[Subject Prev][Subject Next][Thread Prev][Thread Next][Date Index][Thread Index]

[hts-users:02057] Re: Problem Feature Extraction


Heiga, do you know how to make the extraction of mgc-lsp right?
Except the method of -e and -f,
I tried "$mgcep -e 1000000 -f 1e-1000000000"  or "$mgcep -e 1 -f 1e-1000000" but it doesnot work.



what is difference between mgc-lsp and mfcc?
seems the mfcc extraction donot need the theq() iteration.

Thanks


2009/6/25 Heiga ZEN (Byung Ha CHUN) <heiga.zen@xxxxxxxxxxxxxxxxx>
Hi,

wangxiang wrote (2009/06/25 3:48):


But I used Straight synthesizer but got also silence of the synthesized wave
and I found something strange in the mgc-lsp coefficient, the first
dimention of one frame is a strange number like -8 or -10

They are (log) filter gain rather than LSPs so they are not strange.


Regards,

Heiga ZEN (Byung Ha CHUN)

--
Heiga ZEN (Byung Ha CHUN)
Speech Technology Group
Cambridge Research Lab
Toshiba Research Europe
phone: +44 1223 436975

______________________________________________________________________
This email has been scanned by the MessageLabs Email Security System.
For more information please visit http://www.messagelabs.com/email ______________________________________________________________________




--
Xiang Wang
Institue Of Acoustics, CAS


References
[hts-users:02034] Re: Problem Feature Extraction, Geoffrey Wilfart
[hts-users:02040] Re: Problem Feature Extraction, wangxiang
[hts-users:02041] Re: Problem Feature Extraction, Geoffrey Wilfart
[hts-users:02043] Re: Problem Feature Extraction, wangxiang
[hts-users:02045] Re: Problem Feature Extraction, Geoffrey Wilfart
[hts-users:02046] Re: Problem Feature Extraction, wangxiang
[hts-users:02050] Re: Problem Feature Extraction, Heiga ZEN (Byung Ha CHUN)
[hts-users:02051] Re: Problem Feature Extraction, Heiga Zen (Byung Ha CHUN)
[hts-users:02054] Re: Problem Feature Extraction, wangxiang
[hts-users:02055] Re: Problem Feature Extraction, Heiga ZEN (Byung Ha CHUN)