作者: Tobias Friedrich , Matthias Gruhne , Gerald Schuller
DOI: 10.1109/ICASSP.2008.4517585
关键词:
摘要: We describe an efficient conversion method, which directly converts a desired spectral representation from compressed audio material. The method provides feature extraction algorithm with suitable complex frequency of signal. presented allows us to trade off computational complexity accuracy. then test several operating points MPEG system. That leads, in general, reduction the O (N log N) O(N), compared conventional first decoding and applying DFT resulting time domain