A Comparison of Wiggly Techniques for 3D Images

As we all know, not all techniques are created equal. Sometimes, you've got crazy things that work amazingly well and you've got highly complicated things which don't really reached the level of success you would like.

I've tinkered around with some of the images that I have in my library to see what can be done in the intersection between these notional spaces. All of the images are from my Lytro, light-field library, meaning that the native source is roughly 1030 x 1030 with the addition of a 330 x 330 grayscale depth map. Given that data, I can create an image from the seven 1030 x 1030 slightly offset image slices that Lytro can export, align and stack them in Photoshop, then feed them into another tool for presentation altogether: in this case, Depthy.

Let's start with something simple, the key ring that came up with my Homeworld Collectors Edition preorder:

Homeworld on a chain. Pointedly, not in use.

This is animated with #Depthy from the aforementioned source, then hung out on Glfycat for the folks who need more constrained bandwidth.

Compare to the same piece from the Lytro library:

Roughly the same size when tossed to the blog and the latter has a lot heavier JavaScript rendering skin, Notably, beyond the expected considerable improvement in smoothness and shifting the point of view, the pre-animated GIF version doesn't look that bad.

Of course, it doesn't have assuming or panning or anything interactive going on, which gives it a significant advantage. But if you're trying to convey information in 3-D, you don't necessarily need a fully interactive presentation process. Though it never hurts.

Let's look at something a little more complex and complicated, not least reason being that we need a harder test on a less quality image.

I took this next image in a graveyard in Snellville, GA, and it's particularly difficult to work with because the depth map is more than a little spotty. Because of the organization of elements it was extremely difficult to get clean depth information for each plane.

This time, let's try three different visualization methodologies.

First, a simple, straight up Depthy re-visualization.

Game over, man. Game over.

Not bad, really. Dramatic, but the sense of motion is somewhat muted (Depth map creation on the first generation Lytro cameras is not particularly impressive at rtange; awesome macro, not so good at length.)

Let's compare to an export from the Lytro Desktop app in MP4 / video format.