Text to Speech by using pyttsx3 - Python

Converting text to speech can add a new level of interactivity to our Python applications. Whether we want to create a virtual assistant or simply make our program more engaging, pyttsx3 is a library used for converting text into speech. This offline tool offers flexibility with male and female voice options and different TTS engines. Now lets understand how to implement it.

1. Installing pyttsx3

To install pyttsx3 module first of all we have to open the terminal and run the following command:

pip install pyttsx3

If we receive errors such as No module named win32com.client, No module named win32 or No module named win32api we will need to additionally install pypiwin32 using below command:

pip install pypiwin32

2. Importing and Initializing Speech Engine

Once the library is installed, we can import it into your Python script and initialize the pyttsx3 engine. This engine is responsible for handling text-to-speech conversion. init function is used to get an engine instance for the speech synthesis

pyttsx3 supports three different text-to-speech engines depending on your platform:

1. sapi5: Default engine on Windows.
2. nsss: Available on Mac OS X.
3. espeak: Available on Linux and other platforms.

Library automatically selects appropriate engine based on our operating system, but we can switch between them if needed.

import pyttsx3
engine = pyttsx3.init()

3. Customizing Speech Properties

We can adjust several properties of the speech such as rate, volume and voice.

• engine.getProperty('rate'): This fetches current rate (speed) at which the speech is being spoken. Rate is measured in words per minute (wpm).
• engine.setProperty('rate', rate - 50): This sets the new rate (speed) for the speech engine.
• engine.setProperty('volume', 1): This controls the volume (loudness) of the speech. Value ranges from 0.0 to 1.0.
• voices = engine.getProperty('voices'): This retrieves the available voices from the speech engine. The voices is a list of available voice objects and each object contains properties such as the voice’s ID, name, gender and language.

rate = engine.getProperty('rate')
engine.setProperty('rate', rate - 50)  
engine.setProperty('volume', 1)  
voices = engine.getProperty('voices')
engine.setProperty('voice', voices[0].id)  

4. Converting Text to Speech

Now that properties are set we can convert any text to speech. Use the say() method to add text we want to be spoken, then call runAndWait() to make the engine process the speech.

engine.say("Hello sir, how may I help you, sir.")
engine.runAndWait()

Output :

From the above code it would be a voice saying

'Hello sir, how may I help you, sir.'

The pyttsx3 allows us to easily convert text to speech in Python with full control over speech’s speed, volume and voice type. Whether we're working on a project that requires audio feedback or simply want to add an interactive feature, customizing these properties gives us the flexibility to make the speech sound exactly the way we want.