AI-Powered Automated Video Transcription
With the advancement of artificial intelligence, complex programming tasks have become increasingly accessible. For instance, the maturity of Automatic Speech Recognition (ASR) technology allows individual developers to easily create efficient tools. This article will demonstrate the initial steps of generating English subtitles for a video, showing how to automatically extract and transcribe audio from video files by combining open-source technologies like Python, FFmpeg, and OpenAI Whisper.
Core Tech Stack Analysis
The construction of this tool relies on three key technologies:
- Python: As the most mainstream programming language in the fields of artificial intelligence and data science, Python boasts a large and active community and a rich ecosystem of third-party libraries, making it the first choice for implementing the logic control and integration of this project.
- FFmpeg: This is a powerful open-source multimedia processing framework capable of handling almost all formats of audio and video files. In this project, it is mainly responsible for losslessly extracting the audio stream from video files for use by the subsequent speech recognition model.
- OpenAI Whisper: An automatic speech recognition model open-sourced by OpenAI in 2022. The model was trained on a massive and diverse dataset, exhibiting extremely high recognition accuracy and robustness for multiple languages, especially English. It can directly process audio files and output text with timestamps, serving as the core engine of this project.
Windows Environment Setup Guide
Before starting to code, you need to configure the development environment on the Windows operating system. It is recommended to run the following operations with administrator privileges to ensure smooth installation of dependencies and automatic configuration of environment variables.
1. Install Python
You can install the latest version of Python with a single command using the Windows Package Manager (winget).
bash
winget install Python.Python3
After the installation is complete, reopen the command line terminal and run python --version to verify.
2. Install FFmpeg
It is recommended to install FFmpeg via the Chocolatey package manager, which will automatically handle the environment variable configuration.
First, ensure Chocolatey is installed. If not, run the following command in an administrator PowerShell:
powershell
Set-ExecutionPolicy Bypass -Scope Process -Force; [System.Net.ServicePointManager]::SecurityProtocol = [System.Net.ServicePointManager]::SecurityProtocol -bor 3072; iex ((New-Object System.Net.WebClient).DownloadString(‘https://community.chocolatey.org/install.ps1’))
Then, use Chocolatey to install FFmpeg:
bash
choco install ffmpeg -y
After installation, restart the terminal and run ffmpeg -version to confirm a successful installation.
3. Install Python Dependencies and Whisper Model
This project depends on several Python libraries, including the core openai-whisper library and auxiliary libraries for audio processing. At the same time, you need to download the Whisper model file in advance. The small model is recommended as it strikes a good balance between recognition performance and running speed.
bash
Install base libraries
pip install ffmpeg-python moviepy librosa soundfile openai-whisper pyannote.audio
Download the Whisper small model
whisper download-model small
The pyannote.audio library is used for speaker diarization. Although not directly used in the current script, installing it in advance prepares for future feature expansion (such as distinguishing different speakers). Please note that using pyannote may require obtaining a free license by following the instructions on its official website.
Speech Transcription Script Implementation
Once the environment is ready, you can write the core transcription script. The following Python code implements the complete process of loading the Whisper model, reading a video file, performing speech recognition, and saving the results to a text file.
Save the code as a transcribe.py file:
python
import whisper
import os
Define the path to the video file to be processed
audio_path = “test.mp4” # Please replace with your target file
output_file = “transcription_result.txt”
1. Load the Whisper model
print(“Loading Whisper small model…”)
try:
model = whisper.load_model(“small”)
print(“✅ Whisper small model loaded successfully!”)
except Exception as e:
print(f"❌ Model loading failed: {e}")
exit()
2. Check if the file exists and perform transcription
if not os.path.exists(audio_path):
print(f"⚠️ File not found: {audio_path}, skipping transcription.")
else:
try:
print(f"Transcribing file: {audio_path}…")
Call the transcribe method for recognition
result = model.transcribe(audio_path)
# Print and save the recognition result
transcribed_text = result["text"]
print("✅ Audio transcription successful!")
print("\
Transcription content:\
“ + ”-“20 + f”\
{transcribed_text}\
“ + ”-"20)
with open(output_file, "w", encoding="utf-8") as f:
f.write(transcribed_text)
print(f"✅ Transcription result has been saved to: {output_file}")
except Exception as e:
print(f"❌ An error occurred during transcription: {e}. Please check if FFmpeg is installed correctly and configured in the system's environment variables.")
When using this script, make sure to replace the value of the audio_path variable with the name of your own video or audio file and place it in the same directory as the script. After running the script, the recognized text will be output to the console and saved as transcription_result.txt.
Conclusion and Outlook
This article demonstrates the basic workflow of using modern AI tools for automated content processing. By following the steps above, any user with basic computer knowledge can efficiently convert the speech content in videos into text. This is just the first step. On this basis, by analyzing the more detailed results output by Whisper (including timestamp information), one can further develop applications that generate bilingual subtitle files in .srt or .ass format, achieving more complex application scenarios.
