Expert opinions needed

2002-04-25 21:46:26

I was idly searching for Ogg Vorbis stuff today and stumbled upon this article.
I'm well aware its an old outdated test but in your opinions does this guy know what he was about? And how far has Ogg Vorbis progressed since then,thanks guys.

http://216.239.39.100/search?q=cache:tlDdy...ersus+mp3&hl=en

Expert opinions needed

Reply #1 – 2002-04-25 22:46:55

Vorbis has progressed a hell of a lot since January 2001. Back then, the Vorbis encoder wasn't even in Release Cantidate stage. I didn't read his evaluations of the quality too closely, but regardless of how accurate they are, they really aren't relevent anymore since Vorbis is much, much better now than it was over a year ago.

Expert opinions needed

Reply #2 – 2002-04-25 23:02:36

His tests were only informal, not blind, and the quality has improved enormously in the last 16 months. His comments about Ogg being patent and license free are accurate.

Expert opinions needed

Reply #3 – 2002-04-26 02:42:03

Well, they did manage to get the fact that Vorbis uses a MDCT (Modified Discrete Cosine Transform) right for tonal noises/stationary signals and this is where the similarities end. Block switching and temporary masking models have sinced improved since that article was published as well. I actually saw something interesting on that article that reminded me of something. Has anybody heard anything about any future versions of the LAME Encoder having any support for doing some experimental things with libvorbis next to experimental MPEG-1 Layer 3 models? or did they decide to scrap that idea for some appearent reason?, because I had noticed that LAME 3.91 and up has support for decoding Ogg Bitstreams back into RAW PCM Samples and that got me thinking why wouldn't the LAME project do any experimental work with Vorbis, that wouldn't be quite a bad idea. I guess I will find out soon enough why they won't or can.