LJSpeech-1

24 hours of single-speaker recordings for text-to-speech (TTS) models.

About Dataset

LJ Speech is a single-speaker speech dataset consisting of approximately 13,000 short audio clips of a female speaker reading passages from non-fiction books. Created by Keith Ito, the dataset includes aligned text transcripts and high-quality audio recordings. It is cleanly segmented and well-documented, making it suitable for speech synthesis research.

Purpose of Dataset

Lj Speech Is Widely Used For Training And Benchmarking Text-to-speech (Tts) Systems. Its Consistency And Audio Quality Make It Ideal For Developing Neural Speech Synthesis Models. Researchers Use It To Study Voice Modeling, Pronunciation, And Prosody, And It Is Commonly Employed In Educational And Prototyping Settings For Speech Generation.