Hello Bruce, > Did you use Jean-Marc's statistical methods in retraining the codebook?
The codebook in Codec2 is just a Cartesian product, so there was no need to retrain anything. Also, this patch does not change anything at that level of the code. > How does this differ from Opus at that rate? Or does Opus not get that low? I had never heard from Opus, but according to opus-codec.org it does not go below 6 kb/s. Please note that I am not proposing any change to the wire format for Codec2 with this patch; I am merely suggesting to exploit the full time domain size of the FFT (processing 512 instead of 320 samples en bloc) and deal with the slightly altered frequency domain that follows from it. It's very simple work, really. It may just have a big impact on quality, that's all :-) Cheers, -Rick ------------------------------------------------------------------------------ Keep Your Developer Skills Current with LearnDevNow! The most comprehensive online learning library for Microsoft developers is just $99.99! Visual Studio, SharePoint, SQL - plus HTML5, CSS3, MVC3, Metro Style Apps, more. Free future releases when you subscribe now! http://p.sf.net/sfu/learndevnow-d2d _______________________________________________ Freetel-codec2 mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/freetel-codec2
