View Full Version : Puzzled on the best way to do this

11-10-2014, 02:22 PM
I've attached a jpg with a logo that has the illusion of a sphere with the letter "P" cut through it.

1. My thinking it to extrude the P and make a hollow sphere with a real thin shell, cut that in half along the Y plane so I have the front half of the sphere and boolean the extruded P from this half sphere.
I'm thinking this way so I can get a reflection that reads like the artwork suggesting a sphere reflecting a horizon (as in the 2D art).

This might read ok looking straight on but not so great from any angled view.

2. I could also just use a flat disk and bump map the P to get the edge that reads like it's cut in and UV map a texture on the disk for the art.

3. Perhaps a slightly curved disk - maybe bump mapped curve? Sort of a marriage of the two concepts

Any ideas on the best way to go about this and end with the closest result would be helpful.

Thanks all.


11-10-2014, 02:31 PM
You could certainly do a boolean cut by cutting the 'P' shape out of the orb. Are you certain it's a sphere and not a squashed ball, like a M&M?

I assume there's some sort of animation you're doing? How does the client actually envision their logo in 3d space?

11-10-2014, 02:37 PM
Thanks for the reply,

The clients I work with are financial advisors so anything we do in 3D tends to blow them away. That being said what ever makes it look nice while animating. For this one I'm thinking transition (dissolve) between multiple dynamic camera angle moves (extreame close ups on various parts) coming to rest on the logo straight on (only because in this case it's going to require that to look right).

I'm leaning toward M&M.


11-10-2014, 02:50 PM
Yeah, I'd say stay away from the full sphere. The M&M shape would be good because that would still sorta work with a 3/4 view. Spinning the logo around like a ball seems really unnecessary for what you're client is trying to portray. It might even be more about dynamic camera layout and subtle camera drifts versus a lot of wide ranging motion.

11-10-2014, 02:54 PM
Thanks for the input. Guess I was just making sure my approach was solid before investing the time. Hate to find out there was something I might be missing before I put pixel to screen.


11-10-2014, 05:03 PM
I think you could whang this out pretty easily. Just model the basic M&M shape for the glass bead or even a meniscus that has the reflective material qualities you like, then simply use alpha masks to cut the shapes in Reflective, Transp, & Specular attribute layers.


11-10-2014, 11:08 PM
I agree the cleanest way to do this would be images and masks. You could do a displacement even based on the image. It would just be a matter of getting the right masks built up in something like Photoshop.

11-11-2014, 06:50 AM
Thanks Ma3rK. Good input.


- - - Updated - - -

Thanks for the response.


11-11-2014, 09:50 AM
the boolean tools from 3rd powers might work?

I just ran a fast test with dpont boolean surface node, but didnīt work out so well with a complex cut object as a text letter wich by itself has a hole "p" not sure if you can work around that and correct it somehow, it gets screwed up with the inside hole for the p letter..

otherwise clip maps might work better.

11-11-2014, 11:39 AM
Have U tried Using 2 Boo cutters on 2 spheres
at o,o,o, then reassemble them back into one layer.

(Yep! 3rd POWERS will do this & More!)

11-11-2014, 01:00 PM
Yeah there are options for modeling. But for something like this you can get much better and cleaner results with mapping. If I was going to model it I would not use high density operations like Boolean. High Density because that is what it would require for the detail. Rather, I'd use subdivision surface which would be controllable at render time. And on that note the image solution also requires the base object to be subD. And you control the displacement density at render time. Of course fine details would only have to be bump or normal.

11-11-2014, 03:31 PM
it works with dp boolean surface nodes, in layout, as I posted previously..I got some artifacts, but bryphi77 solved it and posted the tip, simply triple the geometry and it should work,hereīs the tutorial..and many many thanks to bryph77.



11-11-2014, 05:06 PM
Nice job!

I still think Boolean is overkill and not even the right tool for this. You would still have to solve beveled edges. It is not a question of can you, it is should you. And for a high quality logo animation like in the original post, you'll get much better, cleaner and controllable results with images.

If you really want to make this geometry, use subpatches.

11-11-2014, 06:28 PM
If you really want to make this geometry, use subpatches.

Just curious how you envision that process.

1) Lay it out flat and then heat shrink to a sphere (M&M)?
I tried, quickly, flat and then spherize (I don't have heat shrink), and, as expected the shapes get totally distorted. (I didn't go total hemisphere of course :))
I guess a lot has to do with the severity of the final curvature. Small curve = less distortion, but it still won't be perfect from dead center?

2) stencil onto an M&M and then go around the perimeter fixing ngons? (ouch)

3) put logo image in background and in front view add edges on existing M&M half shape?

4) ?

11-11-2014, 10:45 PM
Yeah it is a tedious process no matter which way you go. I'd probably map an image onto a sphere with UVs and use that as a reference to draw out the initial outlines of the poyflow on a fairly dense mesh.(Add Edges) Once I had that in place I'd go in and clean it up as quads and all of that. Probably just fix it all by hand and maybe use some kind or shrink wrap process if needed once all was in place.

11-12-2014, 08:02 PM
I kind of read the logo as a flat disc with extruded features covered with a glass lens or encased in glass. That doesn't explain the gradient in the center, though. :)

No matter what way you go, you'll have to pay attention to the environment you put it in since much of the shiny glass effect comes from a faked reflection.