joint sparse representation (JSR)方法用于车内语音增强的特征降噪算法-address reducing the mismatch between training and testing conditions for hands-free in-car speech recognition. It is well known that the distortions caused by background noise, channel effec