Cross-fading weight function algorithm

eagle-i ID

Resource Type

Properties

1. Resource Description

   "We propose a new algorithm that uses a unit-dependent trainable parameterized cross-fading weight function to generate more natural-looking formant trajectories and, it is hoped, better-sounding output speech. The proposed algorithm: - uses a perceptually-based objective function to capture differences between cross-faded and natural trajectories across the whole region of the phoneme, and - uses phoneme identity, prosodic contexts, and acoustic features of the units to predict optimal cross-fading parameters to generate more natural formant trajectories."

2. Used by

   Center for Spoken Language Understanding

3. Related Publication or Documentation

   Perceptual Cost Function for Cross-fading Based Concatenation

4. Developed by

   Kain, Alexander, Ph.D.

5. Developed by

   van Santen, Jan P.H., Ph.D.

6. Developed by

   Miao, Qi