WarrenM, no problem I just explain my opinion about this :)

I just took a little offense to the "it's all PR" angle. It's not your opinion that's the problem, it's spouting off information that is factually incorrect that annoyed me. But anyway...

Hello,

I agree with Shapenoid, it would be great if that tech could make it into C4D. It seems like it would be very useful for tweaking models for concept and final rendering rather than fiddling with all details in NURBS that may not be necessary for manufacturing but will show up in the finished product.

Hi Stefan, it's definitely a very useful tool! But it's also not magic without any problems whatsoever, you wrote:

> and then don't forget that there are also no fillet problems there like we often encounter in MoI.

This isn't really quite accurate, there are examples of failed fillets in their own tutorial video here:
http://vimeo.com/85143356

It looks like in a large complex project that you could run into many badly formed fillets with it, it will probably depend a lot on the particular characteristics of the shape though.

It certainly looks like it's very useful for people doing sub-d modeling though, it should allow someone who is already very comfortable with sub-d modeling to be able to do some boolean operations to build their final shape, no doubt that is helpful!

It's a little bit odd that the final output does not look like it's ready to be a sub-d cage itself though, I had thought that was a focus of the Groboto mesh generation stuff. This may be a significant problem for some workflows since it looks like you can't just go in and tweak a few points of the generated result to adjust things...


I'd say that there are still many benefits to NURBS modeling for mechanical object construction, some of the good parts about NURBS modeling are not just that you can do booleans but also that you can generate large chunks of your model just from drawing 2D curves, and also that there is a high level of accuracy, like when you cut a circular hole you know it is really a circle that you're getting to a high level of tolerance. Also relating to accuracy, I've heard many times from Modo users that they greatly prefer MoI's snapping functions to those in Modo, so that's another pretty big consideration as well...

But it all depends on what in particular someone is looking for - if someone is doing sort of blobby semi-mechanical shapes and they're already very comfortable with sub-d modeling, MF looks like a great tool to help them while staying within Modo.

With MoI, you get an overall much broader tool that has a lot of things in the workflow that are tuned for mechanical model construction...

It's probably not a bad idea to have both! :)

- Michael

Mesh fusion is really an amazing exploration tool, thanks to the live boolean tree system, and the freedom from sds modeling. The integration in Modo is very well done.

This kind of crazy shape is really hot !
https://dl.dropboxusercontent.com/u/23530633/FusionImg/fusionWidjetsB.jpg

But ... the final mesh quality is not that great. As you can see in the video, you have to push the sub-d level from input object very high to avoid troubles, and even then, it doesnt have the surface quality like you can have with MoI. Reflection are allways a little bit wavy. Fillets (or stripes like they calles it) are sometimes a little bit out of control, radius is not constant, and there is for the moment no way to create various radius.

Of course it only concern picky prick like me, and you have to consider it's the 1.0 version of the product.

Also I hate retopo job, I think it's the biggest annoying 3d task ... and that's something I really like with MoI, I never have to retopo anything. Exports are clean for the rendering, and the mesher gives enough option to keep polycount low.

For someone who never used nurbs, it's a wave of fresh air, but for us, MoI user, it doesnt look that 'amazing'.

Hi Michael,

"""""""This isn't really quite accurate, there are examples of failed fillets in their own tutorial video here:"""""

Yeah, but there must have been some type of precursor to that video. He seems to be working on an "all tri's" model? Is there some type of rule about all tri's and subd?

"""""It's probably not a bad idea to have both! """"""""""

A man after my own heart!

Hi BurrMan,

>> Yeah, but there must have been some type of precursor to that video. He seems to be working on an "all tri's" model? Is there some type of rule about all tri's and subd?

No, the input geometry need to be all quad. If there are triangles in the input, mesh fusion will add an extra sub-d level on it to resolve the topology. If there is a n-gone, it will just not consider the input mesh valid.

Now, the result mesh preview is a procedural mesh, and for some reason, modo shows it triangulated. When you 'freeze' the result, you will only get quads on the final mesh (and few triangle here and there).

Hi PaQ,

> Now, the result mesh preview is a procedural mesh, and for some reason, modo shows it
> triangulated. When you 'freeze' the result, you will only get quads on the final mesh (and
> few triangle here and there).

Hmmm, I was wondering why the output was looking totally triangulated - so it isn't really.

But I'd think that the output probably has enough topology deviations from the original in collision areas that applying further subdivision on top of it will tend to introduce unwanted shaping side effects, as described here:
http://community.thefoundry.co.uk/discussion/post.aspx?f=4&t=84105&p=756925

- Michael

Hi Michael,

You are right indeed.
Here's a couple of images from a 'simple' boolean case.
It was a test I've done during the beta, so there a chance the final mesh is better now.



Draft Working Mesh : That's the procedural mesh you see during the modeling by default.
You can also choose the preview an airtight and an final airtight version, however it will still look triangulated at that poind.

Next images are the model when 'exported' or 'commit'.

Draft : Few triangles, fillets/stripes are still separeted mesh
Airtight : Everything is merged, but the topology is not very optimal (looks a bit like the MoI quad and tris export option)
Airtight Final : Much 'better' topology, lost of nasty micro triangles are gone.

However, there are still some triangles, and lots of high valence vertex, meaning than adding an extra subdivision surface will create wavy distortion.

Hi Michael !

Yeah !!!
I've both and I am very happy
one unique example in MoI : Circles with Tangent
Thank you
Thierry

I wouldn't get too excited about this video. Its a promotional/advertisement piece. Thats all. I'm telling you, there was an enormous amount of production, preplanning, and 3d modeling work put into this that they never show. If you think you're going to jump into this software and whip up a design like that, well, you're kidding yourself.

"""If you think you're going to jump into this software and whip up a design like that, well, you're kidding yourself""""""""

Really beanworks? This may be true for the guy's who load a software for 20 minutes then make global declarations about it as "useless", but, the tech you see is not "new".... What's new is Darrel got it integrated into Modo....

For someone who can kitbash with that tech already, the watch example is not very complex.

But a lot of people miss the point of a lot of it. Like the dynamic nature of the Boolean tree. It would be like MoI with parametrics.... Can I do it in MoI now, by "keeping all my objects" during Boolean processes and managing those? Yes! But that's not realtime either.

Seems like a lot of dissing the product from some who's take on things is "It cant replace this! or, It doesn't do everything!"

It's a great tool for Modo, for someone that wants that. Is it a "one click wonder tool for quad mesh final output?" No, but you'll have to show me one of those to start with, before that's even a discussion.

Seeing this really got me thinking of just how fantastic working with Booleans can be for modeling complex objects, we tend to think CSG Boolean methods (whether it's applied on nurbs or polygons) is old news and not relevant anymore ... but this video convinced me otherwise. I feel with explicit surface control point pushing a user can get carried away using very low level tools, that sometimes a few simple Booleans could achieve much faster.

This will marvelous if MF will be integrated in Groboto!
Because it's the same creator(s) but alas this next times his health was not so good!
http://www.groboto.com/v3/resources.html

I think with every 3d software there are always pro's, con's and workarounds, we're all individuals so naturally we all have a different opinion there isn't going to be one perfect tool, even the high end stuff have their glitches, I know, I use one everyday.

Just model stuff and be happy :)

Cheers