Neat video

[Dave Haynie]

If I want to composite video, there are of course a large number of options. But the goal here is a straight averaging of the video. You can set up compositing groups, but the simplest is just to add 1/N of each of N videos together.

If you have two captures, put them on two tracks, one over the other, then set the top track to 1/2 transparent. This effectively adds half of each together. Or start with three, and set the top two to 1/3 transparent.

Alternately, you can create a compositing group, and set each video in the group to a level of 1/N, then set the compositing operation to add. This, of course, lets you actually add to more than 100% brightness, whereas with the transparency method, you’ll always have your video at exactly 100%, the transparency of the top (N-1) layers is simply balancing the contribution of each layer to that total.

It’s simple to do a few short captures and play around with it. Again, if the noise is actually random, each video will have the same content but different noise. There could be damage to the tape that causes more patterned noise, but in general, you get random noise. Another application of this is low-light noise reduction. With video or still, take N shots of the same still scene in very low light. You’ll probably see sensor noise on each image, but when you average them, the noise itself largely cancels out (particularly with modern digital still cameras, the photo process itself already cancels out any patterned noise that might be present in the sensor).

-Dave

[Allan Boes]

Hey Dave

Sorry for the probly stupid question, but what does 1/N mean?

Also how do I get my track to perfectly align, cause my two recordings do not start at the same place.
I can use the audio to get it close,but how do I really fine tune it?

Allan Boes!!!
Working in Vegas pro 10 and Adobe photoshop CS5

[Dave Haynie]

N = total number of captures
1/N = weight of each capture in the final render

So, if I capture twice, N=2. I want each capture to contribute to 1/2 of the total image… I spoke of two different ways to do that. If I have three samples, that’s N=3… each will contribute 1/3 to the total of each frame.

For sync, I use audio primarily. Look for a good sync point — a clapboard, a cymbal crash, some other impulse. You have to zoom way in in Vegas to be able to align things on a frame-by-frame basis. There are tools like PluralEyes that can align video by audio, but that’s pretty much overkill for this exercise. Once you’re close and looking at single frames, you may be able to eyeball the video sync directly — look for a high motion event, with lots of change between the frames. But the audio’s probably necessary to get you close.

The one danger with tape is that it stretches. Hopefully, the analog to digital capture device is acting like a time-base corrector, giving you an honest frame by frame capture with nothing dropped.. not always the case, depends on the hardware.

-Dave