the old plugins and decoders they sound
horrible. You haven't updated them. Plugins
on Case's page should work, although I haven't
test it.

Ah yeah.. Didn't notice that you need also the new plugin, sorry!
Yeah at least fossiles sounds better now, haven't tested other files.

[edit] Hopefully the new encoder binary will be available soon?

Btw. Frank. Can you explain what is this new technique you implemented in thumb?

when will the windows encoder binary be available ?

I got some great info from Frank via email.
The new technique behind the new thumb-profile is perceptual noise substitution (PNS).

New encoder binary is not gonna be there before next Tuesday according to Frank.

fwiw-the new 24bit decoder (available on case's page) works fine-I "tested" it all through last night...

Sounds similar to what MP3Pro and PlusV do ?

--
GCP

It would be interesting to see how well this new thumb-profile competes with Ogg/MP3pro at low bitrates!

Is the original for "fossiles" available somewhere in lossless format?

Yes, from ff123's samples webpage. http://ff123.net/samples/fossiles.flac

No, it's not like spectral band replication.

The following is by Menno about AAC PNS, I think it applies pretty much for MPC's PNS as well:
The technique of Perceptual Noise Substitution is based on the observation that one noise sounds like the other. This means that the actual fine structure of a noise signal is of minor importance for the subjective perception of such a signal. Consequently, instead of transmitting the actual spectral components of a noisy signal, the bitstream would just signal that this frequency region is a noise-like one and give some additional information on the total power in that band. PNS can be switched on a scalefactor band basis so even if there just are some spectral regions with a noisy structure PNS can be used to save bits. In the decoder, a randomly generated noise will be inserted into the appropriate spectral region according to the power level signaled within the bitstream.

from Frank:

I introduced PNS. Higher subbands can be encoded as pure noise if this
sounds not too bad. For that

- A subband must be spectral unformed, i.e. does not contain any tonal components
- A subband frame must be temporal unformed, i.e. spectral density must be
constant

This saves some data rate, because the encoded subband samples are not
encoded, only the power density.

I wonder if high profile, such as standard and xtreme can benifit a little from the technique of PNS without suffer much quality loss, or it's only useful with low bitrate profile?

PNS is a techonolgy for low bitrate only.