Audio/Video to Text Extractor with Google Cloud Speech-to-Text

This Python script extracts speech from audio or video files using Google Cloud Speech-to-Text and stores the transcribed text in a file. It automatically handles uploading long audio files to Google Cloud Storage (GCS) and processes both short and long audio files using Google Cloud's transcription APIs.

Features

• Extract speech from both audio and video files (MP4, MKV, MOV, AVI).
• Converts audio to the required format (16kHz, 16-bit mono PCM WAV).
• Automatically uploads long audio files to Google Cloud Storage (GCS) for transcription.
• Uses Google Cloud Speech-to-Text for transcribing speech from audio/video.
• CLI-based usage for easy audio/video file processing.

Requirements

• Python 3.12+
• FFmpeg (for audio extraction and conversion)
• Google Cloud Speech-to-Text API and Google Cloud Storage
• Google Cloud Credentials (Service Account JSON)

Google Cloud Setup

1. Enable the Google Cloud Speech-to-Text API in the Google Cloud Console.
2. Create a Google Cloud Storage bucket to upload audio files for transcription.
3. Set up a service account with appropriate permissions and download the service account JSON key file.
4. Set the environment variable for Google Cloud credentials:

   export GOOGLE_APPLICATION_CREDENTIALS="/path/to/your/credentials.json"

Installation

First, ensure FFmpeg is installed:

• On Linux:

  sudo apt-get install ffmpeg

• On macOS:

  brew install ffmpeg

• On Windows, download the installer from the FFmpeg website.

Install the required dependencies:

poetry install

Usage

You can run the script to extract speech from an audio or video file and save the transcribed text to a file.

Command-Line Interface (CLI) Example:

poetry run audi input_file/video_file.mpg -o output_text.txt --bucket your_gcs_bucket

• audio_path_to_audio_or_video_file: The path to the audio or video file you want to transcribe.
• output_text.txt: The output file where the transcribed text will be saved.
• your_gcs_bucket: The Google Cloud Storage bucket where the audio file will be uploaded for transcription.

License

This project is licensed under the MIT License.