arXiv subjects

Peter Jax

Publications and source records attributed to Peter Jax.

4 recordsLinked to original sources

arXiv2026

Sensing Bone-Conducted Speech with Earbuds

Clear capture of the wearer's own voice (OV) is essential when using earbuds for mobile communication. However, OV capture remains challenging in noisy environments. Bone-conducted (BC) speech, which can be sensed as vibrations of the earbud housing, can be used to improve OV capture. However, neither bandwidth nor spatial characteristics of OV-induced earbud vibrations have been analyzed in detail, despite both characteristics being relevant, e.g., for sensor choice and placement. This study investigates both characteristics, based on measurements with two earbud models. Spectrally, results indicate that OV-induced earbud vibrations exhibit a low-pass characteristic, with a steep roll-off of -93 dB per decade above 400 Hz. Thus, sensors with comparatively low noise floors are required to sense the vibrations above \SI{1}{\kilo\hertz}. Spatially, results indicate that the earbuds mainly vibrate in and out of the ear canal entrance, with high consistency between subjects and fits. Simulations confirm that this enables capture of the high-power vibrations below 400 Hz by a single-axis sensor with less than 1.5 dB mean attenuation.

eess.AS↗

arXiv2025

Data-Driven Uncertainty Modeling for Robust Feedback Active Noise Control in Headphones

Active noise control (ANC) has become popular for reducing noise and thus enhancing user comfort in headphones. While feedback control offers an effective way to implement ANC, it is restricted by uncertainty of the controlled system that arises, e.g., from differing wearing situations. Widely used unstructured models which capture these variations tend to overestimate the uncertainty and thus restrict ANC performance. As a remedy, this work explores uncertainty models that provide a more accurate fit to the observed variations in order to improve ANC performance for over-ear and in-ear headphones. We describe the controller optimization based on these models and implement an ANC prototype to compare the performances associated with conventional and proposed modeling approaches. Extensive measurements with human wearers confirm the robustness and indicate a performance improvement over conventional methods. The results allow to safely increase the active attenuation of ANC headphones by several decibels.

eess.SY↗

arXiv2021

Towards Faster Continuous Multi-Channel HRTF Measurements Based on Learning System Models

Measuring personal head-related transfer functions (HRTFs) is essential in binaural audio. Personal HRTFs are not only required for binaural rendering and for loudspeaker-based binaural reproduction using crosstalk cancellation, but they also serve as a basis for data-driven HRTF individualization techniques and psychoacoustic experiments. Although many attempts have been made to expedite HRTF measurements, the rotational velocities in today's measurement systems remain lower than those in natural head movements. To cope with faster rotations, we present a novel continuous HRTF measurement method. This method estimates the HRTFs offline using a Kalman smoother and learns state-space parameters, including the system model, on short signal segments, utilizing the expectation maximization algorithm. We evaluated our method in simulated single-channel and multi-channel measurements using a rigid sphere HRTF model. Comparing with conventional methods, we found that the system distances are improved by up to 30 dB.

eess.AS↗

arXiv2021

Visualization of Linear Operations in the Spherical Harmonics Domain

Linear operations on coefficients in the spherical harmonics (SH) transform domain that again yield SH-domain coefficients are an important toolset in many disciplines of research and engineering. They comprise rotations, spatially selective filters, and many other modifications for various applications, or describe the response of a MIMO system to an excitation. It is of particular importance to characterize these operations both qualitatively and quantitatively, and make them accessible for people to work with. In this paper, we identify different key properties of such operations and propose a method for their visualization. With our proposed method, we succeed to show many important aspects of an operation in a single plot and give rise to a comprehensive interpretation of the behavior of a system. In our evaluation, we show the potential of the proposed method on the basis of various practical examples from spatial audio signal processing, where SH-domain filtering is used to modify acoustic scenes given by higher-order Ambisonics signals.

eess.AS↗