Planet LSU Computer Society

I've been thinking for a long time about B-roll footage. It takes a long time to edit video so that the audio lines up in both sources, then to cut between them, i.e./ One audio source, two video feeds and you choose "the best".

From what I understand, to edit a simple 5 minute piece of footage like this takes about an hour, replaying the same piece over and over until you can find the right splice points - or you're just very experienced and can split it at the right moments, risking a potentially dodgy cut.

I've been learning Python3.0 recently, with the eventual goal of being able to do the following in a nice little programme:

• Play a section of footage, roll A. Hit a cue button when something simultaneously happens on both cameras/audio feeds.
• Play another section of footage, roll B. Hit a cue button in the same circumstances. You now have the two pieces of film synced.
• Create a grid of four spaces: Top Left is Roll A. Top Right is Roll B. Bottom Left is Output, Bottom Right is a summary of useful information (Timecodes, Project Name, Intended Duration, Editors Notes etc)
• Play the footage, and have be able to hit the space bar when you'd like a cut - i.e./ you think a cut is needed there.
• When reaching the end, hit a button that allows you to easily fine tune the cuts - it takes you to 2 seconds before the first cut, and you can pan through the footage with the mouse wheel to the exact frame you want to splice, and move the cut there. Rinse and Repeat.
• Now optionally choose an audio feed you'd like to use, be it seperately recorded audio, or a feed from one of the cameras.
• Then, you're handed an XML save file, which has a list of the attributes associated with that footage.

You can then go and edit the next piece of footage. What would happen is that on a second machine, you can drop in the two reels, the XML file and run a programme. You run a piece of software, which reads in the footage as necessary and either re-renders it to an output file, or in the case of formats that supports it, just chops and concatenates.

What you're left with is a piece of footage you can import into a video editor and remove the flubbs etc, without having to worry about desyncing the video stream.

Seems quite simple to implement, just wondered what other people thought - a worthy project? It could quite easily be extended to use more than two streams, or even be turned into a "gallery" style programme dealing with live programming outputing to a save file and a stream out to VLC or something...

Thoughts, questions, comments welcomed.