|
|
Speech spectrum smoothing by cepstrum and pseudo-cepstrum weighting
Plšek, Martin
In the paper an experimental vocal tract model of the 6.sup.th.sup. order excited by a synthetic voiced excitation signal with variable fundamental frequency is used as a testing signal for several cepstral and pseudo-cepstral smoothing approaches. The aim of the investigation is to find a reliable method for speech spectrum smoothing convenient also for female and child voices with high fundamental frequency.
|
|
|
Design of FIR vocal tract models with linear and nonlinear phase
Vondra, Martin ; Vích, Robert
Experiences with the IIR cepstral model based on Padé approximation of the exponential function indicate some problems with stability. An experiment in designing a FIR cepstral vocal tract model has been made. The computation of the impulse response of the FIR filter that realizes the vocal tract model can be accomplished either using the window method of digital filter design or by a recursive computation directly from real cepstrum using the minimum-phase approximation.
|
|
| |
|
|
Multi-function Realization of the TTS System
Přibil, Jiří
The research report describes the TTS system realized in the form of the DLL library for Windows 95/98//NT/2k/XP platform and also in the form of the MS ISAPI/NSAPI script for Internet applications. This TTS system works with the synthesis based on LPC, cepstral, and harmonic speech models and on the impulse response of the cepstrum.
|
|
|
Czech-speaking artificial face
Chaloupka, J. ; Nouza, J. ; Přibil, Jiří
This paper describes an artificial face that can speak and articulate Czech. It employs the computer-controlled 3D model of virtual head, know as Baldi, which was developed in USA for communication in English. Here we present methods and algorithms that allowed us to use Baldi with Czech spoken language, both synthesised as well as natural. The methods are based on speech synthesis and speech recognition techniques.
|
|
|
Pseudokepstrální analýza řeči
Vích, Robert
Real generalized cepstral analysis is introduced and applied to speech deconvolution and vocal tract modeling. Finite impulse response pseudo cepstral vocal tract models with linear and minimum phase are derived and applied to speech synthesis.
|
|
|
Konverse hlasu založená na nelineární transformaci spektra
Vích, Robert ; Vondra, Martin
A new voice conversion algorithm is presented, which transforms the utterance of a source speaker into the utterance of a target speaker. The voice conversion algorithm is based on spectral speech analysis, nonlinear frequency transformation, spectral envelope warping, spectrum interpolation and cepstral speech synthesis. Prosodic transformations, i.e. fundamental frequency, time and intensity scale modifications are also shortly mentioned.
|
|
|
Realizace TTS jako systémového ovladače pro Microsoft Speech API
Přibilová, Anna ; Přibil, Jiří
The paper deals with implementation of voice modification in the Czech and Slovak text-to-speech /TTS/ system as Microsoft speech application programming interface /MS SAPI/ TTS engine based on diphone inventory. The original male voice database serves for creation of three derived databases with modification factors given by the average formant and pitch shift between male, female and childish voices. The source-filter model as well as the harmonic speech model, both with cepstral description, were used.
|
|
|
Zpracování řeči
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, enhancement and phonetics.
|
|
|
Inclusion of the Czech and Slovak TTS System to the MS SAPI 5 Standard
Přibil, Jiří
The report is aimed at analysis of structure and basic functions of the Microsoft SAPI 5 standard for Windows platform. A new developed TTS Engine called TTSKubec to be included in our Czech and Slovak TTS system is mainly described. Modifications of current version of the TTS system based on the diphone speech inventory working with the cepstral and the harmonic speech synthesis are also mentioned in the report.
|
guest ::
