MeshFusion: from beginnings to future thoughts

July 12, 2016

Before joining Luxology (later The Foundry), I worked for decades as a freelance illustrator and designer. For me, art was, and is, a vehicle for visualizing ideas. I do enjoy—and often obsess over—simply creating a visually satisfying picture. However, there has always been a left/right brain mix in my work.

The right brain side was always a bit of a wild card, though. For example, I dabbled in homemade software, often creating my own digital art tools. One bit of dabbling was GroBoto, designed to be a playful, easy introduction to 3D graphics for kids. That one actually made it out into the world. Eventually my brother Jeff and pal Boris Tsikanovsky (guys who actually understand mathematics and code), joined that project. It grew into something more than a kids’ app—into a kind of semi-algorithmic 3D tool.

Boris and I teamed up with Luxology in 2013 to develop a third-party plug-in. We signed on with The Foundry at the end of 2014.

Toward MeshFusion

As developers we understood the need to move the unique capabilities of GroBoto into an environment suitable for reaching a larger audience. GroBoto was based on quirky, non-standard technologies, placing a barrier between it and the larger world of 3D CGI. We saw MODO as having the right sort of environment, including a community with enthusiastic interest in our efforts, and an application architecture that opened the necessary technical doors.

So we began the development of MeshFusion. The earliest versions required fairly specific, fairly dense meshes to work well. Over the various releases, many enhancements, big and small, have made it a more practical, productive tool. While still based on Catmull-Clark surfaces, MeshFusion is far more accommodating of all meshes of that type. That mesh-handling robustness has been critical for MeshFusion to work effectively with 10.1 procedurals.

Big strides in the tool have also been made with ‘Strips’—the geometry that joins fused surfaces. Strips are a crucial, unique feature of MeshFusion. Improved core tech and local control of strip and corner properties make creation of high-fidelity strip geometry possible.

The evolution of modeling

With the addition of procedural modeling in MODO 10.1, the true potential of MeshFusion is being realized. At a base level, simply using procedurals as Fusion source meshes allows and encourages form exploration. Those meshes can both radically change, and adapt to radically changing, Fusion models. Procedural meshes are also easily optimized for workflow or final output.

Going further, MODO’s architecture—including rigging, scripting and customizable UI—brings these two nondestructive, nonlinear systems together. We are able to craft tools that enrich and simplify modeling tasks and workflows.

B-Splines are a great example. When used as the basis of procedural meshes, they give the designer very direct control of form. Manipulating that form as a spline, and seeing those changes applied dynamically to a complete procedural/Fusion model, is powerful. B-Splines and Catmull-Clark surfaces use the same math—so it’s a perfect fit with MeshFusion.

In the future I see us possibly moving away from Booleans. MeshFusion has always transcended traditional Boolean modeling—allowing use of open surfaces and joining surfaces with smooth transitional strip geometry. At its core, it does not depend on, nor is it limited to, the intersection of volumes.

We see the opportunity to capitalize on that independence. MeshFusion could support fusing of surfaces along arbitrary curves. The entire seam network found in a typical MeshFusion model could itself be mutable, yielding true freedom and flexibility of form. Nice curves may still be found via intersection, or created manually. Either way, the artist would be free to reshape them.

The entire model could be both plastic and structured. A mix that doesn’t exists in any current modeling toolset.

The ‘holy grail’ of 3D modeling

I think ultimately, though, artists and designers should have as much concern with the component details of a mesh as digital painters have with individual pixels … just about zero. It’s vital that artists and designers be able to work directly in the 3D medium free of resolution or mesh structure concerns. The workflow should not require resolving technical issues or a separate, left-brain skill set.

Clearly this is possible: it essentially exists now with digital sculpting. It can be achieved for all forms of modeling. The structure of the mesh will still be crucial. Attempting to overcome the demands of 3D modeling with brute-poly force—like pixels do in 2D—is both impractical and ill-advised. Intelligent mesh topology—created on the fly behind the scenes—is key. The artist/designer must be free to focus on form, to explore and iterate without penalty or impediment.

And like 2D, 3D modeling needs to be responsive and fluid. We believe MeshFusion technology—supported by the right toolset—can get us there. But it is, indeed, a quest.

Leaving a legacy

In the meantime, I’m really appreciative of The Foundry and their user and developer communities who have been so supportive and patient as we’ve grown MeshFusion. As an artist and designer, I’m as eager as anyone to have the tool fully realized and in hand.

What inspires me to keep chasing MeshFusion is the notion of tool building itself. Naturally, I want it for my own use, but it’s more than that. I’ll be quite satisfied with my creative life if these tools help other artists find satisfaction in theirs. They are something that can outlive me in a truly active sense.


Darrel Anderson

Darrel is an artist, designer and coder with no formal education or training in any of those disciplines—group-taught through collaboration and association with many very bright people.

Share this story 

blog comments powered by Disqus