Text-to-Speech Script: Guide

The provided Python script converts text from a file into speech using a specified Text-to-Speech (TTS) model. Below are detailed instructions on how to use this script effectively:

Prerequisites

Before running the script, ensure you have installed the required libraries. Run

pipenv install

or install the libraries with pip:

pip install coqui-tts
pip install pydub
pip install torch

Save the provided Python code in a file named storyteller.py or any preferred name.

Input Text File

Prepare a text file with your desired content. For example, save it as input.txt.

Running the Script

Open a terminal or command prompt and navigate to the directory containing storyteller.py and your input text file. Run the script using appropriate parameters:

Available parameters include:

• --in-file: Path to the input text file (required).
• --out-file: Path for the output WAV file (default: "book.wav").
• --start-line: Line number to start reading from (default: 1).
• --to-line or --process-lines: To specify either an end line number or a count of lines to process. If --to-line is set, it specifies the last line number; if not, --process-lines determines how many lines are processed starting from --start-line.
• --pause-between-lines: Duration in milliseconds for silence between lines (default: 1000).

Example usage:

> ./storyteller --start-line=3 --to-line=5 --process-lines=130 --in-file=~/Temp/alice.txt --out-file=alice.wav --pause-between-lines=750 --model-name=tts_models/en/vctk/vits --speaker-id=p230

This command converts lines 3 to 5 (inclusive) from the alice.txt file into speech, with a pause of 750 milliseconds between each line. The output is saved as "alice.wav".

Model Selection

The script automatically loads a default TTS model but can be customized using:

• --model-name: Path or name of the TTS model to use (the default value is set to "tts_models/en/vctk/vits").
• --speaker-id: Speaker ID for multi-speaker models.

If the requested model cannot be found, the script lists available models and prompts for selection.

In case of multi-speaker model, the script lists available speaker IDs and prompts for selection.

Device Selection

The script determines a suitable device for processing the TTS output automatically.

Note on Usage

Ensure that your TTS model is compatible with the input text. The script currently targets the English language using VCTK dataset-based models.

With these steps, you can generate audio files from textual content seamlessly using this Python script and the provided README guide.