Bendy Straw topology

yankeedoodle

The “bendy straw” topology is probably the most basic method for connecting a character’s limb to its body. But apart from non-sub D low poly characters, or beginner modeling tutorials, I rarely see it used.

I understand that it could be an issue if you want to accentuate the shape of the shoulder or something like that, but is it really problematic for a “smooth” cartoon character? Maybe it creates limitations in the way the mesh can deform?

When looking at Pixar characters, they also seem to be using a more robust topology, even if they have relatively simplified / cartoony bodies:

https://imgur.com/a/jEF4CPw

But then again, I just saw a Blender Talk ("Level up by creating a short film" by Nurbek Nurlanbekuly), where they used the bendy straw topology:

Does anyone have more insight on the subject? Thanks!

Eric Chadwick

We have some more examples here 
http://wiki.polycount.com/wiki/ShoulderTopology

This is probably the single most complex joint for deformations, besides the groin. The trouble comes when simulating what human bones and muscles do as a character raises their arms above their heads. There are a lot of additional movements to "sell" the motion, including the clavicle and the shoulder blade.

The bendy straw topology just doesn't look that great in extreme deformations. But it really depends on the character style and what it needs.

Best way to learn what works is to try it and see!

Alex_J

the daz base model might be the opposite end of the spectrum. it is especially designed for extreme morphs and the topology to support that, including things like the lat raising when arm is lifted over head.

I just start with simplest and then test your most extreme use cases. since the daz base model is freely available (and many, many others) it is usually possible to test a lot of different setups without a big time expense. e.g. grab base models with different shoulder topology styles and rig them all to quick rig, move it around and see what the difference is

just be sure to test your most extreme use cases but also not to let unrealistic use cases misguide you into trying to solve problems that you wont actually face (e.g. doing splits for a character thats only ever going to run)

yankeedoodle

Thanks for the advice!

I'll study these examples and also the daz model topology, and see what works for me.