Digital signal processing for multichannel audio equalization and signal cancellation.

摘要

This thesis presents new algorithms in important areas of acoustics: (i) multiple listener equalization, and (ii) selective audio signal cancellation.; Equalization of acoustical responses at multiple locations is important for delivery of high quality audio to listeners in an enclosed environment (e.g., theaters, automobiles, etc.). An equalization filter designed to compensate for the room effects at a single location performs poorly at other locations in a room since room responses vary significantly with changing source-receiver positions. A good equalization filter should minimize the effects of the room at multiple listener locations in the room.; The thesis proposes a multiple listener equalization filter using the fuzzy c-means clustering technique. Since the length of the measured room response can be quite large (depending on the room reverberation time), the thesis further shows that, by applying psycho-acoustic techniques, the filter orders can be reduced significantly, for practical Digital Signal Processing (DSP) implementations, without significantly affecting the equalization performance at the multiple listener positions. Finally, it is shown that the proposed filter outperforms existing standard filters in terms of objective measures of equalization performance and subjective listening tests.; Selectively canceling signals at specific locations within an acoustical environment with multiple listeners is of significant importance for home theater, automobile and other applications. The traditional sound absorption/noise cancellation approaches are impractical for such applications because they require either, bulky or uncomfortable ear-defenders to be inserted into the ear canal for source signal absorption, or multiple secondary sources to anti-phase the primary audio source for relatively low frequency acoustic monotone signals, or high volume passive sound absorption materials for relatively high frequencies in the audio band.; In this thesis, an alternative method is presented for signal cancellation, by pre-processing the acoustical signal with a digital filter known as the eigenfilter. The eigenfilter uses information based on the positional variations in the room impulse responses so as to achieve audio signal cancellation. Also examined are the theoretical properties of an eigenfilter, and the performance tradeoffs (e.g., spectral distortion and variation of room impulse responses).; Conclusions and future directions are provided appropriately at the end of relevant chapters.