Judith-

> There is a lot of overlap of values(a confusion) in the
> features you extract from the emotions.so classifying
> emotions  & tones is a big job.

I'm not the original poster, but yes I agree 100%.

-Jeff

______________________________
New Code Sharing Section now Live on DSPRelated.com. Learn about the Reward Program for Contributors here.



(You need to be a member of speech-recognition -- send a blank email to speech-recognition-subscribe@yahoogroups.com )

Indeed a big job. Not even having started with the job, there are few
things to ponder
1. emotion recognition is very much speaker dependent and culture dependent.2.
the emotion classes categorization. i.e. what kind of emotion you wish you
classify.
During my time working on speech emotion recognition, I have gone through
something interesting named as "valence-arousal" dimension, where the
emotion is categorized continuously, instead of in discrete, scaled on VA
dimension (the psychologist says so). lol
I am not working in that area now. However, I wish those who are interested may
move on with the study/research and keep us posted with the update.
-Chong

To: j...@yahoo.co.in
CC: s...@yahoogroups.com
From: j...@signalogic.com
Date: Wed, 7 Apr 2010 13:01:38 -0500
Subject: Re: [speech-recognition] Classifying 'tone' of voice during speech     
     recognition

      Judith-

> There is a lot of overlap of values(a confusion) in the

> features you extract from the emotions.so classifying

> emotions  & tones is a big job.

I'm not the original poster, but yes I agree 100%.

-Jeff

______________________________
New Code Sharing Section now Live on DSPRelated.com. Learn about the Reward Program for Contributors here.



(You need to be a member of speech-recognition -- send a blank email to speech-recognition-subscribe@yahoogroups.com )

Thanks Chong, Jeff and Judith.

This indeed is a big job. Apart from lite joke I am not going to kid myself
that I can do it just like that (education and skill wise), that will be
childish. I would love to work on something infinitely challenging like this
with no expectations of getting rich, haven`t done so since I became a slave
to paying bills in a one income family. With current commitments I don`t
think I will ever get to even try it. I was actually looking to collaborate
with someone (or may be a group) who has the required education, skills and
what else is required and who can work on this piece on the side.

A full proof emotion recognition system is involved task as everyone have
pointed out. We can start small and expand. e.g. start with one language-one
culture and limited emotions. Also pick the major overtone over a sentence
where there are mixed emotions etc. etc.

Unless the research is well funded we will have to develop bits and pieces
that can be commercialized in stages to generate funding for next stages.

Paresh
On Wed, Apr 7, 2010 at 11:40 PM, Chong Tze Yuang <c...@hotmail.com>wrote:

> Indeed a big job. Not even having started with the job, there are few
> things to ponder
> 1. emotion recognition is very much speaker dependent and culture
> dependent.2. the emotion classes categorization. i.e. what kind of emotion
> you wish you classify.
> During my time working on speech emotion recognition, I have gone through
> something interesting named as "valence-arousal" dimension, where
the
> emotion is categorized continuously, instead of in discrete, scaled on VA
> dimension (the psychologist says so). lol
> I am not working in that area now. However, I wish those who are
interested
> may move on with the study/research and keep us posted with the update.
> -Chong
>
> To: j...@yahoo.co.in
> CC: s...@yahoogroups.com
> From: j...@signalogic.com
> Date: Wed, 7 Apr 2010 13:01:38 -0500
> Subject: Re: [speech-recognition] Classifying 'tone' of voice during
speech
>           recognition
>      Judith-
>
> > There is a lot of overlap of values(a confusion) in the
>
> > features you extract from the emotions.so classifying
>
> > emotions  & tones is a big job.
>
> I'm not the original poster, but yes I agree 100%.
>
> -Jeff
>
>
>
> _____________________________________
>
-- 
Thanks
Paresh
416-688-1003
“It is not because things are difficult that we do not dare; it is because
we do not dare that they are difficult.”
Seneca

______________________________
StarterWare provides C-based no-OS platform support for the C6000 DSP devices.
Click here for details.



(You need to be a member of speech-recognition -- send a blank email to speech-recognition-subscribe@yahoogroups.com )