Blizzard Challenge 2027: Accented Zero-Shot TTS in Spontaneous Portuguese Speech (Training Material)
This report details the training material for the Blizzard Challenge 2027, which focuses on accented zero-shot Text-to-Speech (TTS) in spontaneous Portuguese. We describe three primary datasets: NURC-SP (São Paulo), NURC-RE (Recife), and FalAR-TTS (European Portuguese). More Info: https://nilc-nlp.github.io/ssw-2027/blizzard-challenge-dates.html and https://nilc-nlp.github.io/ssw-2027/blizzard-challenge-rules.html
Authors
- Alberto Abad (ORCID: https://orcid.org/0000-0003-2122-5148)
- Marli Quadros Leite (ORCID: https://orcid.org/0000-0002-8417-0140)
- Miguel Oliveira (ORCID: https://orcid.org/0000-0002-0866-0535)
- Sidney Evaldo Leal (ORCID: https://orcid.org/0000-0002-8817-2063)
- Flaviane Romani Fernandes Svartman (ORCID: https://orcid.org/0000-0002-9941-3934)
- Sérgio Paulo
Institutions
- Instituto de Engenharia de Sistemas e Computadores Investigação e Desenvolvimento (PT)
- Universidade de São Paulo (BR)
- Institute for Systems Engineering and Computers (PT)
- Instituto Superior Técnico (PT)
- Universidade Federal de Alagoas (BR)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-06-27
- DOI
- https://doi.org/10.5281/zenodo.20975209
- Primary Topic
- Natural Language Processing Techniques
- Type
- article
- Field-Weighted Citation Impact
- 0.00