DocumentCode :
2005133
Title :
An improved spectral subtraction method in compressed domain
Author :
Liang Liu ; Xiaoqing Yu ; Wengen Wang ; Yunhui Wang ; Haifeng Yu
Author_Institution :
Sch. of Commun. & Inf. Eng., Shanghai Univ., Shanghai, China
fYear :
2011
fDate :
14-16 Nov. 2011
Firstpage :
294
Lastpage :
298
Abstract :
For audio noise reduction, the current research is mainly focused on the non compressed domain. As for the MP3 (MPEGl-layer3) compressed audio, the common noise reduction method is first fully decoding and then for the decoding audio carrying out noise reduction. This method costs a lot of time for decoding and it inevitably reduces the processing efficiency for compressed audio. In this paper, we present an improved spectral subtraction method based on MDCT-entropy activity detection method for noise estimation in compressed domain. The proposed algorithm works directly on MP3 frames. The experimental results show that the proposed algorithm can remove noise directly from noisy MP3 audio and its process efficiency is higher than common method. And the signal-to-noise ratio (SNR) of the compressed audio is improved obviously.
Keywords :
audio coding; data compression; decoding; media streaming; MDCT-entropy activity detection method; MP3 compressed audio; MPEGl-layer3 compressed audio; SNR; audio noise reduction; common noise reduction method; decoding; improved spectral subtraction method; noise reduction; signal-to-noise ratio; Compressed domain; MDCT entropy; Spectral subtraction;
fLanguage :
English
Publisher :
iet
Conference_Titel :
Wireless Mobile and Computing (CCWMC 2011), IET International Communication Conference on
Conference_Location :
Shanghai
Type :
conf
DOI :
10.1049/cp.2011.0895
Filename :
6194852
Link To Document :
بازگشت