🎤 Bark AI Text-to-Speech (TTS) Project

🚀 Bark AI is a powerful AI text-to-speech program. This program, bark_textfile.py, converts the text content from the text.txt file into speech and saves it as output.wav.

📌 Features

👉 Automatic text processing – Reads the content of text.txt and converts it into speech.
👉 Supports long texts – The script automatically splits long texts into meaningful paragraphs.
👉 Single audio file – All sections are merged into one file.
👉 Automatic playback – The audio file is played automatically after processing.

---

🔧 Installation

1️⃣ Requirements

• Python 3.8 or newer
• pip (Python package manager)
• Windows, Linux, or macOS
• Powerful hardware recommended

  • The processing time can be significant. In a test using an Nvidia RTX 4070 and 96GB RAM, the conversion of a text file containing:

    • 811 words
    • 4,897 characters (without spaces)
    • 5,690 characters (with spaces)
    • 27 sentences
    • 30.0 words per sentence on average
    • Estimated reading time: 6.2 minutes

    still took over an hour to complete. We strongly recommend using a high-performance GPU and ample RAM.

2️⃣ Clone the repository

If you have Git installed:

git clone https://github.com/your-github-username/bark-ai-tts.git
cd bark-ai-tts

Alternatively, you can download the ZIP file from GitHub and extract it.

3️⃣ Create and activate a virtual environment

python -m venv venv

• Windows (CMD/PowerShell):

  venv\Scripts\activate

• Linux/macOS:

  source venv/bin/activate

4️⃣ Install dependencies

pip install -r requirements.txt

---

🎤 Usage

1. Write the desired text in text.txt.
2. Run the script:

   python bark_textfile.py

3. The script processes the text and saves the speech output as output.wav.
4. The file is played automatically.

🔍 How does bark_textfile.py work?

• The script reads the content from the text.txt file.
• If the text is too long, it is automatically split into smaller chunks.
• Each section is converted into speech using Bark AI.
• All generated audio data is merged into a single file: output.wav.
• After processing, output.wav is played automatically.

---

🛠 Troubleshooting

If the script is not working:

• Make sure you are using the correct Python version (3.8 or newer).
• If Bark AI has trouble loading the models, check the file:

  venv\Lib\site-packages\bark\generation.py
  

  and replace the following line (around line 212):

  checkpoint = torch.load(ckpt_path, map_location=device)

  with:

  checkpoint = torch.load(ckpt_path, map_location=device, weights_only=False)

• If the file does not play, open it manually (output.wav).

---

🌟 License

This project is licensed under the MIT License.
More information can be found in the LICENSE file.

---

💡 Credits

• Bark AI – Open-source text-to-speech model
• PyTorch – Deep learning framework

---