Open 1973Blunt opened 5 years ago
Thank you for your interest to my project,
I just used already existed code so that I didn't carefully concern about the difference between the implementation that I used and that reference you mentioned.
However, I think the difference is minor when we compare our method to others
Thanks for sharing your code!
This issue is about LogSpectralDistance calculation. Source code is:
LSD=mean(sqrt(mean((log(CL(RangeBin,1:N))-log(NO(RangeBin,1:N))).^2)));
It's not consistent with the reference:
Difference is: