I am user of windows 7, I use Audacity 2.0.5 which I downloaded in its zip version.
I have checked different ways to compress files into OGG and I have used Audacity, oggenc2 and MATLAB’s function ‘audiowrite’. I have found out that the results between Audacity and MATLAB’s ‘audiowrite’ are very similar (both number of samples and same amplitude but when miminum differences). Whereas when using oggenc2 the output contains 22 more samples and the amplitude is clearly different.
Oggenc2 is not the official OGG encoder. As far as I know its main difference is that it can accept input from a wider range of lossless codecs and has better support for multi-channel input.
Audacity uses the official OGG reference encoder.
Audacity does not develop libvorbis or oggenc2. If you want to find out more about oggenc2 I suggest you look at its source code or contact its developer John33 via the Hydrogenaudio forums: http://www.hydrogenaudio.org/forums/index.php?showtopic=47538 .
I couldn’t help but smile. “22 more samples” and we’re getting into very fine distinctions between reference codecs and reference implementations.
Side stepping slightly, the “official” (Ogg) Vorbis encoder by the Xiph.org Foundation (http://www.xiph.org/) is Oggenc and is is part of vorbis-tools.
Oggenc2 is an enhanced version of oggenc developed by John33.
Xiph.org and the the Hydrogenaudio forums will be the best sources for detailed information.