At high bitrates ~200KBit/s the quality is extremly good in my opinion, I couldnīt hear any difference, I also looked at the spectral view with cooledit: no cutoff at all, also with a view zoomed to yust one second I couldnīt see the difference between both graphs (normally the graph of encoded files are very blocky, exspecially with encoders like Blade, Xing, Plugger, Shine etc..). I tested castanets to check pre-echo: No pre-echo at all. I couldnīt hear/see any difference.
This format seems to differ from others very much, you also cannot set bitrate, you can set Quantization, Predictor Size, can also set downsampling (1:1 - 1:10). You can also activate/deactivate joint-stereo. My settings were: Quant: 0.40; Predictor Size: 107, Downsampling: 1:1, Joint Stereo activated. Maybe someone with very good hearing and a better system can prove this.
But at lower bitrates (~130kBit/s) it sounds a bit noisy for me, the background seems to get louder.
It also has an lossless mode, but the files are about 5% (from the original) bigger than monkeyīs audio.
QUOTE
But they obviously don't know anything about quality mp3s as they offer a link to Blade...
But the encoder inculded in the program are Lame, Vorbis and Bonk.
Edit: The site posted by Rommel is not the main-site of the Bonk-Format, just of the frontend, this is the main-site:
http://yoyo.cc.monash.edu.au/~pfh/bonk/The format is multi-platform open-source and licensed under GPL.
There is also a XMMS-plug-in and a other player. Unfortunately the winamp-plug-in doesnīt work.