1 August 2021 in The Journal of the Acoustical Society of America
We present a method for computing a spherical harmonic representation of a sound field based on observations of the sound pressure along the equator of a rigid spherical scatterer. Our proposed solution assumes that the captured sound field is height invariant so that it can be represented by a two-dimensional (2D) plane wave decomposition (PWD). The 2D PWD is embedded in a three-dimensional representation of the sound field, which allows for perfectly undoing the effect of the spherical scattering object. If the assumption of height invariance is fulfilled, then the proposed solution is at least as accurate as a conventional spherical microphone array of the same spherical harmonic order, which requires a multiple of the number of sensors. Our targeted application is binaural rendering of the captured sound field. We demonstrate by analyzing the binaural output signals that violations of the assumptions that the solution is based on—particularly height invariance and consequently also horizontal propagation—lead to errors of moderate magnitude.
27 September 2022 in IEEE/ACM Transactions on Audio, Speech, and Language Processing
Spherical harmonic (SH) representations of sound fields are usually obtained from microphone arrays with rigid spherical baffles whereby the microphones are distributed over the entire surface of the baffle. We present a method that overcomes the requirement for the baffle to be spherical. Furthermore, the microphones can be placed along a circumferential contour around the baffle. This greatly reduces the required number of microphones for a given spatial resolution compared to conventional spherical arrays. Our method is based on the analytical solution for SH decomposition based on observations along the equator of a rigid sphere that we presented recently. It incorporates a calibration stage in which the microphone signals due to a suitable set of calibration sound fields are projected onto the SH decomposition of those same sound fields on the surface of a notional rigid sphere by means of a linear filtering operation. The filter coefficients are computed from the calibration data via a least/squares fit. We present an evaluation of the method based on the application of binaural rendering of the SH decomposition of the signals from an 18/element array that uses a human head as the baffle and that provides 8th ambisonic order. We analyse the accuracy and robustness of our method based on simulated data as well as based on measured data from a prototype.
5 September 2024 in IEEE/ACM Transactions on Audio, Speech, and Language Processing
Smart glasses are increasingly recognized as a key medium for augmented reality, offering a hands-free platform with integrated microphones and non-ear-occluding loudspeakers to seamlessly mix virtual sound sources into the real-world acoustic scene. To convincingly integrate virtual sound sources, the room acoustic rendering of the virtual sources must match the real-world acoustics. Information about a user's acoustic environment however is typically not available. This work uses a microphone array in a pair of smart glasses to blindly identify binaural room impulse responses (BRIRs) from a few seconds of speech in the real-world environment. The proposed method uses dereverberation and beamforming to generate a pseudo reference signal that is used by a multichannel Wiener filter to estimate room impulse responses which are then converted to BRIRs. The multichannel room impulse responses can be used to estimate room acoustic parameters which is shown to outperform baseline algorithms in the estimation of reverberation time and direct-to-reverberant energy ratio. Results from a listening experiment further indicate that the estimated BRIRs often reproduce the real-world room acoustics perceptually more convincingly than measured BRIRs from other rooms of similar size.
30 January 2023 in IEEE/ACM Transactions on Audio, Speech, and Language Processing
Psychoacoustic experiments have shown that directional properties of the direct sound, salient reflections, and the late reverberation of an acoustic room response can have a distinct influence on the auditory perception of a given room. Spatial room impulse responses (SRIRs) capture those properties and thus are used for direction-dependent room acoustic analysis and virtual acoustic rendering. This work proposes a subspace method that decomposes SRIRs into a direct part, which comprises the direct sound and the salient reflections, and a residual, to facilitate enhanced analysis and rendering methods by providing individual access to these components. The proposed method is based on the generalized singular value decomposition and interprets the residual as noise that is to be separated from the other components of the reverberation. Large generalized singular values are attributed to the direct part, which is then obtained as a low-rank approximation of the SRIR. By advancing from the end of the SRIR toward the beginning while iteratively updating the residual estimate, the method adapts to spatio-temporal variations of the residual. The method is evaluated using a spatio-spectral error measure and simulated SRIRs of different rooms, microphone arrays, and ratios of direct sound to residual energy. The proposed method creates lower errors than existing approaches in all tested scenarios, including a scenario with two simultaneous reflections. A case study with measured SRIRs shows the applicability of the method under real-world acoustic conditions. A reference implementation is provided.
19 November 2021 in IEEE/ACM Transactions on Audio, Speech, and Language Processing
Additive noise produced by the recording hardware will contribute to streamed signals from spherical microphone arrays under practical conditions. For the application of binaural reproduction and under the assumption that the noise is uncorrelated between the array channels, the spectral properties and the overall level of the rendered noise in the ear signals have been shown to be strongly influenced by the configuration of the array as well as of the processing pipeline. In a previous investigation, we determined the audibility thresholds for changes in the rendered noise due to listener head rotations as a function of the differences in noise level of individual array channels. In this article, we calibrate the instrumental metric of Composite Loudness Level to the perceptual data and predict audibility of changes in the additive noise due to head rotations for a broad set of array configurations and distributions of the noise levels across array channels. We demonstrate that some types of microphone layouts can produce audible variations even if the noise level is equal in all channels. This is particularly the case for sampling grids that exhibit negative quadrature weights such as the Lebedev and Fliege-Maier grids for some spherical harmonic orders. The analysis of configurations with unevenly distributed noise contributions show that the influence of the noise from individual array channels is determined by the proximity of their virtual location to the relative trajectory of the ears.