Introduction

eSpeak NG is a compact, open-source text-to-speech (TTS) engine that supports multiple languages, including German and English. It is lightweight, fast, and works offline, making it ideal for systems where internet access is limited or unnecessary. While its default voices are synthetic, you can significantly improve voice quality by integrating MBROLA, a high-quality speech synthesizer that provides more natural-sounding voices.

Personally, I use eSpeak NG to read out text files like Markdown documents or (summarized) news articles, and the improved voice quality with MBROLA makes it much more enjoyable to listen to.

This article guides you through installing and configuring eSpeak NG on Debian, including how to add MBROLA voices for enhanced German speech output.

Installation of eSpeak NG

Step 1: Install eSpeak NG

Open a terminal and run the following command to install eSpeak NG and its data package:

sudo apt-get install espeak-ng espeak-ng-data

• espeak-ng: The core TTS engine.
• espeak-ng-data: Additional voice data for multiple languages.

Step 2: Verify Installation

Check if eSpeak NG is installed correctly by listing available voices:

espeak-ng --voices

This command displays a table of available voices, their languages, and genders (e.g., de for German, en for English, M for male, F for female).

Adding MBROLA Voices

MBROLA is a speech synthesizer that works alongside eSpeak NG to provide higher-quality voices. Debian’s repositories include several German MBROLA voices, both male and female.

Step 1: Install MBROLA

First, install the MBROLA binary:

sudo apt-get install mbrola

Step 2: Install German MBROLA Voices

Debian provides multiple German MBROLA voices. To list all available German MBROLA voices, run:

apt-cache search mbrola | grep de

This returns a list like:

mbrola-de1 - German female voice for Mbrola
mbrola-de2 - German male voice for Mbrola
mbrola-de3 - German female voice for Mbrola
mbrola-de4 - German male voice for Mbrola
mbrola-de5 - German female voice for Mbrola
mbrola-de6 - German male voice for Mbrola
mbrola-de7 - German female voice for Mbrola
mbrola-de8 - German-Bavarian male voice for Mbrola

To install all German MBROLA voices, use:

sudo apt-get install mbrola-de*

This installs all available German voices, including both male and female options.

Using eSpeak NG with MBROLA

Step 1: Test Default eSpeak NG Voices

Try the default German voice:

espeak-ng -v de "Hallo, das ist ein Test mit eSpeak NG."

• -v de: Uses the default German voice.

Step 2: Use MBROLA Voices

To use MBROLA voices with eSpeak NG, specify the MBROLA voice in the command. For example, to use the **German female voice de1**:

espeak-ng -v mb-de1 "Hallo, das ist ein Test mit MBROLA." -p 20 -s 150

• -v mb-de1: Uses the MBROLA German female voice de1.
• -p 20: Adjusts the pitch (higher values sound more feminine).
• -s 150: Adjusts the speed (words per minute).

Available MBROLA German Voices

Customizing Voice Output

eSpeak NG allows you to fine-tune the speech output:

• Pitch: Use -p to adjust pitch (e.g., -p 30 for a higher pitch).
• Speed: Use -s to adjust speed (e.g., -s 120 for slower speech).
• Volume: Use -a to adjust amplitude (e.g., -a 200 for louder output).

Example:

espeak-ng -v mb-de3 -p 30 -s 130 -a 150 "Das ist ein angepasster Test."

Saving Speech to a File

You can save the speech output to a WAV file for later use:

espeak-ng -v mb-de1 -w output.wav "Hallo, das ist eine Audiodatei."

• -w output.wav: Saves the speech to output.wav.

Read a Text File

To read text from a file, use:

cat /path/to/your/file.txt | espeak-ng -v mb-us2

• cat: Reads the content of the specified file and pipes it to eSpeak NG.
• -v mb-us2: Uses the MBROLA US English voice us2.

Alternatively, you can directly specify the file with the -f option:

espeak-ng -v mb-en1 -f /path/to/your/file.txt

• -f: Specifies the input file to read.
• mb-en1: Uses the MBROLA English voice en1.

Conclusion

eSpeak NG is a versatile and lightweight TTS solution for Debian, and integrating MBROLA voices significantly enhances the naturalness of the speech output. Whether you need a simple command-line TTS tool or a more customizable solution with higher-quality voices, eSpeak NG and MBROLA provide a robust and flexible option.

Experiment with different voices and settings to find the best fit for your needs!