Neural models improve distant speaker diarization in noisy conditions
Neural Multichannel Distant Speaker Diarization with Heavy-tailed Source Separation Model
SoundArtificial Intelligence
Summary
Separating and identifying who is speaking in a room with multiple people talking at once and from a distance is hard. The authors improved a method that uses multiple microphones and advanced math models to better separate voices and label who spoke when. They replaced the usual assumptions about voice signal patterns with more flexible ones that better fit real speech. Their experiments show this change reduces mistakes in identifying speakers.
What this means in practice
- •For voice assistant developers: Improve voice assistant accuracy by better distinguishing multiple distant speakers in noisy environments.$Commercial implications: Better speaker separation enables commercial voice assistants to handle overlapping speech and distant voices more reliably.
- •For conference system engineers: Enhance meeting transcription quality by more accurately identifying who spoke when in multi-participant distant microphone setups.
Authors
Sicheng Mao, Baihan Li, Mathieu Fontaine, Anthony Larcher, Roland Badeau
Abstract
Distant speaker diarization remains challenging due to difficult acoustic environments, varying numbers of speakers and overlapping speech. Model-driven methods are proposed to exploit the speech source features in multi-channel recordings that help diarization. This paper generalizes a neural model that jointly learns to perform blind source separation and diarization over speech mixtures (neural FCASA) with heavy-tailed models. The popular Gaussian distribution has been applied for variance modeling in the original source separation model, which we replace with two families of heavy-tailed models (Leptokurtic Generalized Gaussian distribution and Student's t distribution) to better capture the heavy-tailedness in speech signals. Thanks to the Gaussian scale mixture model, we are able to unify the proposed method and the original one under the same form of learning objective. Our experiments show consistent large improvements in Diarization Error Rate (DER) and Jaccard Error Rate (JER) compared to the baseline on various corpora.