From Voice To Shell: Speech To Luminaires Dataset

Dataset with 1,060 audio files, 1.18 hours, comprising voice recordings of 5 subjects (2 women and 3 men) while enunciating the prompts of 'text-to-luminaires' dataset test samples in both English (EN) and Spanish (ES). Followed protocol has been approved by the Social Research Ethics Committee (SREC) of the University of Castilla-La Mancha (UCLM) under reference number CEIS-2025-119450. Human participation is justified and the benefits and risks have been adequately assessed to participants, developing a participation mechanism which warrants equal opportunities for research collaboration. Every participant has been provided an informed consent document including the necessary information of research purposes and requirements. It also complies with the regulations in force regarding the personal data protection. Audio files are encrypted, which could be extracted using the decrypt.py script: user@host:#~/data/$ lsen es decrypt.py metadatauser@host:#~/data/$ pip install cryptography[... install dependencies ...]user@host:#~/data/$ python3 decrypt.py -p <password> .[... decrypt files ...]