A Comparative Study of Modular and Joint Approaches for Speaker-Attributed ASR on Monaural Long-Form Audio

TL;DR

This study compares modular and joint approaches for speaker-attributed ASR on real monaural recordings, showing the joint system's superior performance after fine-tuning with limited data.

Contribution

It provides a comprehensive comparison of modular and end-to-end SA-ASR systems on real data, including a new pipeline combining speaker clustering with E2E models.

Findings

Joint system outperforms modular after fine-tuning.

Modular system performs better without fine-tuning.

Error analysis reveals remaining challenges in monaural SA-ASR.

Abstract

Speaker-attributed automatic speech recognition (SA-ASR) is a task to recognize "who spoke what" from multi-talker recordings. An SA-ASR system usually consists of multiple modules such as speech separation, speaker diarization and ASR. On the other hand, considering the joint optimization, an end-to-end (E2E) SA-ASR model has recently been proposed with promising results on simulation data. In this paper, we present our recent study on the comparison of such modular and joint approaches towards SA-ASR on real monaural recordings. We develop state-of-the-art SA-ASR systems for both modular and joint approaches by leveraging large-scale training data, including 75 thousand hours of ASR training data and the VoxCeleb corpus for speaker representation learning. We also propose a new pipeline that performs the E2E SA-ASR model after speaker clustering. Our evaluation on the AMI meeting corpus reveals that after fine-tuning with a small real data, the joint system performs 8.9--29.9% better in accuracy compared to the best modular system while the modular system performs better before such fine-tuning. We also conduct various error analyses to show the remaining issues for the monaural SA-ASR.