A Speech synthesis (text to speech) service. This service expects a text and outputs an audio file. Different voices are implemented.
This package contains multiple implementations to create spoken language for any written text. The different voices implementeda are:
- WaveNet
- Mozilla-TTS
This repository uses Python >= 3.7. To set ip up you can run
conda create --name cltl-speechsynthesis python=3.7
conda activate cltl-speechsynthesis
pip install -e .To use this voices you need to create a project on the Google Cloud Platform supporting Text-To-Speech APIs.
Put the google_cloud_key.json file in the config folder of this repo.
To use the Mozilla TTS services please get the docker image by synesthesiam running beforehand, for example like this:
docker run -it -p 5002:5002 synesthesiam/mozillatts:en- Fix logging
- Fix config of language, save/play audio file, audio directory
- Event infrastructure
- App
- Check if we can switch voices via different APIs
- Check implementation middle layer