[This is rob -- still in Germnany]

> I have done that, and found out that only layers 1 and 2 are implented in
> fixed point arithmetic? Am I right? A short answer will help a lot.

I will take another look, but I suspect that layer 3 integer may not be supported. 
Here is how things went back at Xing: 

1) The first decoders were written around the time of the i386 when floating point was 
2) Then with the 486DX and the Pentium code was written to support floating point 
3) When layer 3 decode was written a Pentium had become standard, thus the target 
customer base has a floating point unit available. For that reason I think the integer 
decode was not implemented.

I'll take a look at the code again.



Reply via email to