3D pairs captured from video clips (cross-eye viewing method)

I have prepared to show you some pairs of 3D images extracted from videos (ready to be viewed with VR devices). In the original format they are placed in parallel mode, for the presentation I reversed them so they can be viewed in cross-eye mode.
Filming was done in 4k and captures were resized/reduced.
I used a DIY system with two lenses with an inter-axial distance of 9mm.
I also use a system with two cameras in parallel and telephoto lenses for close-up from a greater distances, and another system with two cameras and beamsplitter (less often, because it is large, heavy, and difficult to adjust).

I like them and I like cross eyes.

Thank you!

Amazing quality.

Nice pairs. Too bad for the bee... Grasshopper is pretty interesting.

Regards.

Ron

Thanks!

Another one with the same grasshopper.

Take-off...

Héhé!!! Nice shot! Had it been only 2D, I thing I'd never have noticed that it is scratching its right wing with its middle leg.

Ron

I'm not sure if it's scratching or have other significance. The front small leg is also put in vertical position. If studied frame by frame, seems to be part of take-off procedure.

Here is a frame capture from a 3d clip

The clip was made with two cameras. The WB can be adjusted to look similar, but focus plane needs to be carefully set to be the same for both cameras.

wow these are awesome!! The 3D effect is amazing and they look too beautiful

Are the cameras fully parallel or is there some toe-in?

As far as I'm aware, all other "professional" photos of insects in 2D are obsolete and might as well be black and white on newspaper-quality paper compared to this

Thanks!

The macro ones are made generally with one camera and a special device with two lenses (example with the bee ), sometimes with two cameras and a small 3D beam splitter.

Here, they are for cross eye view mode, but initially are intended to be viewed with VR devices in parallel mode.
The example with the waterfall is made with two cameras in parallel (without axial convergence)





.

Here is a link to the 3d clip


The 3d close up of the Beetles are done with two camera on a rig. The rig was made by myself and has a half-mirror ( beam splitter ) in it. There are many problems with macro filming using two cameras. One is getting both cameras to be in focus on the same point in the scene. For example one camera might focus on the insects feet and the other camera on the insects head and with a small depth of focus which you get on macro , it can effect the overall 3d effect. Another is white balance should be the same on both cameras. Another is alignment between the two cameras - there needs to be a bit of toe-in otherwise when aligning the two clips in software some of the image is wasted at the sides. This means lost pixels and lower resolution on the final combined image. For example the two cameras might be 1280x720 pixels ( 120fps to allow slow motion and better alignment in timing which is necessary ). In software the two images are aligned and after that the useful remaining image may only be say 900x 530 pixels for good 3d alignment.

Here ia a link to the image of the rig with half-mirror and with two cameras

Brian F Flint said:

There are many problems with macro filming using two cameras....

Click to expand...

I like 3D materials to have as many details as possible.
In 2012 I bought two Nikon D5100 cameras and built a beam splitter with a 50T / 50R mirror. I made a series of 3D videos shot in 1920x1080p, but I was not satisfied with the quality of the details provided.
The first capture is from a 3D clip with a size of 3840x1080 pixels (1920x1080/eye), reduced to 3000x844 pixels, so 1500x844 pixels / eye.
Not very bad, but I wanted more.





So, in 2014 I bought two Panasonic GH4 cameras that shoot in 4K, and in 2016 two Sony a6300 cameras that also shoot in 4K, but at a higher quality than the one offered by GH4.

The second capture is made from a single 4K frame that contains the images for both eyes, half a frame for each eye. I chose to contain an ant for comparison. Although it is 60% crop of the original content, it still offers a much higher level of detail than shooting with two cameras in 1080p in 2012. This system is easier to use, but the working distance is short.





With two 4k cameras I can use long focal length lenses, which allow a considerably longer working distance, useful for scared subjects, and the image quality resulting from 2x4k is significantly better.

My new videos are designed to be viewed with VR devices. As a result, the framing of the subjects makes it almost centered and with generous space around the subjects. This allows the image to be magnified so that the subject in the field of view is isolated (as example to the size of the grid marked in blue), and the head tracking system allows some observation of the environment around the subject, for better immersion.