基于LMA声道模型的语声合成新方法

A new synthesis method based on the LMA vocal tract model

介绍一种基于LMA（LogMagnitudeApproximate）声道模型的语音合成新方法．采用这种方法，合成系统不但具有更好的合成音质，对韵律参数的调整能力也大大增强．合成实验表明，新的系统在自然度和可懂度上均比以前显著提高．文中首先介绍了LMA滤波器的基本原理和合成器的构成方案，然后详细讨论了用它对韵律参数进行调整的方法．最后，以一个较为成功的PSOLA合成系统KDTALK－1为对比，给出了系统性能的定量评价．

In this paper, a new synthesis algorithm based on the LMA filter in Chinese text-to-speech system is introduced. Using this method, the system can not only produce speech with higher quality, but also have a more powerful ability to modify the prosodic parameters, which ensures a far more natural and intelligible synthesized speech than ever before. Here, we first present the fundamental principles of the LMA filter and the construction of the synthesizer, then describe minutely how to modify the acoustic parameters with this synthesizer, finally, the quatitative evaluation of the system's performance is shown while comparing with a relatively successful PSOLA synthesizer KDTALK- 1.