UCLA Speaker Variability Database

AbstractIntroduction UCLA Speaker Variability Database was developed by UCLA Speech Processing and Auditory Perception Laboratory and is comprised of approximately 34 hours of English speech and orthographic transcripts. This corpus was designed to sample variability in speaking within individual speakers and across a large number of speakers. Participants took part in six different tasks: vowel sounds, reading sentences, giving instructions, neutral conversation, happy conversation, a phone conversation, annoyed conversation, and responding to a video. Data Speakers (101 female, 101 male) were recruited from the UCLA university community and represented a variety of language backgrounds. Speaker metadata is included in the documentation for this release. Audio files are presented as 16kHz, 16-bit single channel, flac compressed wav files. Most speech files have two corresponding Praat TextGrid files: an orthographic sentence/utterance transcription and output from forced alignment.