How to get my own mu.mat？

[NewEricWang]

Hi, @anicolson ,
Thanks for your great work!
I want to train DeepXi model using my own data now. But I don't know how to get mu.mat and sigma.mat based on my data. Could you give me some help?
Any advices will be appreciated!

[anicolson]

Hi, thank you, mu.mat contains a vector of size NFFT/2 + 1 (257). xi_dB(n,k)=10*log10(xi(n,k)), where n is frame index and k is frequency index from k=0 to k=NFFT/2+1. The mean for each frequency component is taken over the training set, i.e. for the k^th component, mean(xi_dB(:,k)). Similarly for the standard deviation: stddev(xi_dB(:,k)) once you compute this vector, it can replace mu.mat, just store in in a numpy array or a pickle file. more details can be found in the manuscript. https://maxwell.ict.griffith.edu.au/spl/publications/papers/spcom19_aaron_deep.pdf hope this helps please contact me again if this is not clear Get Outlook for iOS<https://aka.ms/o0ukef>

…

________________________________ From: NewEricWang <notifications@github.com> Sent: Wednesday, July 10, 2019 7:58:04 PM To: anicolson/DeepXi Cc: Aaron Nicolson; Mention Subject: [anicolson/DeepXi] How to get my own mu.mat？ (#4) Hi, @anicolson<https://github.com/anicolson> , Thanks for your great work! I want to train DeepXi model using my own data now. But I don't know how to get mu.mat and sigma.mat based on my data. Could you give me some help? Any advices will be appreciated! — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub<#4?email_source=notifications&email_token=AGHGZ7RJQPR676QQ4REL55DP6WXCZA5CNFSM4H7NT3SKYY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4G6J4S5Q>, or mute the thread<https://github.com/notifications/unsubscribe-auth/AGHGZ7W2CRGQRCY6WR54IALP6WXCZANCNFSM4H7NT3SA>.

[NewEricWang]

@anicolson, thank you again!