Publication detail
Target Speaker ASR with Whisper
POLOK, A. KLEMENT, D. WIESNER, M. KHUDANPUR, S. ČERNOCKÝ, J. BURGET, L.
Original Title
Target Speaker ASR with Whisper
Type
article in a collection out of WoS and Scopus
Language
English
Original Abstract
We propose a novel approach to enable the use of large, single-speaker ASR models, such as Whisper, for target speaker ASR. The key claim of this method is that it is much easier to model relative differences among speakers by learning to condition on frame-level diarization outputs than to learn the space of all speaker embeddings. We find that adding even a single bias term per diarization output type before the first transformer block can transform single-speaker ASR models into target-speaker ASR models. Our approach also supports speaker-attributed ASR by sequentially generating transcripts for each speaker in a diarization output. This simplified method outperforms baseline speech separation and diarization cascade by 12.9% absolute ORC-WER on the NOTSOFAR-1 dataset.
Keywords
target-speaker ASR, diarization conditioning, multi-speaker ASR, Whisper
Authors
POLOK, A.; KLEMENT, D.; WIESNER, M.; KHUDANPUR, S.; ČERNOCKÝ, J.; BURGET, L.
Released
6. 5. 2025
Publisher
IEEE Biometric Council
Location
Hyderabad
ISBN
979-8-3503-6874-1
Book
Proceedings of ICASSP 2025
Pages from
1
Pages to
5
Pages count
5
URL
BibTex
@inproceedings{BUT198049,
author="Alexander {Polok} and Dominik {Klement} and Matthew {Wiesner} and Sanjeev {Khudanpur} and Jan {Černocký} and Lukáš {Burget}",
title="Target Speaker ASR with Whisper",
booktitle="Proceedings of ICASSP 2025",
year="2025",
pages="1--5",
publisher="IEEE Biometric Council",
address="Hyderabad",
doi="10.1109/ICASSP49660.2025.10887683",
isbn="979-8-3503-6874-1",
url="https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=10887683"
}Documents