There are other corpora that have been prepared using public domain Librivox data.
| Corpus | Description |
|---|---|
| Librispeech | LibriSpeech is a corpus of approximately 1000 hours of 16kHz read English speech. |
| LJ Speech | Approximately 24 hours of read 22.05 KHz English speech from a single female speaker |
| Tundra | The Simple4All Tundra Corpus is a collection of 14 audiobooks in 14 languages. |
| CSS10 | A collection of single speaker speech datasets for 10 languages. |
| M-AILABS | A collection of nearly 1000 hours of 16KHz read speech (male and female speakers) for 10 languages. |
QuoteLi3 is a manually annotated dataset for quote attribution in literary text. In this dataset, direct speech is tagged with respect to the character that is speaking. It was prepared by Muzny et al (2017), who proposed an automatic method for quote attribution. One of the annotated books was Emma, by Jane Austen. This labelled data could be used together with the parallel recordings of Emma distributed with the Parallel Audiobook Corpus. This might be helpful if, for example, you wish to investigate character voices in audiobook recordings.