> 
> Hi all,
> 
> Just to inquire more about MS Stereo. I am confused about how the block is
> decided between long and short blocks. 
> 
> In the previous algorithm, the blocks are decided based on LR thresholds
> after L3psycho_anal. Is that true? Can't we use the final LR/MS thresholds ie
> fthr for the block decision.
> 
> One more thing on bmax(b) from 13818-7 MS Stereo section. Is that applied to
> both short and long blocks? In lame 3.21, I notice that it is use on the
> scalefactorbands instead of the partitions. So when the standard states to
> calculate the LR/MS and final thresholds, it actually means the LR/MS ratios in
> the after calculating the PE. Is that right?
> 
> Will appreciate all assistance.
> 
> Hann Long
> NUS Student
> 

Keep in mind that the ISO documentation is required since the format
is going to be some kind of standard.  But the companies writing
it are surely *not* trying to make it easy for other people 
to impliment a good encoder, and I'm sure they leave out as much
as possible.  So if something in the ISO docs does (like the completely
broken M/S stereo "sparsing") not make sense, try other strategies.
Tonality estimation is another good example - the default tonality
estimater in the FhG mp3 encoder is not what is given in the ISO
docs.

If you do experiment with other algorithms and find something that
works well, I hope you'll share your experiences with this list.

Mark





_______________________________________________
mp3encoder mailing list
[EMAIL PROTECTED]
http://minnie.tuhs.org/mailman/listinfo/mp3encoder

Reply via email to