DVEdit: Features

Home | News | To-Do | Features | Examples | Tutorial | API Documentation | Download
There are hundreds of open source video editing programs and frameworks out there. DVEdit is just another, but it could quite possibly be the best tool for you.

Here are some of the features which tend to make DVEdit a little different from the rest. Hopefully this section will help you to decide if DVEdit is a good fit for your video editing needs:

Easy to learn - get started in minutes!
100% script based - no battling with fiddly, flaky or segfaulty graphical interfaces
Highly pythonic 'video clip as list' paradigm
Rich catalogue of Python extension classes encapsulating video content, and operations on video content
Simple but powerful class hierarchy offers great leverage and abstraction
Unlimited nesting and top-down structuring of video content
Uses DV format video - but will automatically/transparently decode from and render to other formats as needed
Write your video project files in Python, not some esoteric language or XML grammar
Gives high-level control over the structure of videos - from the raw input clips through to final rendering
Focus on performance - frames are only decoded from the raw DV format when absolutely necessary
Written in Cython (a Python-C interface language, successor to Pyrex) for speed
Plugin architecture is easily learned without sacrificing features or performance - start writing custom filters in minutes
Full frei0r plugin support - load and use without fuss
Create reusable components - code and media - using Python's leverage and rapid development facilities
If you want, you can build your own GUI over DVEdit
PIL (Python Imaging Library) interface - manipulate individual frames as PIL objects, save them to disk, load from disk etc

Aims

In designing DVEdit, I have aimed for a combination of characteristics which I believe are uncommon amongst the existing range of video editing software:

Easy to learn
Good performance
Simple to script
Very approachable plugin architecture, welcoming to new plugin authors
'Just-in-time' and 'only when necessary' approach to decoding and encoding:
- If a frame doesn't need any processing, it will be passed through in its encoded state, without incurring quality loss or wasting CPU cycles
- 'Pull' model - no processing actually happens until, when rendering or viewing a final output clip, a frame is requested. This request will bubble up through all applicable effect stacks until it reaches a physical Clip.