luganda_dataset

The corpus is helpful in projects related to NLP and luganda language

The file contains the phonetic file, transcription file, lexicon dictionary and 511 database sentences

Align the sentences with corresponding wav files ie recording

Added Luganda Languange and voice https://github.com/Nandutu/marytts

The complete research is a speech synthesis systems that detects Luganda text to Luganda speech using MARYTTS engine. The Luganda downloadable voice is here https://goo.gl/cmRFFi.

Instructions to Install and Test the Synthesized Voice

Install Java and add it to system path

Clone https://github.com/marytts/marytts or version 5.2 SNAPSHOT.

Download Luganda target folder from dropbox link https://goo.gl/Qi6xDo

Unzip it and place it at the root of marytts project

Navigate to target/marytts-5.2-SNAPSHOT/bin/marytts-server in command line. When the server is running; browse http://localhost:59125/

https://github.com/Nandutu/luganda_dataset . The little dataset used to train the voices. test/.txt and lg.txt

Paper

Link to the paper https://arxiv.org/ftp/arxiv/papers/2005/2005.05447.pdf