search

google-code-export / voiceid

Automatically exported from code.google.com/p/voiceid
0 stars 2 forks source link

Never outputs gender=U (Unknown) #10

Open GoogleCodeExporter opened 9 years ago

GoogleCodeExporter commented 9 years ago 
What steps will reproduce the problem?
1. use any of the get_gender functions
2. provide a file with noise or silence

What is the expected output? What do you see instead?
Expected output: Gender : U
Actual output:   Gender : F

What version of the product are you using? On what operating system?
Ubuntu 12.04. Revision 192

Please provide any additional information below.
Sample audio file attached.

Original issue reported on code.google.com by Millenni...@gmail.com on 8 Aug 2012 at 6:01

Attachments:

GoogleCodeExporter commented 9 years ago 
This is a side effect of a bad use of the system, it should never get in input 
a noise or silence wave file, it should be rejected by the voice/silence 
detection. We will take a look at this problem as soon as possible.

Original comment by maurome...@gmail.com on 12 Sep 2012 at 1:49

GoogleCodeExporter commented 9 years ago 
Any update on this?

Original comment by Millenni...@gmail.com on 27 Nov 2012 at 8:15

GoogleCodeExporter commented 9 years ago 
We tried to fix it but it's a lium library bug. We have reported it to the lium 
SpkDiarization team. We hope to have a solution as soon as possible.

Original comment by michelaf...@gmail.com on 11 Dec 2012 at 10:46

GoogleCodeExporter commented 9 years ago 
Any updates on this?

Original comment by Millenni...@gmail.com on 26 Apr 2013 at 9:11

GoogleCodeExporter commented 9 years ago 
We are dealing with it and, as said before, maybe it's an issue about Lium lib 
(no answers from them). By the way, the sample you attach was not recorded by a 
microphone, it is an unnatural no energy signal, not sure this make any 
difference.
Is it really useful? The project aims to recognize voices, the U gender should 
never be the desired output... What's your application domain? Have you single 
files or a big file to split by diarization?

You can alternatively rebuild the models of silence and genders used by the 
system, they should be located at /usr/local/share/voiceid/ in your system.

Original comment by maurome...@gmail.com on 6 May 2013 at 3:19