Trends in adaptive MISO system identification for multichannel audio reproduction and speech communication

Abstract

Online identification of multiple-input / single-output (MISO) acoustic systems is one of the long-standing and continuing challenges in multichannel speech and audio applications. Fast and robust estimation of the impulse response of an acoustic system is a key requirement for several adaptive solutions in time-varying scenarios, such as stereophonic acoustic echo cancellation, room equalization, or crosstalk cancellation. The inevitable presence of cross-correlated loudspeaker signals that is implied by multichannel applications, however, entails the well-known non-uniqueness problem of MISO system identification. Apart from this fundamental issue, a more practical problem already consists in the lack of techniques to evaluate the estimated impulse responses properly. Since well-established measures are often not capable of accounting for all aspects of online MISO system identification, we revert to the recently proposed spectral-importance weighted misalignment (SIWM) to assess MISO identification. In this contribution, we review SIWM and its relation to well-established evaluation tools. On this basis, we provide an insight into the problem of MISO system identification in applications driven by real stereo data. We also analyze and compare a traditional and a very recent approach to deal with the non-uniqueness problem.

The generalized frequency-domain adaptive filtering algorithm as an approximation of the block recursive least-squares algorithm

“…To overcome this obstacle, it was proposed in [39] to invert a sparse approximation ofŜ(n) rather than inverting W H 10Ŝ (n)W 10 in (48).…”

“…Note that the actual microphone signal was only used for the adaptation of the filters, while the noise-free microphone signal was used to determine the ERLE. This measure was termed "true ERLE" in [48] and allows to assess the echo cancellation performance also during perturbation. On the other hand, the NMA is defined by…”

Steady-State Performance of an Adaptive Combined MISO Filter Using the Multichannel Affine Projection Algorithm

“…In particular, the combined structure that we take into account is composed of a convex combination of two adaptive multiple-input single-output (MISO) filters [8]. MISO filters are widely used in several applications, such as adaptive beamforming [21,22], OFDM systems [23], channel equalization [24], and audio and speech applications [25,26]. These have been chosen as a baseline scheme of the multichannel case, but the combined MISO structure, and the related analysis thereof, can also be easily extended to multiple-input multiple-output (MIMO) filters, as in [11], since an MIMO filter can be seen as a parallel of MISO filters.…”