원문정보
초록
영어
In order to verify the integrity of the original voice and determine the voice match, researches on analyzing speakers' own voice have been undergoing. However, spectrogram and waveform analysis technologies that are currently used have issues of lack of accuracy and reliability when determining the voice match. In this paper, ways of determining and performing voice match are suggested and proven by confirming the original voice frequency form of each individual with voice analysis methods using spectral contents. Also, the circuit for correcting or improving the quality of voice is designed and proposed. When the form of voice recording signals was analog, improved voice analysis data was secured by eliminating unnecessary noise and increasing property during PCM process. Thus, portability of voice analysis more enhanced than existing research results was verified and studied.
목차
1. Introduction
2. System Design of Spectral Contents Voice Analysis
2.1 Spectral Contents Audio Analysis Methods
2.2 System Design of the Spectral Content Speech Analysis
2.3 Detailed Configuration of the Speech Spectral Content Analysis System
3. The Experiment of Voice Matches Analysis and Signal Enhancement of Spectral Contents
3.1 Settings of Spectral Contents Analyzing System
3.2 Experiments on Spectral Contents Voice Matching
3.3 Using the Spectral Content ‘a, e, i, o, u’ Pronunciation Analysis
4. Discussion and Conclusions
References