论文部分内容阅读
提出了一种判别语音静、清、浊三态S/U/V的新方法———语音波形图像模板匹配法。它的基本原理是首先利用语音信号的波形图 ,导出静、清、浊三态的标准矢量模板 ,然后将输入信号波形转换成矢量图 ,与标准模板进行比较。此法不仅简单、高效 ,而且对于过渡帧的判别精度高于其它一些判别法 ,并具有一定的抗突发噪音性能。
A new method of judging speech static, clear and turbidity tristate S / U / V is put forward, that is speech waveform image template matching method. Its basic principle is to first use the waveform of speech signal to derive the static, clear and turbid three-phase standard vector template, and then convert the input signal waveform into vector and compare with the standard template. This method is not only simple and efficient, but also has higher discrimination accuracy for transitional frames than some other discriminant methods and has certain anti-burst noise performance.