Transcript generation from audio file

---

Generating a transcript from an audio file with the Google Cloud Speech-to-Text API

In this Python project the Google Cloud Speech-to-Text API is used to generate a transcript from an audio file.

---

Files inculded

This project includes the following files:

• main_demo.py - main script
• transcript_generator.py - class for transcript generation

The folder 'audio_raw/' contains audio files that can be used to test the code (see 'Usage - Run demo'). This folder also contains the true transcripts of the provided audio files.

Requirements

Python-Version: 3.7

All python packages needed to run the code are included in the 'requirements.txt' file. Read the section 'Usage - Run demo' on how to set up the python environment and on how to run the demo.

Usage

Setting up he python environment

1. Download this repository.

2. Create a virtual environment with Python 3.

   virtualenv -p python3 env1
   

3. Activate the environment.

   source env1/bin/activate
   

4. Install the required dependencies.

   pip3 install -r requirements.txt
   

Run demo

1. Set parameters in main_demo.py (line 5 to 7).

2. Set location for your Google credentials JSON-file in main_demo.py (line 11).

   • How to get the Google credentials JSON-file: click here

3. Copy the audio files (.mp3 or .wav) into the folder 'audio_raw/'.

4. Run main_demo.py. The generated transcripts are stored in the folder 'transcripts/' (folder will be created automatically).

References

• Google Cloud Speech-to-Text API