Open Manothave opened 4 years ago
Thanks for the suggestion, it would be interesting to get the pronunciation of some voice synthesizer like Google for the translator and collect the sound of each kana.
To avoid interaction with non-free network services, it would be best to store it in the application, I will have to carry out tests on how to download those sounds and then play them in the app. It would be interesting not to increase the size of the application too much to avoid overloading it.
OsmAnd have japanese TTS, if it helps
You could probably grab voices from an Anki deck.
Please add pronunciation