Meshes vs surfaces and curves for solid modeling

Support Modeling
#1

Believe it or not I have been doing some of the dreaded “RTFMs” and tutorials etc. I am embarking on a project of having an animated shipyard of the 17-1800’s. All the objects will be solids, with many ropes, pulleys and chains. The materials will be rope and wood and some terrain features. The animation will be basic with ropes moving solid objects. I’ll worry about any ‘characters’ later. These shots are from one of my reference works.

001.jpg2300×1582 499 KB

I would like to build some experience and am wondering, for this particular environment, it would be better to concentrate on meshes or non-mesh modeling. I have a CAD and some C4D experience, but am no means proficient.

Any advice would be appreciated.
Thanx,
Ed

#2

I think that overall workflow would be much faster in blender if you ask me.

#3

There are terms used in the opening post that might mix things up.

Solid modeling is a modeling paradigm of its own (Solidworks, brl-cad) that is used to describe objects with volume, or solids in other words. Commonly used for technical purposes and solid nature of the objects allows for datapoints inside the objects for stress simulation etc.

Polygonal modeling deals with meshes, originally developed and still mostly used for visual purposes. It’s approximation of curves and surfaces, meaning that it’s a surface type modeling paradigm and we are approximating them instead of describing them with absolute precision. Meshes are approximations and refinement algorithms (like Catmull-Clark with subdivision surfaces) are also approximating, not interpolating ones.

Surfaces might mean polygonal surfaces, or it might be in the context of NURBS, which is also a surface type modeling paradigm. Blender has them but it’s mainly a polygonal modeling application. NURBS tools are old and not that great.

Curves like bezier and others can also be used as helpers for polygonal modeling or animation, and they can be used to define limited kinds of surfaces. Tube-like objects are very easy with those. Could also use them as guides for another curve shape to create extrusion-like structures.

As for the question, if the project is for visual purposes, could model and animate in Blender. If you need precision parts or accurate simulations of a real world, use something else. If you use Blender, you could use polygonal modeling tools and curves. Could also use NURBS surfaces on some parts if it’s easier for you, but that would be just to help creating meshes (eventually converting to a mesh).

#4

JA12
Thanks for the reply
Sorry for the confusion - there are really too many terms floating around. I want to use Blender. CAD is poor for animation and I have been priced out of upgrades for C4D. You are correct I really don’t mean solids. The work is just for visualization. I do not need data points within objects, but it will be necessary to define the objects accurately.( for example frames made up of timbers that are of a given thickness so as to relate accurately with other timbers which are also a specific dimension) I’ve worked with the Bezier curves and find the handles very easy to manipulate for odd shaped objects. ( later I will add a ship or two using nurbs and lofting)
Ed

#5

I am embarking on a project of having an animated shipyard of the 17-1800’s

but it will be necessary to define the objects accurately.( for example frames made up of timbers that are of a given thickness so as to relate accurately with other timbers which are also a specific dimension)
How accurate would you need something from an 18th/19th century shipyard ? It’s not something you need with the accuracy of a CAD application so well within what blender can do

#6

Yea, you can be pretty loose with the accuracy since it’s for visual purposes. Relative accuracy is enough. Relative to distance to the camera, other objects, features on objects themselves, detail level, scale, and even render resolution.

Visually it’s easy to see something is right or wrong when there is something to compare to. That’s why it’s often necessary to include scale reference in an image so that the viewer knows how big things are without necessarily even thinking about it.

For example, you could model a plank with a hole in it. You could say that the hole is 5cm in diameter, even though it’s actually 42mm, or even 30mm when measured, and no one would know any better if they don’t have a reference point. But if you put another hole next to it and that actually is 5cm while the other is 4.2 cm, the viewer might spot that immediately if the camera is close enough. One would be able to tell they’re not the same size, but wouldn’t know the actual dimension.

Same when the distance to the camera increases. Objects and detail gets smaller the further away they are from the camera, so it’s harder to see and compare. Have them far enough and they disappear from the render altogether because there isn’t enough pixels to have them in. They’ll be smudges on the image before they disappear though, so all fine detail you might have used hours to model or texture might be in that smudge.

One example of that. Did a menger sponge a while back. The number of different size holes is also the level of the menger sponge. Scaled down version for the forum nearly lost one level because they were so small /uploads/default/original/4X/5/6/3/563c20723b5444d3da03d797e930d3c01c1a9f51.jpgd=1396009299
The original had four times more pixels which also was on the limit because the smallest holes weren’t looking particularly square http://www.pasteall.org/pic/show.php?id=68951
That also meant that the next level would’ve needed ridiculously more geometry and to top that, it would’ve needed much larger render to show all the levels while keeping the whole thing within the frame.