Speech Synthesis Datasets

This dataset contains 20 hours of American English male voice recordings. It is recorded by Americans (native English speakers) with authentic accent. The phoneme coverage is balanced. Professional phonetician participates in the annotation. It is suitable for text-to-speech (TTS) model training, phoneme recognition research, and AI voice development.

This dataset contains 19.46 hours of American English female voice recordings. It is recorded by American (native English speaker) with authentic accent and clear, sweet tone. The phoneme coverage is balanced. Professional phoneticians participate in the annotation. It is suitable for text-to-speech (TTS) model training, phoneme recognition, and AI voice development requiring natural-sounding female speech.

This dataset contains 10.4 hours of Japanese female voice recordings. It is recorded by Japanese native speaker with an authentic accent. The phoneme coverage is balanced. Professional phonetician participates in the annotation. This corpus is ideal for tasks such as Japanese text-to-speech (TTS) training, speech synthesis research, and AI voice model development.

This dataset features recordings from 2 native Australian English speakers with authentic accents. The phoneme coverage is balanced. Professional phonetician participates in the annotation. It precisely matches with the research and development needs of the speech synthesis.

This dataset includes recordings from 2 native Spanish speakers with authentic Castilian accents. The phoneme coverage is balanced. Professional phonetician participates in the annotation. It precisely matches with the research and development needs of the speech synthesis.

This dataset includes recordings from 3 native Italian speakers with authentic accents. Covering both customer service and general speaking styles. The phoneme coverage is balanced. Professional phonetician participates in the annotation. It precisely matches with the research and development needs of the speech synthesis.

This dataset contains recordings from 10 native British English speakers with an authentic UK accent. The phoneme coverage is balanced. Professional phonetician participates in the annotation. It precisely matches with the research and development needs of the text-to-speech (TTS) systems and AI voice synthesis models.

This dataset contains recordings from 2 native Canadian French speakers with authentic accents. It is ideal for researchers and developers seeking natural Canadian French voices.

Indonesian Natural Conversational Speech Synthesis Corpus. Recorded by two native voice talents (one male, one female), this corpus comprises two modules: free dialogue on specified topics and scripted emotional/paralinguistic text performances. It covers single-emotion, multi-emotion, and paralinguistic expressions (including a variety of modal particles). Professional linguists have provided three-tier annotation—on text content, emotion labels, and paralinguistic phenomena. The total duration is approximately 11 hours, with audio specifications of 48 kHz sampling rate, 24‑bit depth, WAV format, mono channel. This dataset is designed to fully meet the diverse requirements of TTS research and development in terms of naturalness, expressiveness, and precise annotation.