Microsoft Speaker Diarization System for the VoxCeleb Speaker Recognition Challenge 2020
This work addresses the problem of accurately identifying and segmenting speakers in noisy, real-life audio for applications like transcription and analysis, representing an incremental improvement in a specific domain.
The paper tackles speaker diarization in real-world multi-talker audio recordings, achieving a diarization error rate of 3.71% on the development set and 6.23% on the evaluation set, ranking first in the VoxSRC 2020 challenge.
This paper describes the Microsoft speaker diarization system for monaural multi-talker recordings in the wild, evaluated at the diarization track of the VoxCeleb Speaker Recognition Challenge(VoxSRC) 2020. We will first explain our system design to address issues in handling real multi-talker recordings. We then present the details of the components, which include Res2Net-based speaker embedding extractor, conformer-based continuous speech separation with leakage filtering, and a modified DOVER (short for Diarization Output Voting Error Reduction) method for system fusion. We evaluate the systems with the data set provided by VoxSRCchallenge 2020, which contains real-life multi-talker audio collected from YouTube. Our best system achieves 3.71% and 6.23% of the diarization error rate (DER) on development set and evaluation set, respectively, being ranked the 1st at the diarization track of the challenge.