Mapping a spherical animation onto a 3D physical Globe

[emotion-friend]

I'm going to be doing some content for a real life physical LED Sphere . The question I have is..... Can I create content in c4D (like chocolate dripping down a spherical blob of ice cream) and then export/render it as an equirectangular/dome animation which can them be played on the LED sphere.

It sounds easy at first but the challenge is that there can't be a specific point of view (camera angle) as the animation happens on all sides of the globe.

[Dr. Sassi]

Hi emotion-friend,

I need more information, as this question triggers many questions.

Let me go quickly and simplified through the problem:

Your description does not tell if it is a VR Sphere, i.e., like a LED Volume that a person can enter, or a sphere that resembles in a smaller size something that the audience can walk around. In other words, is the LED pointing inwards or outwards?

Inwards, the audience can enter.
This would be the 360º camera, which can produce an equirectangular. It would sit in the exact center of the Sphere. OR sit where the eyes of the audience would be, and another position creates some kind of distortion.
However, I believe you want to have not a glass sphere.

Outwards, the audience can walk around it.

This is like a model to VR, or much simpler baking the surface. I have no idea if that would resample your question.
The drop down the Sphere can be really seen without distortion from any point. When it is on the Sphere, it has little distortion; as there is on a sphere, only one point can look perpendicular.
Typically you can build a camera rig in MoGraph, with a target in the center. Then make that editable, which creates all the cameras you need. Those need to go then in Takes to make the process simpler. Then stitched, but also here, anything that is not flat on the surface will create parallax, which will show on the seams.
Not suggested, but for illustration purposes
CV4_2023_drs_23_MGcr_01.c4d

I suggest using a high-res Sphere with a 100% reflective material (Standard Render) and baking an animation, leaving the Use Camera Vector to influence OFF.
This would be the simplest way
CV4_2023_drs_23_TXsb_01.c4d

All the best

[adult-garden]

Good morning Dr. Sassi, I am following your advice, as I also have to transfer the animation of a sphere that rotates in a static position in C4D to a real physical sphere covered by LED screens.

I am having trouble because I cannot find a software that can stitch all the videos created by the different shots onto the C4D sphere. Do you have any advice for me?
I can't figure out if there is a way to do it with After Effects.

[Dr. Sassi]

Hi adult-garden,

As I wrote above, I need more information. Is it a complete sphere, or more a Hemisphere? What does the scene look like, and perhaps a little sketch?

Here is an experimental suggestion. This is a little bit twisted, but it is the closest to representing what is inside the Sphere on a 2D spherical surface while not randomly stitching views from the outside.
The simple idea is: (Invented terms are in Italic)
We need to find a way as if the content inside the sphere projects from the middle point, the closest surface to the Sphere, pixel by pixel.
We need to apply a trick I haven't seen anywhere else to do that. We place a 360ºx180º camera in the center of the Sphere. All objects are set to be visible not in the Primary Ray but in reflections (Secondary Ray), which means the camera produces an Inverse point of view to create an "Omni-Perspectically" equirectangular image. This can reproduce an image from the content inside while not creating a "Plural Perspective", which means there is an equal representation of the inside as far as a surface can represent a spherical volume as a surface image. many cameras from the outside of the sphere will result more in a look of what we were used from Cubism.

Check the scene file, render it, and then have a look at where the Torus and the Type are placed. What is in front of what? This might be, for a minute, mind-twisting, but then it works.

CV4_2025_drs_24_VRcb_01.c4d

If you take many cameras from outside of the Sphere, there is no matching perspective content from one image to the next image or another camera, as they are at a distance from each other. Even if those get merged into something, they have areas where a certain point of view creates an illusion of being able to look into the Sphere, but anything around that spot will be heavily distorted. There is no stitching app that could fix that. All images I have seen from the Las Vegas Sphere are either relatively flat or have graphical elements without perspective details that would allow those to create sweet spots. The only thing that can come close is a Sfumato-like (1-dimensional Perspective)).

This means the perpendicular view to the surface of the Sphere allows for the closest representation of the inside. At the same time, the tangential is naturally less visible but also farther away from the idea. A good compromise when all possible points of view are relevant to serve.

So, thanks for the question; the time invested today to explore it allowed me to develop this new approach. Love it.

Cheers

[adult-garden]

Thanks for your help. I am attaching [https://we.tl/t-m9RKECDeIt] the animation that I should reproduce on the sphere, and an image of the sphere on which the animation will be projected.
Thank you very much.

[Dr. Sassi]

Hi Adult-garden,

I have not all textures here. Is that object a "glass ball" with just the iris textured?

All the best