• Corpus ID: 2726169

Model based analysis of a diphone database for improved unit concatenation

  title={Model based analysis of a diphone database for improved unit concatenation},
  author={Karl Schnell and Arild Lacroix},
One crucial point of concatenation approaches using diphones is to handle the discontinuities between the concatenated units. This problem is treated by a suitable analysis of the diphones for a parametric synthesis. The model of the parametric synthesis is the lossy tube model, which is an extension of the standard lattice filter considering frequency dependent vocal tract losses. The parameters of the tube model are estimated from diphones by an optimization algorithm. The discontinuities of… 
Combination of LSF and pole based parameter interpolation for model-based diphone concatenation
A comparison between the LSF and the pole based interpolations shows that the L SF interpolations can be improved by the corrected pole assignments and by the trajectories of the poles.
Joint analysis of speech frames for synthesis based on lossy tube models
Investigations show that, especially for extended tube models, the joint analysis of frames improves the quality of the synthesized speech signals.


Speech production based on lossy tube models: unit concatenation and sound transitions
The investigations show that the lossy tube model is advantageous compared to the lossless standard tube model for speech analysis and speech production.
A comparison of spectral smoothing methods for segment concatenation based speech synthesis
Analysis of lossy vocal tract models for speech production
For the analysis of speech sounds the parameters of a lossy vocal tract model are estimated from speech signals by an optimization algorithm and the spectrum of the analyzed speech can be approximated well by the estimated magnitude response of the lossyVoice tract model.
Modelling of LIP radiation impedance in Z-domain
  • U. Laine
  • Mathematics, Computer Science
  • 1982
Three z-domain models for lip radiaton impedance are introduced. Two of them are based on the observation that the normalized acoustic impedance can be modelled as z(\omega) = C*[1.-\cos(\omegaT)] +