Evaluating the Santa Barbara Corpus: Challenges of the Breadth of Conversational Spoken Language

Evaluating the Santa Barbara Corpus: Challenges of the Breadth of Conversational Spoken Language

Paper in proceedings (conference paper)

As speech technology has matured, there has been a push to- wards systems that can process conversational speech, reflect- ing the so-called "cocktail party problem," which includes not only more challenging acoustic conditions, but also necessi- tates solutions to new problems, such as identifying who spoke when and processing multiple concurrent streams of speech. Such problems have been approached primarily via corpora comprising business meetings and dinner parties, overlooking the broad range of conversational dynamics and speaker de- mographics that fall under the category of multi-talker speech. To this end, we introduce the use of the Santa Barbara Corpus of Spoken American English for evaluation of speech technol- ogy-including preparing the corpus and annotations for auto- matic processing, demonstrating the failure of state-of-the-art systems to withstand the heterogeneity of conditions, and high- lighting the situations where standard methods struggle to per- form at all

As speech technology has matured, there has been a push to- wards systems that can process conversational speech, reflect- ing the so-called "cocktail party problem," which includes not only more challenging acoustic conditions, but also necessi- tates solutions to new problems, such as identifying who spoke when and processing multiple concurrent streams of speech. Such problems have been approached primarily via corpora comprising business meetings and dinner parties, overlooking the broad range of conversational dynamics and speaker de- mographics that fall under the category of multi-talker speech. To this end, we introduce the use of the Santa Barbara Corpus of Spoken American English for evaluation of speech technol- ogy-including preparing the corpus and annotations for auto- matic processing, demonstrating the failure of state-of-the-art systems to withstand the heterogeneity of conditions, and high- lighting the situations where standard methods struggle to per- form at all

conversational speech, diarization, speech recognition

conversational speech, diarization, speech recognition

MACIEJEWSKI, M.; KLEMENT, D.; HUANG, R.; WIESNER, M.; KHUDANPUR, S.

2025

01.09.2024

International Speech Communication Association

Kos

Proceedings of Interspeech 2024

1990-9772

Proceedings of Interspeech

2024

9

French Republic

2155

2160

5

https://www.isca-archive.org/interspeech_2024/maciejewski24_interspeech.pdf

maciejewski_2024_interspeech