THE PERFECT ARCHITECTURE
In the spring of 2018, six months before the appearance of the camera INSTA TITAN, I attended the IVRPA TOKYO 2018 as a speaker. My goal was to introduce a new stereoscopic panorama capture technique that Jürgen Schrader, Roman Dudek and I had developed.
As represented in the workshop presentation leaflet, the capture consists of taking 8 photos around in the horizontal plane, with nodal misalignment in the longitudinal axis.
So to sum up, let’s shoot to the 8 cardinal points. This is an omnidirectional capture on a single plane.
The result is spectacular, but it has some shortcomings. The stereography is very good looking at the horizon, but it diminishes when we look up and down. The solution was very simple, I had to repeat that same capture, but in the three planes that form a Cartesian axis.
If we take 8 photos around, on the three axes, the sum is not 24. It is 18, because the shooting positions overlap when the axes cross.
This is easier to see in a geometric figure. In a Rhombicuboctahedron we will place a camera in each square space. So now it’s easier to count the 18 cameras. And what we have achieved is a 45° omnidirectional capture in three axes.
And if this capture is represented in a sphere, the result is the 6Dof video camera. I have called it “6Dof Leonardo”.
As you can see, we could say that it is an INSTA TITAN, with 8 cameras around it on three axes. It’s that easy and that simple. Optics, photography, stereography, panography are exact sciences, and this is the minimum configuration to obtain disocclusion in the vertical and horizontal axes, as described in the RED ONE MANIFOLD project.
Does this work? Have a look!
So, yes, it works perfectly. If it has worked in photo mode, there is no problem with it working in exactly the same way in video mode.
You will get four panoramas with lateral and vertical disocclusion, and then you will have to make depth maps for each view, and your 6Dof video will be ready for a VR HMD.
PATENTED US74… Active.
This system was patented in 2003 by a genius; Leonard P. Steuart III, who was 20 years ahead of his time.
Leonard P. Steuart III Make the following exposure:
“The digital 3D/360° camera system is an omnidirectional stereoscopic device for capturing image data that may be used to create a 3-dimensional model for presenting a 3D image, a 3D movie, or 3D animation. The device uses multiple digital cameras, arranged with overlapping fields of view, to capture image data covering an entire 360° scene. The data collected by one, or several, digital 3D/360° camera systems can be used to create a 3D model of a 360° scene by using triangulation of the image data within the overlapping fields of view.“
About FOV and resolution.
Leonard said: “It can be appreciated that selection of the camera fields of view involves a tradeoff between maximized coverage and resolution. An increased field of view will, of course, improve the amount of image data that can be captured, and in particular increase the amount of useful stereoscopic data as the size of the stereoscopic field of view is increased. However, as the view angle is increased the resolution is decreased. This is illustrated by considering a hypothetical digital camera with a resolution of 4,000 by 4,000 pixels and a field of view of 100°. Dividing 100° by 4,000 pixels gives coverage of 1/40 of a degree per pixel. This yields a resolution of about 1 cm at 23 meters. As digital camera technology improves the resolution will improve. Digital camera resolution has followed Moore’s law for semiconductors and has doubled about every year or two, and trend is expected to continue for at least the next fifteen years. An alternate solution to increasing the resolution is to simply use more cameras, each covering a smaller field of view”.
It was a surprise to find this patent, and even more so the name of its author. When I proposed this camera I gave it the name of Leonardo, but it was not due to Da Vinci, nor to P. Steuart III, but to a painter/lithographer panographer, Leonardo Barañano. So I reaffirm the decision, but it is Leonard P. Steuart who deserves it.
The technique has evolved a lot since 2003, as Leonard announced.
A patent describes an idea but its application is something very different. For a patent to work, many problems have to be solved that do not need to be detailed in the patent itself.
The architecture is not the same as the one I propose, but it is just as effective. Perhaps by using 18 cameras instead of the 24 Leonard proposes, the cost is reduced, as well as the dimensions of the camera. But the result will be the same.
Leonard proposes a capture in four horizontal planes, so that a horizontal and vertical disocclusion can be obtained. So if we increase the FOV proposed by Leonard to 190º. Its design works for 6Dof.
If you are interested in building this camera you must take into account an important detail, never before exposed, the interaxial distance. Or rather the “off-set” of the 18 entrance pupils, with respect to the geometric centre of the camera.
If the camera is too small, you will look through the eyes of a mouse, and everything will seem big, this steriopsis effect is called GIGANTISM. On the contrary, if the diameter of the camera is too big, you’ll look through the eyes of an elephant, and everything will seem too small, this effect is called miniaturization. So once you decide on the optics, and the sensors, you will need to perform several tests until you achieve the desired dimensionality.
Finally, you will need a rendering system optimised for 6Dof, that is to say a “soft” that allows you to compose the different panoramas.
One step beyond!
6Dof photography is something completely different in the way it tells the story. No radioactive zombie will attack you from any angle. It will be something much more natural and simple; you will simply walk into a photograph. You can visit a museum, play “Scape Room”, take a walk around IKEA, or visit Cologne’s Gothic Cathedral.
Ready player two?