Gimp combined with GAP has some interesting possibilities in creating composite 3d objects. Using an old technique called slit-scan photography and applying it as a concept using Gimp and GAP allows this to be done.

I created this simple pattern:



Applying Map Object I created the front of a cylindrical looking object:



Naturally, you can vary the width (and length) of your cylinder. A narrower cylinder could be also produced from the same base pattern:



Please consider the following animation:



It is a 300 frame animation, where the white line moves by one pixel difference between each successive frame. Almost looks like a scanner light as it moves across an object.

At some layer early in the animation, say layer alpha, it may be at this position:



At some layer "a" later in the animation, it may be here:



And yet at a later layer "b" it would be here:



Imagine applying each of those different layers as Layermasks to the various cylinders.

Layer "alpha" was applied as a layermask to the narrower cylinder resulting in this:



Layer "a" was applied as a layermask to the regular sized cylinder resulting in this:



Layer "b" was applied as a layermask also to the regular sized cylinder yielding this:



Now, if I composite all three of those images, I get this:



Attempting to do this manually for 300 layers, varying the size of the cylinder and using a different "scan animation" layer as a layer mask would be quite a daunting task. However, thanks to GAP, one can create 300 frames, use Frames Modify and Apply Filter on Layer(s) along with Map Object to create varying sizes of cylinders. One can then use Move Path to separately place each separate "scan animation" layer into each separate frame, and then apply that layer as a layermask. When I did that I was able to get a resulting composite image.

My first efforts, used a 1 pixel wide white line "scan animation", where the layers were applied a layermasks to varying size cylinders. The result was, um, different:



I noticed that using the 1 pixel width line resulted in a diffraction pattern being created on the composite object. Interesting to look at, but not the intended effect.

After some experimenting with different sized lines, I settled on a 3 pixel width line in the scan animation. The resulting object was this:



Now, my imagination asks, what if I began combining boxes, spheres, planes, and cylinders in such a procedure?

The next step in the challenge would be to make a 3d rotation of this composite object using Gimp and GAP. It is possible. It would take a lot of frames to accomplish, since it take 300 frames to make a single still. If I wanted a 60 frame simple rotation animation of that composite object, it would take at maximum 300x60 or 18,000 frames. That would be for only a six second animation. Extreme!

Anyone up for the challenge?



Truthfully, once you've gotten to this stage, using a 3d application, such as Blender is best. You sure wouldn't need 18,000 frames just to make that composite object rotate in Blender. It's just more of a personal curiosity of using Gimp and GAP to their limits. As was mentioned in the title, it's "Extreme GAP'ing"

Thank you for checking this out.

This is quite interesting. Though I don't really have much to add, I appreciate your sharing it with us.

One thing that might be useful, you can use onion-skinning to incorporate a previously rendered frame into the next one; thus your last frame in the sequence will be a "composited" version of the all the frames (you can delete the others). It is probably simpler to convert the frames to an image and merge the layers, but doing so can in some cases require more memory than available.

saulgoode,

Thanks. I had wondered previously about doing composite 3d objects based on the basic one's Gimp does. It was six months to a year ago that I was able to first figure out this process. I had some exposure the the slit-scan photography concept, and one day thought I would try it with Gimp, where the slits act like layermasks and for some locations of the slit, I would use varying sized objects. I may try using a curved slit, by making a very thin cylinder, like with the Gimp coin animation to see if that is any more effective.

Using the onionskin I believe would work after I applied the layers of the scan animation as layermasks to the various sized cylinders to get the various slices. I may try that to see if I could then reduce the time required to make the object appear to rotate. Thanks for that insight.

Thank you for commenting and for the helpful suggestion.