how to store extracted features

[yashpal]

hi to all.
i am working on Speech emotion recognition using SVM.
i am using matlab's colea toolbox for extracing the features like
pitch and formanant frequencies. but dont know how to store them in a
single database for training and recognition. plz help me out...
n also sugest any another tool for feature extraction....
thanking you in advance.

[jonas]

Hi Yashpal.
First, why are you extracting F0 and formant features (I guess state
of the art would be MFCC or rasta PLP)?
Second, are you extracting F0 and formant values for the entire
recording? The duration of the recordings?
One idea would in that case be to store them as GMMs. I suppose that
Praat would be a good feature extractor as well http://www.praat.org
Best regards
Jonas

[Md Sahidullah]

Dear Yashpal,

I also agree with Jonas . Why to extract the formants? And is it
possible to model them like conventional speech features like MFCC or
LPC ? I also think it is better to use them and model those features
in SVM or other techniques.
Just follow the basic feature extraction procedure and store those
in .mat format if required.
For your reference you can check the following paper

"Capturing Complementary Information via Reversed Filter Bank and
Parallel Implementation with MFCC for Improved Text-Independent
Speaker Identification "
Author: Sandipan Chakroborty et al.

Proceedings of the International Conference on Computing: Theory and
Applications table of contents
Pages 463-467
Year of Publication: 2007
ISBN:0-7695-2770-1

Thanks & Regards
Md. Sahidullah