Clone me a Modo (procedurally, that is)

If it wasn’t bad enough that we have to update our software every few months or every year, tiered roll-outs of major versions have now become all the latest trend. Aside from the devastating psychological effects – you always feel like you paid for something that you are not getting and have to continue to pay to actually get it – it leaves you with the issue of having to adapt your workflows over and over again. Sadly, Modo has adopted this scheme all too well and so it is only now, about half a year after the initial version of the 10.x cycle has been released, that we’re talking about the “real” thing with Modo 10.2. So what’s new?

First and foremost I want to mention of course the procedural cloning and replication tools. It seems so silly that something fundamental like a radial array wasn’t part of the original feature set in 10.1, which ridiculed the whole idea of procedural modelling on some level. Tweaking the number of spokes on a wheel really is a way more obvious use case to keep things parametric than applying funky beveling to a flat plane based on random selections. If you get my meaning: A lot of what was in there previously had limited value since it still required you to use manual operations way too often. Things are still far from perfect, but at least it’s now possible to clone something, select and delete a bunch of polygons and combine them as one surface with some bevel and bridge operations.

Whether or not this will actually ever become practical to use is another question. One of the things that really rub me the wrong way is the complexity of the procedural networks in Modo. There’s way to many extra nodes required even for what should be simple operations. The usual dance is something like that: You add selection operators to pick out your polygons and apply your operations, but then need to use a Merge Mesh to be able to create a consecutive selection for your next operations. This goes on and on and that way your simple extrusion-based cogwheel from a cylinder ends up looking like you constructed a whole gearbox, at least as far as the node network goes. Assemblies and the new Assembly Aliases, the latter of which can be locked to not expose internal channels and data to the uninitiated, if you so desire, but still, a lot of this just seems wrong to me. You shouldn’t need to obfuscate your internal logic just to allow other people to use it without messing it up.

The funny thing is, the new Schematic-less MeshFusion workflow nicely illustrates that they can make things user-friendly if only they want to. It’s not perfect, either, but, to insert a cheap pun here, it shows that even something like Mesh(Con)Fusion can be tamed in a manner that it’s actually usable without frying your brain over order of operations in a node flow. If there’s one thing, it is that kind of “artist-friendly”, simplistic stuff that would actually win over people from other programs.

The same could of course be said for everything to do with spline tools. They can go all crazy over their new auto-tool-enable and minor things such as closing splines with a right-click, but the truth simply is that the curve/ spline tools in Modo are just completely crooked and wrong from the ground up. From the way the handles are drawn to how you deal with tangents to that weird and unnecessary distinction between different line and curve types without being able to convert between them easily. I’ve always been critical with Cinema 4D, but even before the major overhaul in last year’s R17, the spline tools there were exponentially superior to what Modo has on offer. In that department Maxon‘s implementation beats The Foundry‘s stuff hands down and it only shows how bad, bad, bad Modo‘s spline tools actually are.

Now for the big one: The auto-retopology. As you may remember, I’m not doing any sculpting nor that much organic modeling, so naturally, I have limited use for this stuff. I gave it a quick whirl with some downloaded high polycount meshes, but how it compares to other such tools I cannot say. What struck me as odd is the lack of interactivity. Controlling the polygon size and flow via pre-saved weight maps (and having to call those up in the actual tool interface via pop-ups) certainly works, but in an age where other programs have interactive tools for scaling, orientation etc. this still feels clunky and inferior. On the bright side, Modo has all the prerequisites in place now and if they add it to the procedural pipeline one day, it could even see some cool uses as a shrink-wrap sort of thing to optimize and fix badly hacked together underlying geometry.

There are some other additions like deformer caching to improve interactivity while working, mesh-based lights and a few more “smart” auto-toggle features to facilitate e.g. component selection based on where you click, but overall the impression lingers, that Modo has turned kind of messy and it’s not really clear where they are headed. Many new workflows and tools are not carried through consequently and feel plugged-on and half-assed. The most apparent shortcoming, however, remain the various performance issues combined with an inconsistent user experience. This completely ruins the fun of working with the procedural stuff for instance. It’s always like you have to wait for the program to reorganize its internal scene graph, update some list or load the data into its viewport. They really need to take a deep breath for version 11 and figure these things out or else Modo ends up being the next Lightwave – lots of legacy issues to carry around that prevent it from ever making as big a splash as it possibly could.

%d bloggers like this: