Jump to content

How would you go about modeling this handbag?


Go to solution Solved by Cerbera,

Recommended Posts

Hi everyone,

 

I had a client ask if I would model a handbag similar to this one here. It's basically a very slouchy paper bag style clutch.

 

I'm trying to figure out the best workflow to create something like this. It's slouchy and organic and I'm hung up over how to start and would love to know how you guys would tackle this project.

 

On one end, I think I could do this in Marvelous Designer - the one downside is that it's not the easiest to modify the exported model or work in two programs in general.

 

On the other hand, I think I could model this bag when it is "stiff" and then maybe apply some sort of slouchiness through soft body simulation or cloth simulation and then sculpt further folds.

 

Would love to hear your thoughts! Thanks!

Haruko

Link to comment

I'll pop the pic here so people can see it straight away...

 

image.png.54752db616058fff46b0c0c4eba374e1.png

 

I think your initial thoughts are mostly good ones ! 🙂

 

MD is obviously going to win in terms of realism of folding and collision-free simulation, but as you say it would be nice not to resort to that if you possibly could.

 

You didn't say if this is needed for an animation, so I will presume you need a still shot primarily...

 

So in Cinema, the primary difficulty with simulation on this sort of thing is what is happening with thickness, and not only that, but rolled thickness, which makes stuff much more challenging in the collision avoidance department.

 

So I think would begin a series of tests, working up from a very low poly SDS model to see how both cloth and soft bodies react.

My tests would be initially trying to see if either of those techniques produces reasonable results on just an empty bag (no roll top yet), then one with thickness and so on... next I make a simple roll top mesh, again, without then with thickness, and test what happened to that with cloth / softbody. i suspect that is the bit that may fail !

 

Then I would try model the roll top in directly possibly with the thought in mind that I would apply a simulation only to the lower section, and then manually join that to the unsimulated roll section later, a plan that should be feasible if you take care to match the segmentation in the models.

 

The last challenge is the seam section, and how you will handle that. I do have some ideas here, but they kinda depend on finding out which of the above methods produces the sort of results you need...

 

CBR

Link to comment
  • Solution

This is the sort of topology I think you could reasonably start with - I made this from 2 helix splines in a loft with Cloth SDS adding some thickness, then made that editable fairly early on, and tweaked the bag into a more organic shape with tools like Iron and soft selection. This sort of thing, being relatively even in poly distribution (albeit slightly denser at the top in the spiral where it needs to be) should provide a good basis for a) simulation tests and b) the actual bag initial shape if you discover that sims are the way to go here...

 

image.png.0562d82102aa71460606379fc43d3c27.png

 

I would conduct those sim tests first at this very low poly level, but then again with more and more applied levels of subdivision (so they get included in the sim) until I found the one that a) responds best to sims, and b) gives the best sort of wrinkling on collapse...

 

Note: you will need to apply at least 2 levels of smooth subdivision to get the mesh dense enough to be able to form wrinkles, but the lower levels of subdiv may still be useful to get the initial deformation into a more relaxed organic shape.

 

In the few quick tests I had time for I found that softbody did better than cloth, but cloth wrinkled more convincingly if we only get the whole thing not to collapse ! In the end I went for a just a few frames of cloth sim to get the initial relax, and then binned off the tags, and went softbody after that with some success, but it was becoming clear that this would need  A LOT of experimentation with sim and force settings to get right.

 

Another approach would be to just try and model it without any simulation, which is where the old magnet tool and smoothing deformer trick may come in helpful ! This is expounded upon in great detail by Chris Schmidt here...

 

 

I'm sure some of the techniques he goes over may help you here..

 

CBR

Link to comment

Experience showed me that simulating such models is often a bit over the top.

A rough baseshape and sculpting will get you way quicker and more art-directable in the right direction.

Link to comment
×
×
  • Create New...

Copyright Core 4D © 2023 Powered by Invision Community