Tracking Vocal Tract Resonances Using an Analytical Nonlinear Predictor and a Target-guided Temporal Constraint

A technique for high-accuracy tracking of formants or vocal

tract resonances is presented in this paper using a novel nonlinear

predictor and using a target-directed temporal constraint.

The nonlinear predictor is constructed from a parameter-free,

discrete mapping function from the formant (frequencies and

bandwidths) space to the LPC-cepstral space, with trainable

residuals. We examine in this study the key role of vocal tract

resonance targets in the tracking accuracy. Experimental results

show that due to the use of the targets, the tracked formants in

the consonantal regions (including closures and short pauses)

of the speech utterance exhibit the same dynamic properties as

for the vocalic regions, and reflect the underlying vocal tract

resonances. The results also demonstrate the effectiveness of

training the prediction-residual parameters and of incorporating

the target-based constraint in obtaining high-accuracy formant

estimates, especially for non-sonorant portions of speech.

2003-deng-eurospeech.pdf
PDF file

In  Proc. of the Eurospeech Conference. Geneva

Details

TypeInproceedings
> Publications > Tracking Vocal Tract Resonances Using an Analytical Nonlinear Predictor and a Target-guided Temporal Constraint