typed-ffmpeg offers a modern, Pythonic interface to FFmpeg, providing extensive support for complex filters with detailed typing and documentation. Inspired by ffmpeg-python
, this package enhances functionality by addressing common limitations, such as lack of IDE integration and comprehensive typing, while also introducing new features like JSON serialization of filter graphs and automatic FFmpeg validation.
graphviz
for visual representation, aiding in understanding and debugging.Please note that the following features are under consideration or development for future releases:
typed-ffmpeg
is currently built with FFmpeg version 6.0 in mind, we are working to ensure compatibility across different FFmpeg versions. Feedback and issue reports are welcome to improve version support.typed-ffmpeg
. Continuous updates will be made to include more complex and varied filters.To install typed-ffmpeg
, simply use pip:
pip install typed-ffmpeg
Note: FFmpeg must be installed on your system.
To enable graph visualization features:
pip install 'typed-ffmpeg[graph]'
Note: This requires Graphviz to be installed on your system.
Here's how to quickly start using typed-ffmpeg
:
import ffmpeg
# Flip video horizontally and output
f = (
ffmpeg
.input(filename='input.mp4')
.hflip()
.output(filename='output.mp4')
)
f
For a more complex example:
import ffmpeg.filters
import ffmpeg
# Complex filter graph example
in_file = ffmpeg.input("input.mp4")
overlay_file = ffmpeg.input("overlay.png")
f = (
ffmpeg.filters
.concat(
in_file.trim(start_frame=10, end_frame=20),
in_file.trim(start_frame=30, end_frame=40),
)
.video(0)
.overlay(overlay_file.hflip())
.drawbox(x="50", y="50", width="120", height="120", color="red", thickness="5")
.output(filename="out.mp4")
)
f
See the Usage section in our documentation for more examples and detailed guides.
This project was initially inspired by the capabilities of GPT-3, with the original concept focusing on using GPT-3 to generate an FFmpeg filter SDK directly from the FFmpeg documentation. However, during the development process, I encountered limitations with GPT-3's ability to fully automate this task.
As a result, I shifted to traditional code generation methods to complete the SDK, ensuring a more robust and reliable tool. Despite this change in approach, both GitHub Copilot and GPT-3 were instrumental in accelerating the development process, providing valuable insights and saving significant time.
I would also like to extend my gratitude to the ffmpeg-python
project, which inspired this project significantly. The API style and design ideas from ffmpeg-python
have been influential, and I have utilized these aspects to shape the development of our SDK.
This project is dedicated to my son, Austin, on his seventh birthday (February 24, 2024), whose curiosity and zest for life continually inspire me.
Feel free to check the Documentation for detailed information and more advanced features.