搜索资源列表
chxu
- 本程序用于测量增强语音与纯净语音的对数谱距离，用于语音增强。对数谱距离越小，表明增强语音与纯净语音越接近，增强语音质量越好；反之，则增强语音质量越差。-Calculates the average log-spectral distance between CLEAN and NOISY signals. Frames are 25ms long with 60 percent (15ms) overlap, hamming windowed. RS is the remove
spraal1
- An Automatic Gain Controller (AGC) for speech signals embedded in additive noise requires Voice Activity Detection (VAD) to avoid noise amplification, a peak level detector for computing gain, and a gain controller for adjusting gain. This paper
sift_pitch
- This MATLAB exercise designs and implements a pitch period detector based on detecting and tracking peaks in the autocorrelation of the LPC error signal during regions of voiced speech. The pitch detection procedure is called the SIFT (Simple Inverse
Cepstral_Pitch_Detector
- 探测器基音周期的主要特点是使用二次cepstral峰值检波器的每一帧演讲,为了检测并纠正基音周期检测错误由于效果翻倍基音周期或减半等及相关现象。文件的4.5 Cepstral距探测器。pdf”提供了一个用户的指导对于这个练习。-This MATLAB exercise implements a pitch period detector based on detecting and tracking peaks in the (real) cepstrum during regions of v
lpc_vocoder_rev2
- 这个MATLAB构建一个锻炼LPC声码器,即,执行LPC分析和合成语音文件,导致合成语音近似原始的演讲。LPC分析使用一个标准的自相关分析来确定LPC系数的设置,一帧一帧的基础上,以及框架获得。一个独立的分析方法(cepstral螺距内检测器)把每一帧的言论是要么表示演讲(时间由cepstral峰值的位置在指定范围的音调时期)或无声的言论(模拟随机噪声帧)0帧基音周期的样本。独立的分析提供了一个两国并存的激发函数LPC合成处理的一部分,包括一系列的脉冲(表示帧期间)和/或噪声序列(在无声的帧)。
speech reconstruction+SLP
- This paper proposes a new variant of the least square autoregressive (LSAR) method for speech reconstruction, which can estimate via least squares a segment of missing samples by applying the linear prediction (LP) model of speech. First, we show t