Strategies for modeling ship rigging

[tkhoatson]

I would like to add some rope rigging to my model, like what is seen in old sailing ships. Since “rigging” involves lots of different things, here’s an image of the specific type of thing I’m looking for:

Technically, these are known as ‘shrouds’ or informally as ‘ratlines’. I don’t necessarily need all the stuff in the lower half of the image, just the netting in the upper half.

I’m interested in efficient strategies for modeling this rigging. (Here efficiency is not only concerned with things like minimal edges/faces and file size, but also the time it takes to model it.)

I have thought of a few ideas, but my experience on this forum has shown me that there are sages who are able to come up with creative solutions I never would have imagined. Here’s what I have so far:

1. Use long, narrow cylinders for the rope, and move them into position. This would also require crossing them, i.e. having horizontal and vertical cylinders. Drawbacks: moving the ends of the cylinders into place could be tricky, and the end result would probably not look quite right (e.g. no sag in the ropes, no knots, etc.).
2. Create a short cylinder, ending in a knot of some sort, and duplicating this up the length of the vertical rope. Then use slightly narrower horizontal ropes between the knots on the vertical ropes, which can be drawn using follow-me with circular faces and arcs which approximate the sag. Drawbacks: Lots of fussing around with small components with varying lengths and even more of the movement issues described above.
3. Find a clear flat image of a shroud, in color, and edit in Photoshop to make the spaces between the ropes transparent. Then import this image into SU and paint it on a triangle. Drawbacks: finding an appropriate image online, having the triangle not quite match the contours of the ship’s side and having the result not look good from every angle.

Any better ideas?

Note: I am using SU2015, and I have very few extensions. However, I would be open to adding a free extension which works on this version of SU, or even SU2017.

[DaveR]

It depends on how closely you’ll end up needing to look at it. I might use cylinders as you suggest.Keep the number of sides to 6 or 8 for the starting circle. I wouldn’t bother with the knots, hitches, or bends.

@tkhoatson

don’t necessarily need all the stuff in the lower half of the image

Too bad. The deadeyes are fun to model

[ateliernab]

I mean for the ropes, I would probably use low poly cylindres (6-8 sided with smoothing could be fine) if I have to look at them from the distance - if the geometry of the rigging is more important than the geometry of the rope itself.

In a few days, I would probably try with this extension about to be released

A sneak peek of my upcoming extension, Orienter Express

Done! Two new tools are joining: one is to center on edge, orient and z-scale and the other is to center on edge, orient and xyz-scale. I have to think about the name of these tools [OE_11_5]

trace simple lines, create a “low poly straight rope component” and use the extension if I understand correctly how it works. probably not compatible with 2017 though.

or manually : you need one “vertical” assembly and one “horizontal” one. if you make them components, you can then copy, rotate and scale them up so they match the rigging.

here is a crappy mockup of the manual one :

in your case, after the movement / rotation you would need to readjust and rotate each column. and once it’s done on the ground, you group it all, make it vertical, and angle it as it should.

Oh and your horizontal elements would probably need fake knots at the ends. good thing is, they would hide any position mistake

with a low poly rope (not a twisty one) it’s actually not that difficult to get an ok result, you just need to be precise when making the single elements, then it’s a matter of copying / rotating / scaling

edit : I know what I’m drawing this evening… now the rigging has piqued my curiosity. tsstsstss

[tkhoatson]

Your deadeyes look great, Dave. But it’s the part about integrating them with the ropes that trouble me even more than drawing the deadeyes!

[mihai.s]

[tkhoatson]

Nab, thanks for your suggestions…

In a few days, I would probably try with this extension about to be released

Reading the description of this coming extension, I can’t even understand what it offers, much less imagine how it might help me. I’ve had this problem with lots of extensions…

(I’m a software developer myself, so I am constantly running into software tools that offer to do something, but the developer simply didn’t know how to describe it in a way that the target user would be able to see its applicability or usefulness.)

or manually : you need one “vertical” assembly and one “horizontal” one. if you make them components, you can then copy, rotate and scale them up so they match the rigging.

I have used scale occasionally, but I have learned to avoid it in most cases. Once you paint a texture on a component, the texture looks totally skewed on any copies of that component which have been scaled. This limits the usefulness of painting textures on any faces of a component, which is something I do quite often.

in your case, after the movement / rotation you would need to readjust and rotate each column. and once it’s done on the ground, you group it all, make it vertical, and angle it as it should.

I’m not sure what you mean by “readjust and rotate each column”, especially if it’s still on the ground.

But your image is tempting me to simply draw a profile of the thing - “on the ground”, as you suggest - and then pull it up half an inch. The ropes would have a square cross-section, but if I hid the edges, it might not matter…

The drawback of that approach is that it still would not match the contour of the side of the ship, but it’s still tempting…

I know what I’m drawing this evening… now the rigging has piqued my curiosity.

I’ll be very interested to see what you come up with!

[tkhoatson]

Mihai, that’s cool! I would want the ropes to come together at the top, but you’re giving me some good ideas…!

[mihai.s]

I gave you an idea to start modeling, then adjust as you need, but to start, start with a minimum of edges.

[endlessfix]

An almost completely native method. I used Pipe Along Path to quicken the extrusion process, but this could also be easily done with native follow me.

[ateliernab]

@endlessfix yeah I had that in mind. took another route in the end. but ■■■■, it was quicker and looks really good ! the pipe method was faster, and granted, you don’t have knots, but that should knot be too hard to add on top !
your technique, add my bottom pulley parts, and voilà

---

Well, I did a native one. I used the solid tools at some point, but it could have been done with intersect with model

---

ok, first, I modelled one vertical element. And I realise how little english vocabulary I have when it comes to old rigging.

I made one round thing, drilled holes, added the rope around, did a bit of symmetry… the zig zag ropes are a simple follow me, straight in the middle of the holes.

The whole bottom part is a component. the top part (basically two cylindres) are simple groups. I composed the whole vertical element, copied it and rotated it so it matches the image.

At this point, I modelled a low poly sphere for the knots, and placed it inside the component. I ended up placing them outside of the component so I could adjust them independently.

Then, I drew one long squigly line, follow me, and hop, the horizontal ropes. I placed the bottom one, copied it, scaled it, and bam.

@tkhoatson

I have used scale occasionally, but I have learned to avoid it in most cases. Once you paint a texture on a component, the texture looks totally skewed on any copies of that component which have been scaled. This limits the usefulness of painting textures on any faces of a component, which is something I do quite often.

yeah but you don’t have to keep components as components. you can explode them and regroup them. Sure, they won’t be clones anymore, but if you whole work is done, no need to worry.
In my file, the horizontal ropes are scaled and as you said, texture would be distorted.

But exploding the ropes and regrouping them would solve the distorsion issue.

you can check the file to see how it works, I exported it in SU15
(45min of tweaking around, half of it for the pulley system. once you get the idea, it goes quicker. not as quick as the pipe along stuff though.)

rigging.skp (1,2 Mo)

[tkhoatson]

Riley, very nice. You got a pretty good looking result pretty quickly, so it meets my efficiency goals!

[tkhoatson]

Nab, you went the extra mile and integrated the ropes with the round blocks! Really nice. It’s amazing that you and Riley were able to produce a pretty complete solution much faster than I could. (I’m not even halfway there… )

[ateliernab]

experience. this isn’t the weirdest stuff we’ve had to model
plus, it’s not our project. it’s liberating to model solutions for other.

I actually spent way more time on the bottom part because it was fun. the top is basically straight tubes, curved tubes, and low poly balls.
the knots were eyeballed. we’re talking about a ladder of lines, not a very precise milled out of steel stuff

[francisquitof]

If you don’t have vertex tools to model as Mihai, you can deform it with a plugin called sketchy FFD, it’s free and easy to use as well.

[tkhoatson]

OK, here’s what I ended up with:

I drew the whole thing flat, then used Push/Pull to give it depth equal to the width of a rope (making each rope square in cross-section). Then I hid the edges, so it would look more round.

At the same time, I was intrigued by Nab’s suggested method, because I thought it would better allow me to adjust the netting to the contour of the ship’s side. Then I realized how much work it would actually be to get everything positioned and sized correctly, with every stretch of rope a different length and orientation. But I pressed on, just to get an idea for it:

I kept Nab’s block and tackle components on the bottom, not wanting to recreate that part. And I didn’t bother with having the ropes come together at the top, forming a triangular shroud. When I tried using Scale for the horizontal ropes, as Nab suggested, I found that things didn’t line up right in the middle of the group. So I kept the model rectangular, since I mainly wanted to see how this model would compare in size with my own shroud effort. Here’s what I found:

My shroud method:
Edges - 34,831
Faces - 11,604
Materials - 1
File size - 7.5MB

Nab’s shroud method:
Edges - 352,020
Faces - 162,356
Materials - 2
File - 1MB

I found it interesting that Nab’s method required 10 times more edges and nearly 15 times more faces; yet my method resulted in a model that took more than 7 times as much space on disk! (My model is available on request, but I didn’t want to post a link here if no one was interested in having it.)

I really liked Mihai’s method, but I thought it might also have more edges/faces than mine, and be as difficult to arrange in my triangular shape as Nab’s method. Both methods are relatively simple in concept, but get more cumbersome at scale when the vertical rope lines need to be converging and the horizontal rope lines need to be getting progressively shorter.

The method suggested by endlessfix was intriguing as well. Using the rotate tool seemed like a good option, but I found my rope lines weren’t at consistent angles. It was the spacing along the ship’s side which was consistent.

In the end, I chose a method that was unlike any of the 3 I had thought of in the beginning. But all the suggestions got my creativity going more than it would have on my own.

Now I find myself wishing that I knew how to create an extension in Ruby to make this easier. (Maybe someday, after I retire…) I would make something similar to 1001bit tools’ Linear and Polar arrays: Given 2 arcs of any length and a selected group/component and a given number of repetitions ‘r’, divide each arc into r segments, and orient the ends of the r objects between the 2 segments, using Scale when necessary to make the ends of the object touch each arc. Kinda like this:

Thanks for all the good suggestions! I learn a lot from your experience.

[ateliernab]

@tkhoatson

I found it interesting that Nab’s method required 10 times more edges and nearly 15 times more faces; yet my method resulted in a model that took more than 7 times as much space on disk!

hell yeah baby, the power of components !
to make it simple, when you draw a component, let’s say 1000 elements and 100kb, and you multiply it 30 times, your file will contain 30 000 elements but not 3Mb

because when you place the component, all sketchup has to remember really is “where is the component located, how is it oriented, how is it scaled, how is it painted” - then place a copy on each of these spots.

that’s how you end with a file with more lines and faces - obviously I used circular ropes and elements - but not as heavy, SU doesn’t need to remember all the things at once.

ex : this file is 3747 components (with sub comp and sub groups). no materials, 2,2Mb
if I explode it all, 36,7Mb
it’s still the same, 1M lines, 0,3M faces, 2 materials (neutral and a transparent grey).

the power of components

@tkhoatson

(making each rope square in cross-section). Then I hid the edges, so it would look more round.

I did a spring for a desk lamp a few years ago, it’s square too. and small. and not the center of focus in the model. Videogames designers have done the same over the years, they’re onto something.

[Box]

But be aware that too many components can create a small file that will still bring your computer to its knees! The graphic card still has to render all the edges and faces.

[ateliernab]

@Box

But be aware that too many components can create a small file that will still bring your computer to its knees! The graphic card still has to render all the edges and faces.

yup. that’s why I’ve been assembling one building, group it / hide it, then the next…

and in the end, I’m using a set of tags, one for the roofs, one for the inside walls, one for the inside slabs, one for the outer layer… if I’m planting trees or drawing a road, I don’t need to have the kitchen modelled inside the 6th floor unit. I just need the outside skin (I’ll probably just draw big boxes on a “proxy” tag)

the power of Tags

looking on this forum, you’ll find similar cases where people complain about their file, some are huge, a 500mb file filled with textures, some are surprisingly light, not even 10mb.
but using my file as an example, I could make 18 buildings, 6M lines, 1,8M faces and just reach the 10mb mark. yet my M1 GPU would start glowing red and heat up the room.

[Box]

I did a very rough version as a test for my first use of @alsomar Orienter Express. I didn’t bother with the knots but the vertex option would do them in a couple of clicks

Edit
Throw in a little truebend.

[tkhoatson]

I would make something similar to…

Turns out, there already is an extension very much like what I had in mind! I just found mc-AlignEnds on sketchUcation…!