|
Speech Processing
Vích, Robert
Speech Processing is a periodic publication collecting the contributions presented at the Czech-German Workshops on Speech Processing organized every year in September in Prague. The papers are devoted to speech analysis, synthesis, recognition and phonetics.
|
| |
|
Application-based prosody of synthetic speech
Dohalská, Marie ; Duběda, T. ; Mejvaldová, J.
"Speech styles" are defined as sets of prosodic features characteristic for a particular synthesis application. The studied styles (Address, Spelling, Headings, Dates and Prices) are non-continuous: they have no real syntax, no punctuation, they are nominal, and the information is very compressed and often numeric. To generate the utterances, punctuation and pauses are inserted, local FO contours are chosen and FO declination is applied.
|
|
Automatic speech segmentation using Czech TTS
Horák, Petr
This paper presents automatic phonetic segmentation of natural speech based on the use of a speech synthesizer and dynamic time warping (DTW) algorithm. A speech synthesizer is used to create a synthetic reference speech pattern with phonetic segmentation information (phonemes, diphones, syllables, intonation units etc.). The reference synthetic speech pattern is then used in the alignment process. The segmentation system has been developed for Czech and it uses the Czech TTS system.
|
|
Fundamental frequency and duration modelling by neural networks
Tučková, Jana
This contribution presents the application of a neural network for the fundamental frequency (F.sub.0./sub.) determination and for the modeling of phonemes duration. The minimization of the number of input parameters and the reduction of the structure redundancy is described. The pruning method of artificial neural networks is s suitable tool.
|