Mind the Microphone Gap: Benchmarking Array Upsampling Strategies for Latent Acoustic Mapping
2026-07-27 • Sound
Sound
AI summaryⓘ
The authors studied a method called Latent Acoustic Mapping (LAM) that makes detailed sound maps from many microphones without needing labeled data. They found LAM works well with lots of microphones but struggles when using only a few (like 4). To fix this, they tested different ways to increase the resolution of the sound data before using LAM. Their tests showed the best results happen when the upscaling method is trained separately but aligned well with LAM, and simpler models often worked better than very complex ones.
Latent Acoustic Mappingself-supervised learningacoustic mappingcross-spectral matrixupsamplingconvolutional networksgenerative adversarial networksdirection-of-arrivalspatial resolutionmultichannel recordings
Authors
Philipp Schmidt, Huw Cheston, Juan Azcarreta, Adrian Stepien, Çağdaş Bilen, Iran R. Roman
Abstract
Latent Acoustic Mapping (LAM) is a self-supervised learning method that generates high-resolution spherical acoustic maps from multichannel recordings without labelled data, matching supervised baselines on direction-of-arrival benchmarks. However, LAM degrades significantly with sparse 4-channel arrays, as the low-resolution cross-spectral matrix captures far less spatial information than the 32-channel inputs LAM was designed for. We benchmark a diverse set of upsampling architectures, spanning lightweight convolutional networks, iterative back-projection models, physics-informed networks, and generative adversarial approaches. We also study whether aligning these upsamplers with LAM by training them jointly or in different stages helps preserve the spatial structure that LAM depends on. Results show that the original full-resolution LAM is the strongest, that separately trained lightweight models are the most competitive learned approaches, and that representation alignment between the upsampler and LAM matters more than model complexity.