Speech enhancement algorithm based on variable speech spectral amplitude distribution (信号処理)

概要

論文の詳細を見る
In this paper, we propose a speech enhancement algorithm based on the MAP estimation using variable speech spectral amplitude distribution, under the assumption that the speech spectral amplitude distributions change in various SNR environments. The proposed algorithm achieves that the speech spectral distribution approaches to Rayleigh distribution to keep the quality of the enhanced speech in speech segments. In non-speech segments, it approaches to a exponential function to reduce noise effectively. The proposed speech enhancer based on the variable speech spectral amplitude distribution can extract a high quality speech signal from a noisy speech. Simulation results show the effectiveness of the proposed speech enhancer.
社団法人電子情報通信学会の論文
2009-01-15