: Vocabulaire du TAL: speech quality

applied NLP > speech processing > speech quality

speech quality

A single score for MOS which incorporates scores for a variety of speech impairments serves as a general indicator of speech quality. (Gupta, 2023)
It should be noted that the speech data are faithfully transcribed where pauses and filler words such as um er uh can lower the quality of the speech relative the specificity model which is trained on native textual input hence making it challenging to capture specificity. (Alqahtani, Kayi, Hamidian, Compton & Diab, 2022)
The MOS score shouldn't be used only to describe speech quality because it is a generic statistic that takes numerous aspects into account. (Gupta, 2023)

http://data.loterre.fr/ark:/67375/8LP-MGG8F31X-H

RDF/XML TURTLE JSON-LD Dernière modification le 06/06/2024

Loterre