README_en.md

JP | EN

Coqui TTS for ROS

Table of Contents

1. Introduction
2. Getting Started
   • Prerequisites
   • Installation
3. Launch and Usage
4. Milestone
5. Acknowledgments

Introduction

This repository allows the connection of coqui-ai/TTS with ROS, providing a real-time advanced Text-to-Speech generation. The latest TTSv2 provides of 16 languages and better performance overall.

(back to top)

Getting Started

This section describes how to set up this repository.

Prerequisites

First, please set up the following environment before proceeding to the next installation stage.

System	Version
Ubuntu	20.04 (Focal Fossa) - Local Env.
Python	>= 3.9, < 3.12
Docker Engine	Tested on 26.0.0
CUDA	>=11.8 (If GPU is used)

Note

Docker is required to use this TTS library.

(back to top)

Installation

1. Go to the src folder of ROS.

   $ roscd
   # Or just use "cd ~/catkin_ws/" and change directory.
   $ cd src/

2. Clone this repository.

   $ git clone https://github.com/TeamSOBITS/coqui_tts_ros

3. Navigate into the repository.

   $ cd coqui_tts_ros/

4. Install the dependent packages.

   $ bash install.sh

5. Compile the package.

   $ roscd
   # Or just use "cd ~/catkin_ws/" and change directory.
   $ catkin_make

6. Create a simple alias to launch the TTS server.
   • If using CPU:

   $ echo "alias tts_launch='docker run --rm -it -p 5002:5002 -v ~/{PATH_ROS_WS_LOCAL}/src/coqui_tts_ros/models/:/root/.local/share/tts/ --entrypoint \"tts-server\" ghcr.io/coqui-ai/tts-cpu'" >> ~/.bash_alias

   • If using GPU:

   $ echo "alias tts_launch='docker run --rm -it -p 5002:5002 --gpus all -v ~/{PATH_ROS_WS_LOCAL}/src/coqui_tts_ros/models/:/root/.local/share/tts/ --entrypoint \"tts-server\" ghcr.io/coqui-ai/tts'" >> ~/.bash_alias

Important

{PATH_ROS_WS_LOCAL} needs to be updated to your ROS PATH in the local environment.

Important

You need to run the command 6. in the local environment.

(back to top)

Launch and Usage

1. Launch TTS server from the local environment.
   • If using CPU:

   $ tts_launch --model_name tts_models/en/vctk/vits

   • If using GPU:

   $ tts_launch --model_name tts_models/en/vctk/vits --use_cuda true

Note

Remember that --model_name value can be updated. Please, check the available models in model_list.yaml.

2. Set the parameters inside tts.launch and select the functions to be used.

   <!-- Set Coqui TTS server url -->
   <arg name="url"         default="http://localhost:5002"/>
   <!-- Add period at the end of a sentence (true) -->
   <arg name="addStopChar" default="true"/>
   <!-- Set result sound filename -->
   <arg name="filename"    default="output.wav"/>
   <!-- Set input style_wav if sample voice is given -->
   <arg name="style_wav"   default=""/>
   <!-- Set Speaker ID if multi-speaker model is being used -->
   <arg name="speaker_id"  default="p225"/>
   <!-- Set Language if multi-language model is being used -->
   <arg name="language_id" default=""/>
   <!-- Set sound_audio to true if you want to play the sound -->
   <arg name="sound_audio" default="true"/>

3. Execute the launch file tts.launch.

   $ roslaunch coqui_tts_ros tts.launch

(back to top)

Milestone

• Choose --model_name value through parameter.
• Make available the funtion of style_wav.

See the open issues for a full list of proposed features (and known issues).

(back to top)

Acknowledgments

• coqui-ai/TTS
• coqui-ai/TTS Docker images

(back to top)