.

I used some footage from my friend Lu’s poledance show last year in Tokyo to experiment with datamoshing. I did it the hard way, by encoding each segment with only a first i-frame (keyframe), and then proceeded to remove them too with a different piece of software. I learned about some technicalities of encoding with h264 and avi file formats, which may end up being useful some time in the future.

The way video is compressed for some web formats, there can be graphical errors when skipping through the timeline, due to the infrequency of key frames, or video frames which consist of complete graphical image information.  The subsequent (inbetween) frames contain information related to the movements and changes of graphical information.  To put it as simply as possible, there is a picture followed by information about how to move specific part of the picture, another picture, and then more movement information. When the graphic image frames are discarded, the inbetween frames continue to move and change inaccurate information, leading to an interesting result.

.

.

All the software for osx can be found (for free too) at: http://www.court13.com/datamoshkit.zip

Thanks to the datamosher tutorial vids: http://www.youtube.com/watch?v=tYytVzbPky8 Apparently there’s an easy way, an now that I ‘understand the principle, I’ll try it in AfterEffects.