Speech signals are specific and different from signals used in technical area due to the fact that correspondence of speech signal characteristics to the human being perception is not trivial and fully examined. Additionally lots of existing languages and dialects introduce certain differences in speech signals. Thus concrete speech signal analysis involve various even different methods and techniques which must be used together. However, most modern speech analysis systems are specialized for solving specific scientific problems and are limited in number of applicable methods and determinable characteristics. Extensibility in addition of new characteristics or methods to these systems is problematic and often impossible, too.
The system (further SSCIS - Speech Signal Characteristics Investigation System)
described in this paper is dedicated to speech signal analysis and is free of
mentioned limitations. It consists of software package working under Windows and
any standard sound board. SSCIS lets:
a) to perform visual, audible, correlation, Fourier, cepstrum, linear predictive
and inverse filtration analysis and speech synthesis;
b) to determine such speech signal characteristics as intensity, long and short
time spectrograms, formants, pitch period, linear predictive coefficients;
c) to perform speech signal filtering with linear, adaptive and neural filters;
d) all kind measurements of signals and characteristics: peak amplitude, peak
position, peak duration, duration among peaks, etc.;
e) all kind of signal and characteristics editing: cut, copy, shrink, expand,
interpolate, decimate, etc.
Numerous methods of speech signal characteristics calculation could be easily
selected. At the same time user has possibility to create, use and include in
package his own algorithms.
Material in this paper is organized in a such way. In Section 2 we discuss drawbacks of existing systems and desirable features, operation modes of the system under developing. Acquaintance with already implemented SSCIS functions and their calculation methods are given in Section 3. Section 4 contains a description of SSCIS hardware requirements, system structure and user interface. Finally, in Section 5 discussion and conclusions about system extension possibilities and applicability are presented.