> > Hi all, > > I am a student doing a project to add MS to the dist10 code. I had > implemented the JStereo/MS stereo to the code. The default is now > MS. There is no switch yet on the code. However, the quality of the > sound is similar to L/R and compare to Fhg, it is not as > good. According to the 1172-3 standards, there is something on > limitation of Si channel bandwidth and 'sparsing' of Si > channel. Wondering if this is the reason for the difference in quality > of my encoder? I have read that the 'sparsing' base on 11172-3 is not > workable if used on scalefactor bands. Is this true? Can anyone advice > on improving the quality for MS stereo? I will appreciate any > advice. Thank you. > > > Hann >
I tried and was unable to get any good results using 'sparsing'. Analysis of the FhG codec with our frame analyzer will also show you that FhG is not using sparsing as described in the ISO docs. Adding MS stereo is a good place to start - this is exactly what I did in LAME 3 years ago :-). But MS stereo is just a small improvement compared to fixing many of the fundemental flaws in dist10. The are serious problems with the psycho acoustics, quantization/huffman encoding and noise shaping routines, and these are what causes the poor quality in your encoder. Mark _______________________________________________ mp3encoder mailing list [EMAIL PROTECTED] http://minnie.tuhs.org/mailman/listinfo/mp3encoder
