摘要
介绍了一种更有效的相位内插算法,使正弦模型能够精确而更有效地分析/综合语音信号。将该正弦模型应用于音频的时域修正,通过使用正弦+瞬时冲击+噪声3种模型来灵活表示广泛的音频信号,分别针对不同的信号进行不同的时域修正,实验证明,通过此方法得到的效果明显优于传统时域修正算法。
An effective phase interpolation algorithm is introduced, in which sinusoidal model can analyze/ synthesize speech signal precisely and effectively. The sinusoidal model is applied to time scale of audio, so audio signals are flexibly represented by sinusoidal＋transient＋noise three models, and different time scales are employed according to the signal. Experiment proves that the effect of using the proposed method is significantly better than conventional time scaling algorithm.
出处
《电声技术》
2008年第3期49-52,共4页
Audio Engineering
关键词
正弦
相位内差
时域修正
瞬时冲击
sinusoidal
phase interpolation
time scale
transient
