Publications-Detail

Noise-Shaping for Closed-Loop Multi-Channel Linear Prediction

Authors:
Koep, N. ,  Schäfer, M.Vary, P.
Book Title:
Proceedings of IEEE International Conference on Acoustics, Speech, and Signal Processing (ICASSP)
Organization:
IEEE
Pages:
p.p. 4335 - 4339
Date:
Apr. 2015
DOI:
10.1109/ICASSP.2015.7178789
Language:
English

Abstract

In this paper, a novel noise-shaping method for Multi-Channel Linear Prediction (MCLP) is presented. Without special consideration, the quantization noise of the prediction error poses a serious problem in multi-channel prediction as each noise component distorts the reconstruction of every channel at the decoder. The proposed method diagonalizes the system's quantization error transfer function and thereby limits the influence of quantization noise terms to their respective channels at the decoder. Similar to single-channel linear prediction, a new noise-shaping filter is used to control the trade-off between the objective Signal-to-Noise Ratio
(SNR) and psychoacoustic masking of quantization noise spectra according to the spectral envelope of the predicted signal. It is experimentally shown that the proposed method considerably increases the SNR compared to the open-loop case at the expense of a slight reduction of the effective prediction gain.

Download

BibTeX