Personal audio systems generate a local sound field for a listener while attenuating the sound energy in predefined quiet zones. In practice, system performance is sensitive to errors in the acoustic transfer functions between the sources and the zones. In this paper, a design framework for robust reproduction is proposed that combines transfer functions and error modeling. The framework allows a physical perspective on the regularization required for a system that is based on there being a...

This paper proposes a three-dimensional wave-domain acoustic contrast control method to reproduce a multizone sound field using a circular loudspeaker array. In this method, sound field analysis is based on spherical harmonic decomposition, and the loudspeaker weights are obtained by maximizing the acoustic energy contrast between the predefined bright zone and dark zone. Simulation results show that the proposed method provides good multizone separation performance over a large spatial region and requires lower-order spherical harmonics, resulting in a much lower number of microphones required to measure the acoustic transfer functions.

Automated audio captioning aims to describe audio data with captions using natural language. Existing methods often employ an encoder-decoder structure, where the attention-based decoder (e.g., Transformer decoder) is widely used and achieves state-of-the-art performance. Although this method effectively captures global information within audio data via the self-attention mechanism, it may ignore the event with short time duration, due to its limitation in capturing local information in an audio signal, leading to inaccurate prediction of captions. To address this issue, we propose a method using the pretrained audio neural networks (PANNs) as the encoder and local information assisted attention-free Transformer (LocalAFT) as the decoder. The novelty of our method is in the proposal of the LocalAFT decoder, which allows local information within an audio signal to be captured while retaining the global information. This enables the events of different duration, including short duration, to be captured for more precise caption generation. Experiments show that our method outperforms the state-of-the-art methods in Task 6 of the DCASE 2021 Challenge with the standard attention-based decoder for caption generation.

State-of-the-art audio captioning methods typically use the encoder-decoder structure with pretrained audio neural networks (PANNs) as encoders for feature extraction. However, the convolution operation used in PANNs is limited in capturing the long-time dependencies within an audio signal, thereby leading to potential performance degradation in audio captioning. This letter presents a novel method using graph attention (GraphAC) for encoder-decoder based audio captioning. In the encoder, a graph attention module is introduced after the PANNs to learn contextual association (i.e. the dependency among the audio features over different time frames) through an adjacency graph, and a top- k mask is used to mitigate the interference from noisy nodes. The learnt contextual association leads to a more effective feature representation with feature node aggregation. As a result, the decoder can predict important semantic information about the acoustic scene and events based on the contextual associations learned from the audio signal. Experimental results show that GraphAC outperforms the state-of-the-art methods with PANNs as the encoders, thanks to the incorporation of the graph attention module into the encoder for capturing the long-time dependencies within the audio signal. The source code is available at https://github.com/LittleFlyingSheep/GraphAC.

Personal audio provides private and personalized listening experiences by generating sound zones in a shared space with minimal interference between zones. One challenge of the design is to achieve the best performance with a limited number of microphones and loudspeakers. In this paper, two modal domain methods for personal audio reproduction are compared. One is the spatial harmonic decomposition (SHD) based method and the other is the singular value decomposition (SVD) based method. It is demonstrated that the SVD based method provides a more efficient modal domain decomposition than the SHD method for 2.5 dimensional personal audio design. Simulation results show that the SVD based method outperforms the SHD one by up to 10 dB in terms of acoustic contrast and up to 17 dB in terms of reproduction error for a compact arc array with five loudspeakers, while requiring fewer microphones around the zone boundaries. The SVD based method retains the inherent efficiency of optimizing in a modal domain while avoiding the inherent geometric limitations of using SHD basis functions. Thus, this approach is advantageous for applications with flexible system geometries and a small number of loudspeakers and microphones.