arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 1206.2437cs.CV
收藏

一种新的用于说话人识别的MFCC加窗算法

A Novel Windowing Technique for Efficient Computation of MFCC for Speaker Recognition

Indian Institute of Technology Kharagpur(印度理工学院克勒格布尔分校)

Md. Sahidullah, Goutam Saha

摘要:

在本文中,我们提出了一个新的家庭窗口技术来计算梅尔频率倒谱系数(MFCC)的自动说话人识别从语音。该方法是基于离散时间傅里叶变换(DTFT)的基本性质有关的频域微分。对经典的加窗方法如Hamming窗进行了改进,得到了离散时间傅里叶变换系数的导数。它已被数学证明,斜率和相位的功率谱固有地纳入新计算的倒谱。说话人识别系统的基础上,我们提出的家庭的窗口函数的基础上,以达到实质性的和一致的性能改善基线单锥形汉明窗口,以及最近提出的multitaper窗口技术。

英文摘要:

In this paper, we propose a novel family of windowing technique to compute Mel Frequency Cepstral Coefficient (MFCC) for automatic speaker recognition from speech. The proposed method is based on fundamental property of discrete time Fourier transform (DTFT) related to differentiation in frequency domain. Classical windowing scheme such as Hamming window is modified to obtain derivatives of discrete time Fourier transform coefficients. It has been mathematically shown that the slope and phase of power spectrum are inherently incorporated in newly computed cepstrum. Speaker recognition systems based on our proposed family of window functions are shown to attain substantial and consistent performance improvement over baseline single tapered Hamming window as well as recently proposed multitaper windowing technique.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑