"Well that's a "possible" workflow, but many times you would probably be better off just keeping that model in the poly modeling program throughout and never bringing it into MoI at all - MoI's strengths are more in the area of mechanical object design and not so much in organic creature modeling."

My intent is

Cinema4d > 3d-sketching / concept design
MOI > Construction of the final model

The models I want to build in MOI are not organic, but perfect for Nurbs.

Nevertheless I use the polymodeller Cinema4d for 3d concept sketching and developing the design as polys offer me lots of flexibility and - most important - nondestructive modelling.

MOI comes into play to of the final shape to achieve precision an smooth curves.

Of course I could use renderings of the concept to rebuild the object designed in Cinema, but importing the *obj seemed to be a faster way to rebuild the shapes.

Thank you for the tips.

The overall size of a model imported via script is about 30 times smaller than if I open the same model as Obj-3d, or dxf file directly.

Hi Tom,

> The overall size of a model imported via script is about 30 times smaller than if I open the
> same model as Obj-3d, or dxf file directly.

Ok, by size I think you mean bounding box unit size, sorry I thought you were talking about file size.

I think the ImportObj script by default does some scaling of the object, but you can disable that by passing the parameter "exact" to the command in your keyboard shortcut.

So set up the shortcut like this:

_ImportObj exact


- Michael

Hi Tom,

> The models I want to build in MOI are not organic, but perfect for Nurbs.

Ok, so very different than the elephant one you posted previously, right?

It's because you posted that elephant model that I thought you were possibly trying to do modeling of creatures or animals.

- Michael

Sorry for confusion - I should have posted another model as example
You are very caring about MOI users, a big thank for that !

No problem Tom - although I'm still a little worried from your description of focusing on extracting curves and using the Network command that possibly your actual models are not as NURBS-friendly as you might think.... The most NURBS friendly type of modeling is when you are primarily constructing things with 2D profile curves and doing extrusions, revolves, and boolean cutting operations. If your modeling requires all construction by using network over curves that are warping around in 3D space rather than primarily 2D curves, that often is in the "organic modeling" category that tends to be better done using sub-d modeling instead...

- Michael

"your description of focusing on extracting curves and using the Network command that possibly your actual models are not as NURBS-friendly as you might think...." "curves that are warping around in 3D space rather than primarily 2D curves"

You are right Michael. Seeing what is done with Rhino -and so Moi - like http://gallery.mcneel.com/fullsize/46535.jpg
I did not consider models like this helmet as "organic". So in your defintion and philosophy you would recommend to model sth like the helmet with polys ?

What about jewelry, vehicles or furniture where curves warping around in 3D space are unavoidabl ?. Of course it is possible to model any object with polys, but when it comes to manufacturing Nurbs should be the way to go.

A problem I see is to achieve smooth continuity when joining curved surfaces ...

"""""""""""""""You are right Michael. Seeing what is done with Rhino -and so Moi - like http://gallery.mcneel.com/fullsize/46535.jpg
I did not consider models like this helmet as "organic". So in your defintion and philosophy you would recommend to model sth like the helmet with polys ?""""""""""""

For that helmet model, The question would be more "why are you starting that model in a polygon modeler?"

The recommendation would be to start and finish it in MoI!

Sketching, conceptualizing and completing that model in MoI would be much faster than a polygonal modeler for sure...

Yep. I have to agree. More and more I'm finding it much easier to just start in MoI instead of SketchUp or Lightwave. The workflow is really a thing of genius and I'm continually amazed at how well everything is thought out.

Hi Tom,

> You are right Michael. Seeing what is done with Rhino -and so Moi - like
> http://gallery.mcneel.com/fullsize/46535.jpg
> I did not consider models like this helmet as "organic". So in your defintion and philosophy you
> would recommend to model sth like the helmet with polys ?

Well it kind of falls into a gray area where it requires using a more difficult toolset in NURBS and you are outside of the clearest and strongest areas of NURBS modeling which is when you are able to generate most of the model using 2D curves and booleans rather than trying to work with all 3D swooping shapes only.

But if you can dissect the helmet down to one surface to be generated for each smooth section of the helmet, then that can help a lot to make it work well with NURBS - it's easier to do things with NURBS when each smooth section of the model corresponds to one large sweep or network with possibly some additional trimming done to carve out any irregular boundaries. If you try to approach NURBS modeling in a small "patch by patch" manner as you might with polygon modeling, that does not work very well.


> What about jewelry, vehicles or furniture where curves warping around in 3D space are unavoidabl ?.

Well these all tend to be fundamentally more difficult models to create than mechanical models, ... they are often again in a gray area where they are not necessarily an automatic fit for NURBS modeling unless you are willing to go up a steeper learning curve than what is required for mechanical models.


> A problem I see is to achieve smooth continuity when joining curved surfaces ...

It's this type of "patch by patch" approach trying to build a single smooth skin out of multiple separately constructed surfaces that you really want to stay away from when doing NURBS modeling - if you are getting into a situation where you are worrying about that, it's a big sign that your approach would fit better with sub-d modeling than with NURBS modeling.

If you're going to be doing 3D swoopy surfaces with NURBS modeling you really want each smooth piece of the object to be made of one single large surface that is built in one go, rather than trying to build little pieces at a time that are supposed to be smooth to one another. Often times this means making the surfaces to extend out a ways so that they can follow a simple initial boundary, and then introduce some details by cutting chunks away using trims or booleans, rather than trying to directly construct a bunch of little surfaces "patch by patch" that try to hug complex boundaries directly. Hope that makes sense.... Try to think of NURBS modeling as using surfacing to create large simplified "stock pieces of material" that can then be trimmed to produce the final result. This is very much the case for mechanical objects where the stock pieces are done with simple extrusions. But if your 3D swoopy object can fit within that framework than it can probably be done ok with NURBS too. If your 3D swoopy object has lots of branching in it or little small localized lumpy details, then those are things that don't fit in well with NURBS.

There are some tips here for people coming from a poly modeling background, they may be useful:
http://moi3d.com/forum/index.php?webtag=MOI&msg=4865.2

The key difference between NURBS modeling and polygon modeling is that boolean and trimming operations are a major focus for NURBS modeling and when you see something like a smooth surface that has some zig-zag like overall outline shape, the way you do it in NURBS is to make a big surface out of a few simple boundary curves to generate an extended piece of stock material, and then the zig-zag outline comes from trimming some of that large surface away to get your final result.

- Michael

Thanks Mr MOI for the valuable insights into the strength and weaknesses of a Nurbs-modeller.
indeed I am in the grey zone between poly and nurbs with my current task - will do the work both in C4D and MOi to compare and evaluate the workflow - maybe post and example to see what the nurbsexperts have to say,.
Burr and Chip, do you stick to where MOI is shining (2D curves, boolean, etc) or do you work in the "greyzone" too, as Michael has called the area of "organic" shapes.

Btw, Michael, what about automotive design - a carbody is the fine art of joining many surfaces with different curvatures -precision required but at he same time deep inside the "greyzone" ?

Hey tom,
So "the grey zone" can be defined in different ways. There is also a black and white zone for me. Your elephant if in the white zone, has that helmut in the black zone. There is nothing about that helmut (or a car) that cant be done with NURBS... Although for patch modeling a car, you need a robust toolset. The helmut? Not so much. Also, your level of proficiency comes into play with the grey zone. Alot of NURBS guy may whip that helmut off pretty quick. Alot may struggle with parts. Etc.

A human face? Poly's for sure.

Hi Tom,

> Btw, Michael, what about automotive design - a carbody is the fine art of joining many surfaces with
> different curvatures -precision required but at he same time deep inside the "greyzone" ?

Well, there are some NURBS modeling programs such as Alias Studio that are focused on doing production automotive design using NURBS surfaces. It's also a super expensive program with a very high technical learning curve with a whole lot of difficult to use very specialized commands.

MoI is not focused on that type of use that requires a huge amount of technical training - MoI is much more focused on trying to make the strongest areas of NURBS modeling (again, mechanical design using 2D curves and booleans primarily) to be more user friendly to people without requiring an advanced math degree and a year of training just to operate the program. So MoI does not have the exact same NURBS toolset as Alias Studio for example - it has not been a priority to try and develop those exact same tools in MoI since it's not a fit with what MoI is trying to focus on.

I think it's likely that at some point in the future I'll try to expand MoI somewhat more into that territory and try to add more tools for matching surface continuity, but currently if matching surface curvature between a patchwork of NURBS surfaces is what you need to do, you would want to do that in a different NURBS modeling program such as Alias AutoStudio or Rhino with the VSR shape modeling add-on and not MoI. But those are complex and finicky tools to use so you should expect to spend a lot of time and energy going up a very steep learning curve before you would be very productive with them.

- Michael

You can make some car surfaces with Moi! ;)

Just make some tricky adjustments!



MOI- surface refiniment from M-dynamics on Vimeo.

I have a bit of a different take on the 'gray area.'

For the longest time, I felt it was just easier to use SubD modeling with polys to get what I wanted. I still prefer it for much more organic shapes which are 'eyeballed' for their form. Things like gloves, or soft goods, or cartoon characters are certainly better in a poly program which supports SubD modeling. Especially if it doesn't need to be 'exact.' This doesn't fit your helmet as it need accuracy, so I think NURBS would be best.

I really like SketchUp for poly modeling even though it's SubD packages aren't nearly as sophisticated as dedicated SubD programs like Modo or Lightwave Modeler or nVil. In particular I like it's ability to create quick interiors with instancing-- something MoI doesn't yet have. It also has the best sketch renderer on the planet.

But now I'm pretty much gonna start every project in MoI, move it to SketchUp for 'sketch renders,' and move it to KeyShot for final renders. It's just terribly faster starting in MoI. Unless it has to be a soft organic, I believe it can be faithfully executed in MoI. This includes vehicles as well.

Just so you understand, when Michael talks about 'surface continuity,' he's talking about a very esoteric subject, and one which is most difficult for all CAD Solid modelers to execute. Rhino needs a third party set of plugins. Alias has a decent implementation- in fact it's fairly intuitive and something I do hope Michael takes a look at implementing some day. For now, a 'good' MoI modeler can use some BLEND tricks with small NETWORKs to get pretty damn close. In fact, if you care to look, there have been some stellar car models created in MoI.

So, the trick is to eventually think in NURBS. A whole different strategy is required for modeling something like the helmet, but it's very possible, and I suspect much easier than you may think. Unfortunately, I'm in the middle of a big project or I'd like to take a shot at it and make a video tute. Maybe someday.

If you think about it, it's a symmetrical object, which has different parts which can be 'joined' or 'booleaned' together with fillets added along the seams. The main helmet should be rather simple to create with a few different surfaces --try taking a sphere, show the points and use the points to resize it. Once you have a decent top half, you can chop it up with different boolean curves.

The "gray area" is more "clear" since Maximilian Smirnov did SubD script and its sub-tools...(privet Max :) )
Hope you'll return into this scripts givin'us opportunity to delete splits or weld together two control points ;)
Thanks Max !

M

Thank you all for answering - I appreciate your helpful input, great forum here !

"For the longest time, I felt it was just easier to use SubD modeling with polys to get what I wanted. ....
But now I'm pretty much gonna start every project in MoI ... It's just terribly faster starting in MoI ..."

I tend to stick with subD mainly because I am just very at ease with my SubD modeller, but the more
I look at MOi the more i am fascinated how much power is hidden behind its minimalistic surface.

"So, the trick is to eventually think in NURBS. A whole different strategy is required for modeling"

This is the biggest challenge ... having worked for a long time with SubD my brain is programmed to think of
building up a model following a specific workflow - falling into the "SubD - trap again and again ;-)