gnuspeech-contact
[Top][All Lists]
Advanced

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

Re: [gnuspeech-contact] Question regarding sensitivity functions


From: David Hill
Subject: Re: [gnuspeech-contact] Question regarding sensitivity functions
Date: Fri, 22 Jan 2010 19:10:12 -0800

Dear Francesco,

The best place for you to start work on this would be with the original work on which the Distinctive Region Model is based; 

FANT, G. & PAULI, S. (1974) Spatial characteristics of vocal tract resonance models. Proceedings of the Stockholm Speech Communication Seminar, KTH, Stockholm, Sweden 

and

FANT, G. (1960) Acoustic Theory of Speech Production. Mouton: The Hague328 pp, (Second Printing 1970)

The first reference is the paper on formant sensitivity analysis, and the book is the classic text, resulting from Dr. Gunnar Fant's thesis. I will also forward your enquiry to Dr. Leonard Manzara who developed the tube model we are actually using, based on the work of Julius Smith and Perry Cook at the Stanford Center for Computer Research in Music and Acoustics (CCRMA). He may be able to give you more specific help.

All good wishes.

david
--------
David Hill, Emeritus Professor,
----------
Our lives begin to end the day we become silent about things that matter.
(Martin Luther King Jr.)


On Jan 13, 2010, at 10:21 AM, Francesco DEste wrote:

Dear Gnuspeech Contact Mailing List,
I am a student at the Leiden University, currently developing a thesis in  Vocal Tract Area
Function  estimation. 
I am using the gnuspeech tube waveguide software embedded in an Multi-Objective optimization software.
The method I implemented uses a MO-Evolutionary algorithm to try to match
a target sound.

I would like to ask a couple of informations regarding the Distinctive Region Model
and the sensitivity function calculation.
Given an arbitrary posture, is it possible to calculate its sensitivity function efficiently?
From what I understood, the table shown in the documentation (R1-R8 against F1-F3)
works only when the tube is on a 'close to straight' configuration.

Since the sensitivity is a function of the volume velocity and pressure, is there a way
to keep track of this data inside the tube routines?

If it would be possible, I would like to incorporate this knowledge inside the evolutionary
operators of my system.

Many thanks for the attention,
with kind regards.
Francesco D'Este




Windows Live: Make it easier for your friends to see what you’re up to on Facebook.
_______________________________________________
gnuspeech-contact mailing list



---------



reply via email to

[Prev in Thread] Current Thread [Next in Thread]