You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
PaddleSpeech/demos/TTSCppFrontend/README.md

57 lines
1.4 KiB

# PaddleSpeech TTS CPP Frontend
A TTS frontend that implements text-to-phoneme conversion.
Currently it only supports Chinese, any English word will crash the demo.
## Install Build Tools
```
# Ubuntu
sudo apt install build-essential cmake pkg-config
# CentOS
sudo yum groupinstall "Development Tools"
sudo yum install cmake
```
If your cmake version is too old, you can go here to download a precompiled new version: https://cmake.org/download/
## Build
```
# Build with all CPU cores
./build.sh
# Build with 1 core
./build.sh -j1
```
Dependent libraries will be automatically downloaded to the `third-party/build` folder.
If the download speed is too slow, you can open [third-party/CMakeLists.txt](third-party/CMakeLists.txt) and modify `GIT_REPOSITORY` URLs.
## Download dictionary files
```
./download.sh
```
## Run
You can change `--phone2id_path` in `./front_demo/front.conf` to the `phone_id_map.txt` of your own acoustic model.
```
./run_front_demo.sh
./run_front_demo.sh --help
./run_front_demo.sh --sentence "这是语音合成服务的文本前端,用于将文本转换为音素序号数组。"
./run_front_demo.sh --front_conf ./front_demo/front.conf --sentence "你还需要一个语音合成后端才能将其转换为实际的声音。"
```
## Clean
```
./clean.sh
```
The folders `front_demo/dict`, `build` and `third-party/build` will be deleted.