Direction of Arrival With One Microphone, a Few LEGOs, and Non-Negative Matrix Factorization

Abstract

Conventional approaches to sound source localization require at least two microphones. It is known, however, that people with unilateral hearing loss can also localize sounds. Monaural localization is possible thanks to the scattering by the head, though it hinges on learning the spectra of the various sources. We take inspiration from this human ability to propose algorithms for accurate sound source localization using a single microphone embedded in an arbitrary scattering structure. The structure modifies the frequency response of the microphone in a direction-dependent way giving each direction a signature. While knowing those signatures is sufficient to localize sources of white noise, localizing speech is much more challenging: it is an ill-posed inverse problem which we regularize by prior knowledge in the form of learned non-negative dictionaries. We demonstrate a monaural speech localization algorithm based on non-negative matrix factorization that does not depend on sophisticated, designed scatterers. In fact, we show experimental results with ad hoc scatterers made of LEGO bricks. Even with these rudimentary structures we can accurately localize arbitrary speakers; that is, we do not need to learn the dictionary for the particular speaker to be localized. Finally, we discuss multisource localization and the related limitations of our approach.

Owlet

“…In [64], a small vertical wall of varying shape is placed next to a microphone which changes the frequency response for different directions of sound. A few recent works [17,18] place small structures like legos and cubes around a microphone to produce scattering. These works on monaural localization either keep a dictionary of possible source models or predict the source model before estimating the DoA.…”

Direction-of-arrival and power spectral density estimation using a single directional microphone

“…Although single-microphone source localization has already been performed by employing machine learning and scattering structures [4], [5], as well as by exploiting the Doppler effect obtained from constant circular motion [6], single-channel spatial audio analysis based on movements of a directional microphone remains a largely unexplored problem. Therefore, the method proposed in this paper is developed under a set of considerably strong assumptions, mainly of spatiotemporal stationarity of uncorrelated sources, anechoic conditions and controlled microphone movements.…”

Sound Localization and Separation in 3D Space Using a Single Microphone with a Metamaterial Enclosure

“…First, a learning mechanism should be introduced into the system, which is embodied in the reconstruction algorithm, such as compressive sensing and Hidden Markov Model . Second, an anisotropic scattering mechanism should be designed in order to provide monaural cues, such as placing irregular scatters surrounding the microphone . Among the numerous approaches in recent years, acoustic metamaterials have served as an optimal choice for use in designing an anisotropic scatter mechanism.…”

“…Researchers have attempted to replicate these bionic features in artificial listening systems in an effort to achieve monaural sound localization and separation . To design such a system, one has to deal with two important issues.…”

“…According to bionics, people attempt to demonstrate these features on the artificial listening systems to realize monaural localization and separation. Previous works have sought to address these challenges by mimicking the listening capability of human auditory systems 14,[18][19][20][21] . To design such a system, one has to deal with two important issues.…”