基于频段波形相关的自适应宽带信号基音搜索方法

An Adaptive Bandwidth Pitch Searching Method in Coding Wideband Signals Based on Waveform Cross Correlation

本文针对谱范围扩展后的宽带语音信号长时相关谐波结构在高频段并不显著的特点,提出一种改进的长时预测算法.本算法是基于频段波形相关的自适应带宽选择方法,利用宽带信号频段间的相关度判决对长时预测的激励信号进行选择性低通滤波,改进了传统方法的不足.与3GPP的自适应多码率宽带语音编码标准(AMR-WB)中的长时预测算法相比,新算法的运算量降至54%,且实际编码的主客观结果表明,无论是感知信噪比还是最终音质都证明了新算法的有效性.

The pitch contents in the specn-um of voiced segments in wideband signals do not extend over the whole spectrum range,and the amount of voicing shows more variation compared to narrow-band signals. Thus this paper presents a new paradigm to improve the closed-loop pitch analysis to better accommodate the variations in the voicing level. The paradigm is an adaptive pitch searching paradigm which automatically selects an optimal low pass filter applied to the past excitation based on the inter-band waveform cross correlation. The complexity of new algorithm is only 54 % of that of adaptive multi-rate wideband （AMR-WB） speech codec selected by 3GPP. We verify that the performance, hoth in terms of weighted signal-to-noise ratio （SNR） and perceptual quality, is higher as apposed to that of AMR-WB.