论文部分内容阅读
众所周知,每个人的声调基频是不同的,甚至同一个人在不同情绪下,其声调基频也存在较大的差别。这种差别对于非特定人识别系统的识别率带来了很大的影响,为了减少这种影响,必须采用归一化算法对基频进行处理。文章在几种常用的归一化算法的基础上,提出了一种改进的归一化算法,并对该算法与几种常用算法在相同实验条件下的识别率进行了比较。实验结果表明,该算法有效地提高了声调识别系统的识别率,具有一定的实用价值。
As we all know, each person’s tone fundamental frequency is different, and even the same person in different emotions, the pitch fundamental frequency there is a big difference. This difference has a significant impact on the recognition rate of non-specific person recognition systems. In order to reduce this effect, the normalization algorithm must be used to process the fundamental frequency. Based on several commonly used normalized algorithms, this paper proposes an improved normalization algorithm, and compares the recognition rate of the algorithm with several commonly used algorithms under the same experimental conditions. The experimental results show that the algorithm effectively improves the recognition rate of tone recognition system, and has certain practical value.