Hi,
Very new-pls excuse obvious answers.
I am working on tracking video onto balloons in darkness.
This is what I have done with Madmapper and Modul8 and a Kinect. A bit slow and no way to choose which section of the video gets projected.
I am looking to track video of peoples faces talking as the balloons move slowly across a designated area.
So far I have this file - which is fun and tracks well, but doesn’t do what I am after Balloon Tracking 1.5.toe (3.85 KB)
Question 1 - What is the easiest way to project onto nothing except the balloons ( masking is the answer which I can do in other progs but cant see the easy node or tutorial for this)
How do I project the section of video that I want to rather than just unmask an area of the screen (if that makes sense).
I have been advised by an advanced user that Kamschnapper is the answer to mapping- and I am doing a course in a couple of months
Just looking for a few more pointers so I don’t go down the wrong rabbit hole.
How are your movie files of the faces? Are they full sized assets you need to cut parts out of, or have you done the cutting and it’s a small asset of a face?
I don’t know if I’d use CamSchnapper for this, since you’re not really doing “mapping” (at least in my definition of it), you just want to hit the balloons with light.
If you connect an Info DAT to the Blob Track TOP you’re using, you get ids, x/y position, and width and height for each blob.
You can use the width and height to resize your face videos to the balloon size, then have them transform over your full size canvas at those x/y coordinates. Then you can make masks one of several ways depending on the situation, a quick way to make them is to just subtract a picture of your background from the output of the Blob Track TOP, then push the remainder into white masks with Channel Mix TOP (set everything but constants to 1). You would then multiply those masks by the floating movies you already have, which would just cut out the square edges and make them look like balloons.
The Blob Track has an output mask option, but its generally pretty taxing on the machine.
Tracking INFO is great - thanks
I would like to spherise the video and map it to the width n height (as advised)
Having gone through many of the vid tutorials I can’t quite see how to do this?
Anyone got a moment to point me to relevant tutes or material for this particular one?
Thanks
Any idea how to make Geo read Kinect data? I am trying to get the sphere to track to the blob tracking co-ordinates. If I put 2 instances in the geo and map the kinect this should work?
I understand a basic instancing from a noise CHOP but can’t get the Kinect data to flow through, even through renaming and merging.
