Text to speech

This is a project that aims to convert text from the PDF file to mp3 format.

Table of Contents

• Introduction
• Usage
• Functionalities
  • General
  • Text extraction
  • Saving the extracted text to mp3 format
• License

Introduction

This project is designed to extract text from a PDF file and convert it to an mp3 audio file. It uses PyPDF2 to extract text from the PDF and pyttsx3 to convert the text to speech and save it as an MP3 file.

Usage

1. Navigate to the project directory.
2. Run the script (main file).
3. Enter the name and path of the PDF file you want to convert when prompted.
4. Enter the name and path of the MP3 file you want to save when prompted.
5. Wait for the script to extract the text from the PDF file and save it as an MP3 audio file.

Functionalities

General

This Python script contains two functions and a conditional code block that allows it to be executed as a standalone program. The purpose of the script is to extract text from a PDF file, convert it to speech and save it as an mp3 audio file. It needs approximately 20 seconds to convert 30-page book.

Text extraction¹

The extract_text_from_pdf function takes in the file path to a PDF file as an argument and returns the extracted text from the file as a string. It first opens the PDF file in binary mode using a with statement, reads each page of the file using the PyPDF2 module and concatenates the extracted text from all pages into a single string. If an error occurs while opening or reading the file, the function raises an appropriate exception.

Saving the extracted text to mp3 format²

The save_to_mp3 function takes in the text to be converted to speech, the file path to save the resulting mp3 audio file, the speed at which the audio is spoken and the type of voice to use as arguments. The function initializes the pyttsx3 text-to-speech engine, sets the speech rate and voice type based on the provided arguments, saves the text as an mp3 audio file at the specified path using the save_to_file method, and processes the resulting audio file using the runAndWait method. If an error occurs while saving the file, the function raises an appropriate exception.

License

This project is licensed under the MIT License.

Footnotes

1. This function can still be modified since it doesn't extract the text from pdf file in a perfect way (like we read it). It reads some useless parts like watermarks or warranty information. It is probably possible to find a general way to exclude this but it is hard to achieve this. Additionally, it may truncate a portion of a text which is maybe important (like for example at the very end of the text). ↩

2. This function uses the engine from the pyttsx3 library to convert the text to audio. This engine is free and has no limitations regarding usage unlike Google API for text-to-speech (because it uses your computer's resources). But, it does have limitations regarding parameters which you are allowed to set. I made possible to change the type of voice (male/female) and they are Microsoft's default options. For more languages or different people, you can probably install packages locally and set the parameter. Also, default speed is kind of regular (normal) speed for speaking. If you change the parameter to 100 it would be very slow, and if you change it to 200 it would be very fast. The most important part is that you can modify this function as you like it. ↩