CoChat
Sign Up
Cited by 164Speech Recognition and SynthesisSpeech and Audio ProcessingVoice and Speech Disorders

Zero-Shot Multi-Speaker Text-To-Speech with State-Of-The-Art Neural Speaker Embeddings

Cooper, Erica, Lai, Cheng-I, Yasuda, Yusuke, Fang, Fuming, Wang, Xin, Chen, Nanxin, Yamagishi, Junichi·2020·ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
DOI 10.1109/icassp40776.2020.9054535
Open this paper in CoChat — free account

Abstract

While speaker adaptation for end-to-end speech synthesis using speaker embeddings can produce good speaker similarity for speakers seen during training, there remains a gap for zero-shot adaptation to unseen speakers. We investigate multi-speaker modeling for end-to-end text-to-speech synthesis and study the effects of different types of state-of-the-art neural speaker embeddings on speaker similarity for unseen speakers. Learnable dictionary encoding-based speaker embeddings with angular softmax loss can improve equal error rates over x-vectors in a speaker verification task; these embeddings also improve speaker similarity and naturalness for unseen speakers when used for zero-shot adaptation to new speakers in end-to-end speech synthesis.

Integrity signals

Registry

Crossref · 10.1109/icassp40776.2020.9054535

Cites

3 of 3 references verified0 citation errors · 0 not checkable

Statistics

Recomputing p-values needs the paper’s full text.

Claims the paper makes

This paper’s full text isn’t openly available, so its claims and self-check weren’t run. The registry and citation checks still apply.

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • TTS and VC technology has advanced rapidly in recent years, especially in requiring far less target speaker training data.

    “TTS and VC technology has advanced rapidly in recent years, especially in terms of requiring far less target speaker training data [53][55].”

    From ASVspoof 2021: Towards Spoofed and Deepfake Speech Detection in the Wild (2023) · cites this paper as [53]

  • Some TTS/VC systems operate without fine-tuning and generate speech in the voice of a specific target speaker using a single speaker embedding (zero-shot voice cloning).

    “Some systems even operate without fine-tuning and, instead, generate speech in the voice of a specific target speaker by using a single speaker embedding (Jia et al., 2018; Arik et al., 2018; Cooper et al., 2020; Wu et al., 2022), i.e. zero-shot voice cloning.”

    From ASVspoof 5: Design, collection and validation of resources for spoofing, deepfake, and adversarial attack detection using crowdsourced speech (2026) · cites this paper as Cooper et al., 2020

Cites

Flagged first, then by citations

Cited by

Sorted by citations

Authors
Cooper, E., Lai, C.-I., Yasuda, Y., Fang, F., Wang, X., Chen, N., & Yamagishi, J.
Affiliations
  • National Institute of Informatics (JP)
  • Massachusetts Institute of Technology (US)
  • Johns Hopkins University (US)
ORCID
Venue
ICASSP 2020 - 2020 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)
Publisher
IEEE
Published
2020-05
Pages
6184-6188
ISSN
1520-6149 (ISSN-L)
DOI
10.1109/icassp40776.2020.9054535
OpenAlex
W3015826515
Cited by
164
References
3
Citations by year
2026: 3 · 2025: 18 · 2024: 27 · 2023: 31 · 2022: 42 · 2021: 38 · 2020: 5
FWCI
11.08
Type
conference paper
Language
en
Open access
closed
Licence
Locations
IEEE International Conference on Acoustics Speech and Signal Processing (publishedVersion)
Funders
Japan Society for the Promotion of Science (19K24372, 17H04687)
Topics
  • Physical Sciences › Computer Science › Artificial Intelligence › Speech Recognition and Synthesis
  • Physical Sciences › Computer Science › Signal Processing › Speech and Audio Processing
  • Health Sciences › Medicine › Physiology › Voice and Speech Disorders
Keywords
multi-speaker text-to-speech, end-to-end speech synthesis, speaker similarity, unseen speakers, dictionary encoding, x-vector, speaker verification, naturalness

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review