Skip to content

abdeLKabir-56/Whisper-ASR-Transcription-Project

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

3 Commits
 
 
 
 

Repository files navigation

Whisper ASR Transcription Project

Overview

This project utilizes the Whisper Automatic Speech Recognition (ASR) model developed by OpenAI to transcribe audio from both a locally hosted video and a YouTube video. The transcription results are saved as text files for further analysis.

Dependencies

  • OpenAI Whisper: A powerful ASR model for transcribing speech.
  • pytube: A lightweight, dependency-free Python library to download YouTube videos.
  • moviepy: A video editing library for Python.
  • ffmpeg: A multimedia framework to handle audio and video processing.

Usage

  1. Install the required dependencies:

    pip install -U openai-whisper
    pip install ffmpeg
    pip install pytube
    pip install moviepy
  2. Load the Whisper ASR model:

    import whisper
    model = whisper.load_model("tiny")
  3. Run the provided Python script:

    • Local Video Transcription:

      from moviepy.editor import VideoFileClip
      video = VideoFileClip("video.mp4")
      video.audio.write_audiofile("myaudio.mp3")
      
      result_local = model.transcribe("myaudio.mp3", fp16=False)
      with open("mySound_local.txt", "w") as file:
          file.write(result_local["text"])
      print(result_local["text"])
    • YouTube Video Transcription:

      from pytube import YouTube
      yt = YouTube('https://www.youtube.com/watch?v=Aq92xxwYwSU')
      
      stream = yt.streams.get_highest_resolution()
      video_path = stream.download()
      
      video_youtube = VideoFileClip(video_path)
      video_youtube.audio.write_audiofile("myaudio.mp3")
      
      result_youtube = model.transcribe("myaudio.mp3", fp16=False)
      with open("mySound_youtube.txt", "w") as file:
          file.write(result_youtube["text"])
      print(result_youtube["text"])
  4. Analyze the transcriptions saved in the generated text files (mySound_local.txt and mySound_youtube.txt).