Skip to content
RAG Repo

LJ Speech

LJ Speech is a compact, clean corpus made for training and benchmarking text-to-speech (TTS, generating spoken audio from written text) systems. It consists of 13,100 short clips of one speaker reading passages from seven non-fiction books, with a transcript for each clip and durations between one and ten seconds.

Because it is a single speaker recorded consistently, it is easy to work with and has become the default starting point for TTS research and tutorials. The whole dataset is in the public domain, so you can use it for any purpose, including commercial work, with no restrictions or attribution required.

speechenglishtext-to-speechpublic-domainsingle-speakerbenchmark

Related sources