Sewing and folding limitations? Simulation and Production questions
Is there anyway to stitch directly through a folded 3d pattern without having to calculate and draw the internal lines of complex folds?
For a very basic example, the back pocket of pants.
IRL you fold and press your seam allowances, and stitch through all of the folds, without having to breakout digital calipers and a compass, but in CLO it seems you would have to calculate the exact angle of the secondary fold so you can digital stich from fold to fold and then to the back panel.
Additionally, when I try and do side-seam stitch on a pair of pants, there seems to be a visible gap, as seen here:
And finally, I was trying to do a double-chainstitch on a lapseam, but am seemingly unsuccessful:
I've been playing with the trial for a few days, and it seems absurdly powerful for simulating IRL production garments, but I can't figure out if the features i'm searching for are not-yet-present, or if I'm fundamentally approaching it the wrong way.
Thank you so much!
The gap is coming from the fact that your fabric has added thickness, but it's not being depicted in the current view. Switch to "thick textured surface" (alt+1), you're currently in "textured surface" (alt+2). The options are in the folded fabric icon, top left of the 3D window, and in the menus under Display.
You should change your approach slightly, you want to simulate your patterns without seam allowances. You can have seam allowances on your patterns in the 2D window, so the full pattern is maintained, but seam allowances don't add to fabric physics. E.g. inside a pocket there's absolutely nothing to be gained having that extra mesh folded exactly like the real cloth. For the physical simulation, you need the sewn outline of the pocket and the location of the topstitching, all the other details can be added with textures if they're needed later. You can keep the pocket hem (not essential either), but definitely not double folded. Same goes for felled seams, yes IRL the construction locks the fabrics together in an interesting way, but for simulation you want, the part of the fabric visible on the exterior. If ultimately you need a true to life visual, you can add bulk and puckering to the seams to depict both the thickness and the wear patterns. You can mark all the topstitching with internal lines, or you can use offset topstitching from the pattern outline, including multiple lines.
E.g. IMO this is a reasonable level of detail for simulation purposes, only hems no seam allowances:3
The textured surface sin is all-to-common that I wonder if CLO could do something to make it really hard an un-intuitive to turn it on.
As for using manufacturing techniques like folding seams in CLO, I'd say that you have to first of all define what is your objective. If you want to have a visual 3D representation of your design before manufacturing or maybe create different digital versions with different colors and experiments new designs, then following IRL techniques is not necessary,
If you are planning on testing your patterns for manufacturability, then 3D is not your best tool.2
Thanks for the great responses.0
There are two process flows happening within CLO3D > the technical pack that describes the 2D pattern pieces, cut lines,colorways, trim, grading and detailing with specification and measurement datum's and then the early bird > marketing visualization to describe the garment assembled product as a presentation. It's exactly the same for 3D product design and engineering or architectural interior rendering. For the marketing visualisation aspect of the project, you don't render all the internal components, or unseen edges and faces, that would make the project far to unproductive to simply achieve the final (photoreal) surface visualization images for promotion. A bag of VFX 'tricks' are generally used for expressing textural detail in VFX, and it may also include the implied internal components (eg: glass or clear plastic where line of sight reveals additional surfaces needed to convey the 'whole' products external surfaces). But it is generally limited to as few polygons as reasonable to reach an efficient 3D surface model with textures fit for visualisation purposes). Rendering heavy weight models can be expensive in time and money, so the goal here is to reach a happy medium of what to toss out and what to keep in. Fashion is no different, you do want to show detail where it is not required in the marketing visual, if the construction is not being expressed internally.
Similarly the technical pack, is where the 2D sample maker will require the full detail and annotated directions to make the garment, with all the critical points of measure, sewing lines, and tables with material specifications and tolerances. And that is where you apply that detail typically on the flat form and then on the assembled garment in lower fidelity where clear instruction and assembly or exploded views, simple line drawings of 3D, colorways, etc are used to convey meaning to the maker. Here the 3D photoreal visuals might simply be in support of the 2D technical pack.
It's 'fair' game to use additional tools like texturing and depth of surface to convey surface meaning for any detail under the construction, this means pucker, folded seams, creasing, hidden pleats, ribbon knots, weaving - all might be implied by clever arrangement of external surface texture. And this is exactly how CG product designers and architects use simple aspects of light and shadow within 3D surface models to convey additional meaning. For example in architecture, the chamfer is used to make light travel along a machined or chiseled edge. Similarly in product design, the edge radii is used to convey plastic thickness due to molding case parts, and translucency or ribs (using sub surface scattering in rendering). In fashion you may bake the seam, stitch and edge detail from a library of common (generic) machine fastenings you would commonly use across a garment and from these baked textures apply them to 2D flat pattern pieces on the UV in the UDIM texture atlas layout and wrap these texels (mapped pixels ) across a garment assembly to convey the finite detail down to the woven level if camera closeups are required. That is legitimate use of 3D in the digital domain space of > digital sample making. And it keeps the projects digital assets economical and fit for their intended purposes > the technical pack to describe making & garment assembly/quality > and the marketing visual (photoreal) to offset making a costly real-world sample to convey design/style-way options. So here in early digital marketing the VFX and texturing of made details is used to test market consumption and risk in production - without costly making of real world colorways, fabric alternatives etc. It would be insane to recreate all the unseen detail to simply arrive at a market answer that said - no don't like the product. So it's often best to leverage garment details into quick to apply methods to quickly get from A to B - in order to arrive at that point in the visual response. You don't need to always show all the detail to get to that point. And crucially you want your texture and seam detail library to be highly adaptable and easy to apply across all garments. So there is some pressure on CG artisans to 'rationalize' how to best curate and apply 'unseen' detail so it conveys accurate manufacture and yet is economic to adapt to any garment. One of the best ways is to create tiled textures alongside any garment project you create and save that to a common library to re-use, on the next iteration, at speed, to a photoreal end result. Folding over every seam on every 3D garment you ever create over a lifetime would be an exercise in wasting good time that you might apply to new design/style arrangements instead of pure inconsequential detailing.1
Thank you for such detailed explanations!
@ottoline out of pure curiosity, any idea how far away we are with desktop compute power to make this tiered compute-cost rendering approach unnecessary and include all of the internal material and geometry rendered at a similar fideltiy as offered today with the clo render window?
Thank you again0
You may cheat the detail and still feel good about it. We all do it.
If your computer is not as modern as some of the latest hardware, you can send projects to be cloud rendered at CPU appliance cost - cents per compute hour. That is an easy upgrade that anyone can make - as it means you can simply load the cost onto the final render output for a client. So no upgrade of hardware required.
And if you use texture baking for creasing and pucker maps you can use a bigger particle mesh density on the garment assembly but still get high end detailing. So there are some great ways to speed up simulation without investing in any additional hardware.
Also in simulation settings you can create custom simulation presets that avoid frame-computes that are costly on your cpu/gpu in time. So you can switch off cloth collision to itself for example and get a faster general drape when you are just doing work in progress. So there are ways to speed up the workflow. The best tip is NEVER draft a pattern at small particle pattern cloth distance. As you simply don't need that compute overhead, so create a few new custom presets and adjust how you work and you can speed you workflow up by 300% in many cases - possibly even more. Best hardware upgrade you can make with CLO3D is actually to change how you work and change a few presets. Then - Bingo.1
I can understand wanting to be able to see inside a garment, and having all the construction detail true to life, but where does it end? I've worked on enough garments, that I tend to look inside things every chance I get. But If you do get out the calipers to recreate all the internal folds of that pocket, when are you going to see that? No one sees inside a pocket unless they rip apart the seams, like you've done. Do you want all those frayed threads, too? How about places where the seam allowance has been clipped to aid construction. To some small extend these things do change the exterior of the garment, they cause lumps and pulls, but that's pretty low on the list of things needed to make a garment look life like, and there are easier ways to get that look. I wouldn't expect simulation to get better in this direction, without increased need. Use the computer to decide where to put the pocket, how big it is, how it's ornamented, etc.0
You may actually get any level of texture trim detail for any type of garment detail to any level of complexity, and make it simple to apply to a CLO3D/MD thin mesh garment on both sides. One of the strengths in VFX and vray is to use the right quality textures and mating tiles so they repeat seamlessly. In any of my garments I do this all the time and many are so realistic that you cannot tell them apart from finished product. In fact we do this for marketing just like you might in product design and architectural rendering for buying off the visual, before the final product is made. This includes fabric weaving simulation and complex woven trims eg: lace and computer kit. And I mean very complex weaves not your standard generic and simple fabrics you see in Adobe Substance. Way more complex.
It is possible to do this all at speed also. There is maybe a huge misconception that it's costly - but not so. It's all about the technical digital setup across hardware and software, and then consistency in how you break down the detail, and then recompile the detail synthetically. Above and below you can see the same open weave fabric, true to the real material, and yet I have 2 totally different edge treatments. All of this is virtual - that means the lace trim (with seam and stitching) and elastic are with pucker and seam stitching both use the same fabric repeat and match perfectly to the weave (yarn) diameter so you cannot see how they repeat, along with a random noise crease to add in additional realism. This is drag and drop simplicity done is seconds to swap out a different edge treatment.
And then the ability to simulate random lace at the level of the yarn strand is also possible to any new sketched design. It is simply a matter of breaking the texture detail down into the node graph of repeat and applying to your shader options in vray.
And you are not limited to just trim detail you can also mix it in with elements like tassels and fraying or holes etc.
You may blend in additional creasing frequency over your basic fabric weave to make any fabric look less computer generated and more photoreal, that is simply a function of texture mixing.
If you want shimmer like a silk add in anistropic light behavior, if you want soft (above) add in subtle tissue like creasing.
if you want colorways in weave > add in a master weave ID map. (far right green map)
If you want random fringe detail add it in. All this detail is simple to simulate, repeatable and photo-realistic as you want. Even under close up camera angles. Where many go wrong is by relying on the likes like Adobe substance/ sampler or generic low fidelity service providers to provide this for you - rather than do this yourself taking the time to build the capability inhouse. And if you were to compare what you might get from the freedom to think freely, rather than be stuck in monthly user paywalls - many might jump ship to higher quality and simpler use, that benefits all their digital work. Unfortunately too many service providers limit what you might get in detail level and performance by simply keeping people locked into to toolsets, that limit creativity or homogeneous choices from a small list of basic fabric choices that don't really open out the full range of fabric and textural opportunity.
Now the really outrageous thing is none of this is costly to achieve nor is it slow, all this detail above is drag and drop and tiles seamlessly down to the weave gap. What is missing is the desire to engage at a higher level of expectation with the current crop of texture curation service providers that seem to lack a tight focus on what truly matters when gathering detail and capturing data. And in the 8 years I have been generating these types of garment details and using them in photoreal garments - I have yet to see a commercial service that can match what you can do for - free. That's right - this level of detail can be done for the ridiculous cost of time and knowledge. So the barrier is not .... can you have that detail if it's not in the Adobe curated libraries, or similar, it's about taking the time to look elsewhere, to get what you want when you want. All possible because you are looking at that above.
No level of detail is impossible, organic or structured - it's simply tied to knowledge (not cost) and the desire to have it appear freely in your digital models.0