ttsmyvoice

Minimalist Python TTS using custom vocabulary.

Quick-start (Linux only but with small changes should work on Win and Mac)

First install Python pydub and mpg123

$ sudo pip install pydub
$ sudo apt-get install -y mpg123

git clone https://github.com/francescosalvadore/ttsmyvoice.git
cd ttsmyvoice/src
python prepare_voice.py
python talk.py

To change the read text modify talk.py

To use your own word:

cd ../voices
mkdir -p <your_name>/in

and edit prepare_voice.py and talk.py to use your vocabulary (voice_name variable).

Put a list of words in a .dat file (only letters and numbers in the name of file please) with a word per line, an example is in francesco/in folder, comuni1174.dat.

Then record (e.g. using Audacity and a decent microphone) an interval of words to a .mp3 file, e.g. comuni1174_1_10.mp3

Then run again prepare_voice.py

Suggestion: let at least half a second among the words, and, in case you encounter problems, tune the prepare_voice.py silence time min_silence_len and silence threshold silence_thresh.

For a very good list of italian words look at: https://github.com/napolux/paroleitaliane