@inproceedings{araujo-etal-2026-certas,
title = "Certas Palavras: A 1980s-90s {B}razilian Radio Corpus to Test {TTS} Models in Noisy Multi-Speaker Dialogues",
author = "Ara{\'u}jo, Gustavo Evangelista and
Ponti, Moacir and
Junior, Arnaldo Candido and
Leal, Sidney and
Casanova, Edresson and
Silva, Renato Moraes and
Jr., Miguel Oliveira and
Santos, Adriana Barbosa and
Lopes, Gustavo Wadas and
Aluisio, Sandra",
editor = "Souza, Marlo and
de-Dios-Flores, Iria and
Santos, Diana and
Freitas, Larissa and
Souza, Jackson Wilke da Cruz and
Ribeiro, Eug{\'e}nio",
booktitle = "Proceedings of the 17th International Conference on Computational Processing of {P}ortuguese ({PROPOR} 2026) - Vol. 1",
month = apr,
year = "2026",
address = "Salvador, Brazil",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2026.propor-1.81/",
pages = "819--829",
ISBN = "979-8-89176-387-6",
abstract = "Robust text-to-speech (TTS) systems must be trained on speech that mirrors the variability and imperfections of spontaneous dialogues. However, TTS systems trained on existing Brazilian Portuguese datasets are typically limited to clean, scripted, or studio-recorded speech. Certas Palavras (CP) bridges this gap with 70 hours of spontaneous, multi-speaker dialogues from a Brazilian radio program broadcast in the 1980s{--}1990s. The extensive manual annotation process captures conversational dynamics, including orality markers, filled pauses, and hesitations. For the analog medium, we annotated non-verbal phenomena as musical interference, noise and segmental corrections, describing a challenging acoustic environment for synthesis. Baseline YourTTS and F5-TTS models were trained in a 9-hour subset featuring one of the two main hosts of Certas Palavras. Baseline YourTTS and F5-TTS models were trained on a 9-hour single-speaker subset corresponding to one of the main program hosts. Objective evaluation shows that the synthesized speech remains intelligible, with moderate WER and CER. In contrast, subjective evaluation reveals a clear gap in perceived naturalness, with lower MOS scores and higher inter-rater variability compared to ground-truth audio. Together, these properties make the dataset a strong benchmark for TTS robustness."
}<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="araujo-etal-2026-certas">
<titleInfo>
<title>Certas Palavras: A 1980s-90s Brazilian Radio Corpus to Test TTS Models in Noisy Multi-Speaker Dialogues</title>
</titleInfo>
<name type="personal">
<namePart type="given">Gustavo</namePart>
<namePart type="given">Evangelista</namePart>
<namePart type="family">Araújo</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Moacir</namePart>
<namePart type="family">Ponti</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Arnaldo</namePart>
<namePart type="given">Candido</namePart>
<namePart type="family">Junior</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sidney</namePart>
<namePart type="family">Leal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Edresson</namePart>
<namePart type="family">Casanova</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Renato</namePart>
<namePart type="given">Moraes</namePart>
<namePart type="family">Silva</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Miguel</namePart>
<namePart type="given">Oliveira</namePart>
<namePart type="family">Jr.</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Adriana</namePart>
<namePart type="given">Barbosa</namePart>
<namePart type="family">Santos</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Gustavo</namePart>
<namePart type="given">Wadas</namePart>
<namePart type="family">Lopes</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sandra</namePart>
<namePart type="family">Aluisio</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2026-04</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 17th International Conference on Computational Processing of Portuguese (PROPOR 2026) - Vol. 1</title>
</titleInfo>
<name type="personal">
<namePart type="given">Marlo</namePart>
<namePart type="family">Souza</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Iria</namePart>
<namePart type="family">de-Dios-Flores</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Diana</namePart>
<namePart type="family">Santos</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Larissa</namePart>
<namePart type="family">Freitas</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jackson</namePart>
<namePart type="given">Wilke</namePart>
<namePart type="given">da</namePart>
<namePart type="given">Cruz</namePart>
<namePart type="family">Souza</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Eugénio</namePart>
<namePart type="family">Ribeiro</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Salvador, Brazil</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
<identifier type="isbn">979-8-89176-387-6</identifier>
</relatedItem>
<abstract>Robust text-to-speech (TTS) systems must be trained on speech that mirrors the variability and imperfections of spontaneous dialogues. However, TTS systems trained on existing Brazilian Portuguese datasets are typically limited to clean, scripted, or studio-recorded speech. Certas Palavras (CP) bridges this gap with 70 hours of spontaneous, multi-speaker dialogues from a Brazilian radio program broadcast in the 1980s–1990s. The extensive manual annotation process captures conversational dynamics, including orality markers, filled pauses, and hesitations. For the analog medium, we annotated non-verbal phenomena as musical interference, noise and segmental corrections, describing a challenging acoustic environment for synthesis. Baseline YourTTS and F5-TTS models were trained in a 9-hour subset featuring one of the two main hosts of Certas Palavras. Baseline YourTTS and F5-TTS models were trained on a 9-hour single-speaker subset corresponding to one of the main program hosts. Objective evaluation shows that the synthesized speech remains intelligible, with moderate WER and CER. In contrast, subjective evaluation reveals a clear gap in perceived naturalness, with lower MOS scores and higher inter-rater variability compared to ground-truth audio. Together, these properties make the dataset a strong benchmark for TTS robustness.</abstract>
<identifier type="citekey">araujo-etal-2026-certas</identifier>
<location>
<url>https://aclanthology.org/2026.propor-1.81/</url>
</location>
<part>
<date>2026-04</date>
<extent unit="page">
<start>819</start>
<end>829</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Certas Palavras: A 1980s-90s Brazilian Radio Corpus to Test TTS Models in Noisy Multi-Speaker Dialogues
%A Araújo, Gustavo Evangelista
%A Ponti, Moacir
%A Junior, Arnaldo Candido
%A Leal, Sidney
%A Casanova, Edresson
%A Silva, Renato Moraes
%A Jr., Miguel Oliveira
%A Santos, Adriana Barbosa
%A Lopes, Gustavo Wadas
%A Aluisio, Sandra
%Y Souza, Marlo
%Y de-Dios-Flores, Iria
%Y Santos, Diana
%Y Freitas, Larissa
%Y Souza, Jackson Wilke da Cruz
%Y Ribeiro, Eugénio
%S Proceedings of the 17th International Conference on Computational Processing of Portuguese (PROPOR 2026) - Vol. 1
%D 2026
%8 April
%I Association for Computational Linguistics
%C Salvador, Brazil
%@ 979-8-89176-387-6
%F araujo-etal-2026-certas
%X Robust text-to-speech (TTS) systems must be trained on speech that mirrors the variability and imperfections of spontaneous dialogues. However, TTS systems trained on existing Brazilian Portuguese datasets are typically limited to clean, scripted, or studio-recorded speech. Certas Palavras (CP) bridges this gap with 70 hours of spontaneous, multi-speaker dialogues from a Brazilian radio program broadcast in the 1980s–1990s. The extensive manual annotation process captures conversational dynamics, including orality markers, filled pauses, and hesitations. For the analog medium, we annotated non-verbal phenomena as musical interference, noise and segmental corrections, describing a challenging acoustic environment for synthesis. Baseline YourTTS and F5-TTS models were trained in a 9-hour subset featuring one of the two main hosts of Certas Palavras. Baseline YourTTS and F5-TTS models were trained on a 9-hour single-speaker subset corresponding to one of the main program hosts. Objective evaluation shows that the synthesized speech remains intelligible, with moderate WER and CER. In contrast, subjective evaluation reveals a clear gap in perceived naturalness, with lower MOS scores and higher inter-rater variability compared to ground-truth audio. Together, these properties make the dataset a strong benchmark for TTS robustness.
%U https://aclanthology.org/2026.propor-1.81/
%P 819-829
Markdown (Informal)
[Certas Palavras: A 1980s-90s Brazilian Radio Corpus to Test TTS Models in Noisy Multi-Speaker Dialogues](https://aclanthology.org/2026.propor-1.81/) (Araújo et al., PROPOR 2026)
ACL
- Gustavo Evangelista Araújo, Moacir Ponti, Arnaldo Candido Junior, Sidney Leal, Edresson Casanova, Renato Moraes Silva, Miguel Oliveira Jr., Adriana Barbosa Santos, Gustavo Wadas Lopes, and Sandra Aluisio. 2026. Certas Palavras: A 1980s-90s Brazilian Radio Corpus to Test TTS Models in Noisy Multi-Speaker Dialogues. In Proceedings of the 17th International Conference on Computational Processing of Portuguese (PROPOR 2026) - Vol. 1, pages 819–829, Salvador, Brazil. Association for Computational Linguistics.