4) The Stereopanography

  1. Intro.
  2. Historical review.
  3. The Stereography.
  4. The Stereopanography.
  5. Research and development.
  6. Stereoscopic panorama capture.
  7. Capture templates.
  8. Stereo stitching.
  9. Prosumer
  10. Publication.

  • 4.0)     Antagonist techniques
  • 4.1)     Points of union
  • 4.2)     Interaxial and safety distance
  • 4.3)     Dwarfism
  • 4.4)     Adjusting the parallax in post-production.
  • 4.5)     Native Pixel Parallax
  • 4.6)     The stereo window.
  • 4.7)     Ballon efect.
  • 4.8)     Shooting and adjusting the nadir.
  • 4.9)     Lens & Focal Length selection.
  • 4.10)    Trigger Synchronization.
  • 4.11)    Cromatic aberration and pixel effect.

 
4.0) Antagonist techniques.

 

DSS-Capulets-Montagues-300x221Stereography and panography are two antagonistic photographic techniques, the first requires the existence of the parallax while the second hates it. And their differences do not end here, the panography adores the “fish eyes”  while the stereography does not want to see it. There is even  a synchronization problem, because the stereography needs it, but the panography does not allow… So bringing these two families together requires a lot of tact!


4.1) Points of union.

 

But it’s not all bad news, panography and stereography share important preferences. Both techniques like to have a great depth of field and in both techniques the use of  the high dynamic range is good. And when shooting both also share the idea of making all adjustments in manual mode.

Other aspects to take into account are perfectly known in panography; we will turn off the self-stabilization, manually adjust the white balance adjustment and copy from one camera to another all the capture parameters.

gopro3dgafas copia

oasisanag
HDR 3D 360º

4.2) Interaxial and safety distance

 

The first thing we must make clear is that if it is stereo it is not nodal and if it is not nodal we will have stitching parallax problems, so we will use the correct interaxial distance, in order to use the minimum nodal desalination.

InterTo obtain the minimum possible interaxial distance, we recommend the use of DSLR´s as small as possible. In the case of the Sony Alpha 5 series, the camera body is even smaller than the diameter of the E-mount optics, so your interaxial distance will be equal to the diameter of the optics used, in this case as specified by the manufacturer; 6.2 cm.

 

interior

To calculate our safety distance we will use the constant (30), so multiplying 6.2cm x 30 we get 186cm.

Tripod height: It is independent of the safety distance, but must be kept within its margins, if our safety distance is 1.20cm this will be the minimum height of the tripod but not the maximum.

Imagine that your safety distance is a soap bubble that inflates with the epicentre in your entrance pupil, which should explode when it touches the ground. If it is punctured before touching the ground it is because it has touched some element within this space.

In the following three examples the same configuration has been used, interaxial 5,4cm, 10mm, two rows of 60º, with +-45º each one. In the first one, the double bed, the stereography is correct, in the second one, the bathroom, some tensions appear that translate into disparities and in the third one the room was very small and the camera was too close to the wall, which has generated a great tension in all the seams and the interpolation has had to be stretched more than possible.

In the fourth photo the walls are very close together, almost as much as  a Lego room, but by reducing the interaxial distance to only 2 centimetres, the volumetry is similar to that of the real world and no errors can be seen.

 

Click on the images to activate the VR 3D WEB mode:

2   1   3

4

 


4.3) Dwarfism

 

If we work with twin RIG cameras, it is very important that we mount them in an orthoestereoscopic range, because if we don’t do it an effect will appear that we won’t know why it happens, it is called dwarfism.

It is an effect that occurs at the neuronal level when we visualize a scene in which a greater interaxial distance than necessary has been used, which produces an effect of miniaturization in objects and people. If we see two 2D photographs taken with Ortho and Hyper-interaxial distances, people will look the same size, but when viewed in 3D mode in a magical way, “steriopsis” these people look very small.

This effect was well known after the success of AVATAR and the attempt by videographers to perform stereography using two HD DSLR´s, which when placed side by side show a hyperstereoscopic interaxial distance.

The example shown below has been made using the “rotational” technique, 8 shoots,  and at an interaxial distance of 8.2 cm, “Insta Titan” the model was placed two meters from the camera, and its extended arm extends 50 cm.

 gopro3dgafas copia

6

This does not mean that we cannot use hyperstereoscopic bases, we can use it whenever we decide, knowing that we are going to need a tripod or a pole that reaches the necessary height.


4.4) Adjusting the parallax in post-production.

 

It is important to know that post-production parallax adjustment will cause objects to move closer or further away, but in no case are we varying the interaxial distance,  we are simply adjusting the negative and positive parallaxes to avoid both hyperconvergence and divergence. In the lower panorama we see the area visible on the screen of the Oculus Go:

gopro3dgafas copia

para1

Now let’s take a closer look at this area:

recorte1.jpg

With the naked eye and without the need for anaglyph glasses we can read the halos it produces: The background has hardly any coloured halos so we assume that at this time the zero parallax is in the back door.  At the same time we see how both the model and the cup that she is holding have  marked cyan red halos, so they are in negative parallax and near to producing a hyperconvergence.

Now we are going to distribute the parallax so that the vision is more comfortable for the spectator:

para2.jpg

We can see that now the zero parallax fits the model’s eyes, which will be represented on the screen plane, the cup enters the negative parallax zone without producing hyperconvergence and we feel the sensation that the door is now further back, because we have given a slight positive parallax within a correct range and without producing divergence.

 

 

 

To adjust the parallax in Photoshop, the easiest thing is to create a project with the two panoramas right and left. We will increase to double the width of the project, we will duplicate both the right and the left and we will join them.

 

doble.jpg

Now double-click on the top layer containing the two panoramas on the right side and deactivate the red RGB information.

anag copia

Finally we will look for an alignment in which the anaglyph halos are as discrete as possible in the composition as a whole and we will cut them with an aspect ratio of 2:1.

gopro3dgafas copia

fin


4.5) Native Pixel Parallax

The native parallel pixel indicates the maximum positive parallax we can use for a given screen size. In classic stereography it is easy to calculate, but in 360º it is necessary to do some previous calculations. In the image below, a box of discontinuous red lines marks the visible area in a Oculus Go.

native pixel paralax

According to data provided by the manufacturer the resolution of each screen of the Oculus Go is 1280 x 1440px per screen, so regardless of the size of our stereopanorama 1440px is the width of our screen, the problem is the impossibility of calculating the size of the screen after the enlargement produced by the optical system, so in the stereopanography,  we ´ll try to adjust to zero the parallax of the intermediate parts of the composition.

sl_website-under-construction


4.6) The stereo window & negative paralax.

We have arrived at a complex part. In previous points we saw that the stereo window is the one that will define the neutral parallax, positive or negative, however in the VR mode there is no apparent frame.  We can decide and choose our stereo window, but if the area in negative parallax is greater than the visual field of our VR Gear, we will not be able to find a position in which the effect is not broken.

 

anag2
Equirectangular

In the first photo we see an equirectangular projection with a curious effect of negative parallax that we see amplified in the second photo, while in the third we see that it is not possible to frame the effect, since our visual FOV, is inferior to the zone in negative parallax. We have created a violation of  the 360º “stereo window”.

gopro3dgafas copia

zoom
Zoom
anag3
VR Area

 


4.7  Ballon effect.

StereoWindowgopro3dgafas copia

The steriopsis produces magical effects, the next 3 photos are exactly the same, the only difference is the parallax adjustment in postproduction. In the first we see an adjustment in which negative and positive parallaxes have been distributed equally, in the second the parallax is -15 pixels and in the third +15.

The result is that the stereoscopic spheres inflate and deflate, a curious effect never described before that will help us to adjust the volumetry in the VR environment, because what we are looking for is that the sizes are the same with the naked eye as if we were using a Vr Gear.

 

Click on the photos:


 

4.8) Shooting and adjusting the nadir.

 

Handling the nadir is something that any panographer knows how to do using different means.  However the handling of the stereo nadir is complex because it does not permit manual correction, since a slight misalignment during this patching process will result in an irregular soil with holes and small hills that mysteriously appear during visualization.

Regardless of the capture technique used, in order to manage the nadir we must take 2  nodal photos of the ground at 90º, and one last hand raised photo trying to maintain the nodal alignment.

What we are going to do is prepare the photo of the Nadir, before making the stitching, so that it adjusts automatically.

First we’ll use the two photos of the Nadir with the tripod to clean as much as we can, and finally we’ll use the photo taken without a tripod to clean up the little space we have left.

Now we can proceed with the stitching with the guarantee that our nadir will not have undulations and deformations, and will be stitched automatically.


 

4.9) Lens & Focal Length selection

 

3Dojodepez.jpg

There is a false belief that fisheye lenses cannot be used in stereography, mainly because of their difficulty of alignment. If the cameras are not properly aligned, the fish’s eyes will create a vertical parallax error inside the photo, something almost impossible to repair.

The question we must ask ourselves is different, do we have fish eyes or human eyes?    Well, if we try to look through the eyes of a fish, we will see an increase in stereoscopic sensation, causing that even using orthostereoscopic bases the effect is not natural, it is excessive.

The options are to reduce the interaxial distance or to use an optic more similar to that of the human eye. Moreover, if we are going to portray human figures, it will be better to use smaller aperture optics that provide us with sizes more in line with reality.

In the following photo a 12 mm optics was used.

human


 

4.10)  Trigger Synchronization.

 

First of all, we must bear in mind that in any panography there is going to be a timelessness between takes, so even in the case of the “genlock” of the cameras, there will always be an error of time in the seams.

gopro3dgafas copia

SONY DSCThe “Genlock” can be interesting if we make captures of a single row, using optics of 7.5mm or 8mm, in which case we can capture doing only three double shots and therefore we will have to repair only three seams.

If what we have from those photographs is something in motion, then it will be necessary to use this system, but always without forgetting that we will have to watch the seams.

The capture of moving scenes has to be considered an advanced technique, which is learned over time, so it is not advisable to start here.  it is better to start with simpler photos taken in your own studio.

 

 

 

 


4.11) Cromatic aberration and pixel effect.

 

czjlogo

The vintage Carl Zeiss logo is one of the great inventions of Dr. Abbe, and represents an apochromatic pair consisting of two positive negative lenses with different refractive indices, allowing the creation of magnifiers free of chromatic aberrations, which are still made for jewellers today.

The lenses of high positive prescription carry implicitly the chromatic aberration, and in the case of a Samsung Oculus Gear, we are using 30 dioptres that force us to look exactly through the center of the lens, because if we do not a colour fair will appear, so it is important to use VR Gears with interpupillary distance adjustment, something simple to make mechanically.

lsd

If we look at the screen of the mobile through a magnifying glass, logically we will perceive the plot of pixels, something that is solved by extending the focal distance of the optics, as does the VR OWL, a stereoscope designed to see photographs.


NEXT