Pico TTS

Text to speech voice sinthesizer from SVox, included in Android AOSP.

Build and install in Linux

The source code is inside folder 'pico'

cd pico

Create autotools files:

./autogen.sh

Configure & build:

./configure
make

Install (this install files to /usr/bin, /usr/lib and /usr/share/pico):

make install

Usage

pico2wave -l LANG -w OUT_WAV_FILE "text you want to sinthesize"
aplay OUT_WAV_FILE
rm OUT_WAV_FILE

Languages can be: en-EN, en-GB, es-ES, de-DE, fr-FR, it-IT

Output file must be .wav

TensorFlow TTS Support

Dependencies: Before building with TensorFlow support, install:

• libtensorflowlite-c
• libtensorflowlite-flex

Build with TensorFlow support:

cd pico

./autogen.sh

./configure --enable-tensorflow

make

make install

Run TTS with deep learning models:

pico2wave -d true -t <text_to_mel.tflite> -g <vocoder.tflite> -p processor.json -s "Hello world" -w output.wav

License

License Apache-2.0 (see pico_resources/NOTICE)