Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/darthph0enix7/tts_and_speech_recognition
https://github.com/darthph0enix7/tts_and_speech_recognition
Last synced: 11 days ago
JSON representation
- Host: GitHub
- URL: https://github.com/darthph0enix7/tts_and_speech_recognition
- Owner: Darthph0enix7
- Created: 2024-08-15T16:10:34.000Z (5 months ago)
- Default Branch: main
- Last Pushed: 2024-09-07T19:41:42.000Z (4 months ago)
- Last Synced: 2024-11-05T23:29:03.879Z (about 2 months ago)
- Language: Jupyter Notebook
- Size: 23.6 MB
- Stars: 1
- Watchers: 1
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Voice Synthesizer - TTSStreamer
This project is a Text-to-Speech (TTS) synthesizer that clones voices based on short audio clips, such as famous personalities like Morgan Freeman. It generates speech in the style of the target voice using a custom trained model and allows real-time streaming of generated audio with buffering. The project utilizes `pydub` for audio processing and `TTS` models for speech synthesis.
## Features
- **Clone Voice:** Synthesize a target voice based on a short audio clip.
- **Streaming Audio Playback:** Play generated audio with real-time buffering.
- **Adjustable Speed & Language:** Modify the speed and language of the generated audio.
- **GPU Support:** Leverage CUDA for faster audio generation using GPUs.
## Installation
1. Clone the repository:
```bash
git clone https://github.com/your-repo/voice-synthesizer.git
cd voice-synthesizer
```
2. Install dependencies:
```bash
pip install pydub soundfile torch TTS
```
3. Download or prepare your pre-trained TTS model and place it in the `XTTS-v2` directory.
## Usage
### Initialize TTSStreamer
The `TTSStreamer` class handles the loading of the TTS model and the generation of audio from text.
```python
from tts_streamer import TTSStreamer
tts_streamer = TTSStreamer(
model_path="XTTS-v2",
config_path="XTTS-v2/config.json",
vocab_path="XTTS-v2/vocab.json"
)
```
### Stream Text as Audio
Provide text input to stream audio with real-time playback.
```python
text = """
A series of extracts from Middle English texts of different dates and types...
"""
tts_streamer.stream_audio_with_buffering(
text,
language="en",
speed=1.3,
speaker_path="XTTS-v2/samples/morgan_freeman.wav",
fireup_delay=5.0
)
```
This will generate audio in the cloned voice style, buffer it, and play it with a 5-second delay before starting playback.
## Project Structure
- `TTSStreamer`: Main class that handles model loading, text processing, and audio generation.
- `XTTS-v2/`: Directory for storing model files, configuration, and vocabulary.
- `samples/`: Example audio files used to clone voices.
## Requirements
- Python 3.8+
- [PyDub](https://github.com/jiaaro/pydub)
- [TTS](https://github.com/coqui-ai/TTS)
- GPU (CUDA support recommended)
## Example
Here’s an example of the generated output using Morgan Freeman’s voice:
- **Original Clip**: `samples/morgan_freeman_original.mp3`
- **Cloned Voice**: `samples/morgan_freeman_cloned.mp3`
## License
This project is licensed under the MIT License.
## Acknowledgements
Special thanks to the developers of the [Coqui TTS](https://github.com/coqui-ai/TTS) framework for providing open-source tools for text-to-speech synthesis.