3
$\begingroup$

I'm hoping I can explain this properly.

I'm looking into rendering stereoscopic 3D fisheye images for planetariums. I've previously used 3dsmax, Vray and Domemaster3D created by Roberto Ziche and Andrew Hazelden. It has some functionality that are perhaps planetrium specific, but I'm wondering if there's any way of doing the same in Blender.

From my testing, when rendering a stereoscopic 180 fisheye in Blender the pole merging occurs at the top and bottom of the image, which when displayed in a planetrium dome, is the front and back of the dome.

Illustrating the position and direction of the pole in a stereoscopic 180 fisheye image.

I'd like to know if it's possible to tilt the poles so that the eye separation and stereo effect are aligned towards the front of the dome and the area the audience is looking at.

Illustrating of the poles tilted in a stereoscopic 180 fisheye image.

The image below is an example the seating in a planetarium. The bottom of the fisheye image is the front of the dome, and the rear is the back of the dome. The pole merging would be tilted to be between the centre and top of the fisheye, which would be between the top and back of the dome.

An example of planetarium seating.

Can anyone advise if this is possible with the current Blender stereoscopic controls? Or does anyone have any suggestions or know of any similar Blender development regarding fulldome stereoscopic 3D.

You can find out more about Domemaster3D here: https://github.com/zicher3d-org/domemaster-stereo-shader

[ additional clarification]

I'm talking about tilting the alignment of the stereo cameras left/right offset rather than the image itself. The concentric circles in my images represent the alignment of the eye separation, the alignment of the offset between the left and right eyes.

In the first example, I'm illustrating how it's currently aligned, which is works if you are looking at the centre of the fisheye. The eye separation is left to right until they reach the poles at the top and bottom, where they pinch together.

But in a planetarium, the audience looks at the bottom area of the image with the centre above them. The left and right eye offsets would not be aligned with the audience's viewing direction.

Instead, what I'm asking is, is it possible to tilt the offset so that the left and right offset would be aligned with the audience's viewing direction?

This tilt is illustrated in the second image, where the concentric circles, which represent the left/right eye separation, are aligned with the audience's viewing direction.

$\endgroup$

1 Answer 1

4
$\begingroup$

I do not know if this is exactly what you want, but let's say you have an image projected onto a sphere, for example like the world background with an HDRI as Environment Texture. Placing a 180° fisheye camera in the world's center, rotated 90° on X and 0° on Y and Z i.e. facing in +Y direction, it looks like this (I've added stripes to the HDRI to make the tilt more visible afterwards):

world default

And with Texture Coordinate and Mapping nodes you can now tilt the background on the X axis by -45° and it looks like this:

world tilted

Is this what you want? I'm not sure because it seemed too obvious for me... but maybe this is a starting point to get you closer to your goal.

From your question and comments I'm a little confused about upper and lower pole and the tilt angles etc. Now if I use a sphere instead of the world background, what I see on the right side with the dome looks like this:

sphere tilted backwards

But this would result in something looking like this through the fisheye lens:

tilted backwards 45°

So if you wanted the lower pole to appear in the upper half of the view, you would have to tilt the sphere or dome 135° like this:

tilted backwards 135

And the result looks like what you have on the left side of your schematics. So I'm not sure are the schematics to the right incorrect or the depiction on the left of what you want to see there?

fisheye view of 135 tilt

Anyway I would assume that you either have to tilt the sphere -45°, +45°, -135° or +135° to get what you want. About the stereoscopy: Do do want the pole on both cameras to be in the exact center of the field of view? Of course you could angle them towards the center, but that's usually not how stereoscopy works, the left and right eye should be angled parallel to each other (as long as you do not want to simulate someone squinting strongly).

$\endgroup$
3
  • $\begingroup$ Thanks for your answer, but I'm afraid not. I'm talking about tilting the alignment of the stereo cameras left/right offset rather than the image itself. The concentric circles in my images represent the alignment of the eye separation, the alignment of the offset between the left and right eyes. In the first example, I'm illustrating how it's currently aligned, which is fine if you are looking at the centre of the fisheye. The eye separation is left to right until they reach the poles at the top and bottom, where they pinch together. $\endgroup$
    – Ian Smith
    Aug 4, 2023 at 7:55
  • $\begingroup$ But in a planetarium, the audience looks at the bottom area of the image with the centre above them. The left and right eye offsets would not be aligned with the audience's viewing direction. Instead, what I'm asking is, is it possible to tilt the offset so that the left and right offset would be aligned with the audience's viewing direction? This tilt is illustrated in the second image, where the concentric circles, which represent the left/right eye separation, are aligned with the audience's viewing direction. I hope this clarifies $\endgroup$
    – Ian Smith
    Aug 4, 2023 at 7:55
  • $\begingroup$ I was a bit mislead by the schematics in your question. The hemisphere on the right is just tilted 45° between the top and bottom picture. So to get one of the poles in the upper half of the image and keep an angle of |45°| you would have to tilt it -45° to get the upper pole there. For the bottom pole to appear up there you would have to tilt it 135° which does not look like it on the right, hence my confusion of upper and lower pole. I'll edit my answer to show what I see there. $\endgroup$ Aug 4, 2023 at 8:59

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .