/EPUB-to-MP3

A script to convert an eBook (.epub) to an audiobook (.mp3) using Google Cloud Platform's Text-to-Speech service

Primary LanguagePython

EPUB To MP3

Use text-to-speech synthesis to create an MP3 audiobook of an eBook (in .epub format).

Development Environment

Developed with Python 3 on Ubuntu 18.04, and uses Google Cloud Speech-to-Text API.

Setup

MP3 Codecs

ffmpeg is used by the pydub library to cocatenate MP3s together. This is needed because an audiobook will need to be separated into chunks, sent over to the Google Cloud Platfrom Text-to-Speech API, and the individual chunk's MP3s will be concatenated.

$ sudo apt update
$ sudo apt install ffmpeg

GPU Support for offline TTS

Highly recomended if you plan to do the speech synthesis on your local machine, as this is quite slow without GPU acceleration.

For offline models, if you have a GPU and are on WSL, follow these instructions. You'll want to run

$ sudo apt install nvidia-cuda-toolkit
$ sudo apt-get install -y cuda-toolkit-10-1

after installing the drivers on windows, which is a little unclear from the instructions. Check that everything is intalled by running nvcc --version in the linux terminal, you should have 10.1.

If not, you may also need to follow the vanilla linux isntructions below to add cuda 10.1 to your PATH.

If you are on on plain old Linux, follow these instructions

Setup

Online TTS

Overview

Text will be sent to Google Cloud Platform's Text to Speech synthesis service.

Google Cloud Platform

To use online recognition, set up a Google Cloud Platform account, following the instructions here: https://cloud.google.com/text-to-speech/docs/libraries#client-libraries-install-python

You'll get a JSON file with your Google Cloud Platform credentials. Set it in environment variables.

For example:

$ export GOOGLE_APPLICATION_CREDENTIALS=~/dev/EPUB-to-MP3/GCP-credentials.json

Install required packages

Set up virutalenv and install libraries for:

  • Google Cloud Text to Speech API
  • ePub parsing
$ virtualenv venv
$ source venv\bin\activate
$ pip3 install -r requirements_online_tts.txt

Local TTS

Overview

Text-to-speech synthesis be performed locally.

Instructions

Set up virutalenv and install libraries for:

  • TensorFlow
  • ePub parsing
$ virtualenv venv
$ source venv/bin/activate
$ pip3 install -r requirements_local_tts.txt

Offline model downloads

Option 1: Download pre-zipped models

Download the zip file and extract to this directory (i.e. the root of the repo).

Option 2: Download models from source

Otherwise you can download models to the model_files folder from their respective sub-folders in the repo. For each The necessary files are the configuration.yaml file and the generator.h5 file

Usage

$ python3 convert.py --file="book.epub"

Offline usage

$ python3 convert.py --file="book.epub" --local