Full Version: Considering buying MOI to do this

Show messages:  1-8  9-28  29-48  49-68  69

Hi Marco,

> ...reading your last post looks like Moi has a new display routine....a new and smoother mesh shader.

I'm glad that you like it! That's actually a screenshot of V3, it's just the standard MoI display engine that looks like that.

If you open that file you should see the same thing over there, if you want the exact same shading go to Options > View > "Lighting options" and set it like this:



- Michael

Image Attachments:
Marco_lighting.jpg 

---

Hi Michael,

In Rhino I started to re-construct the model by changing the degrees and control point of the curves (each segment). I managed to make a lot of single span surface.
That is the reason I thought the rials structure do interfiere with the sweep action.

Another doubt that I have is why the outcome of the intersection between my rounded cube and my lines projected, meaning the ''triangles'' that I am using as rials have so many control points on the curved segments. Rhino also gave me these curves with a LOT of control points.

I re-built them and then the sweep was much more cleaner. That Is the reason why I am confused by the ''- Sweep does not use the control point structure of the rail curves directly"

Thanks so much for your dedication

---

Hi Michael. I wanted to ask you a few things regarding to the new proposed technique.

1. How did you find the rial curves for the segments (attachment)

2. Is there any reason why you connect one of the 'radius' of the curve to the perpendicular part of the other one and not with the 'radius' of the other one ?


Thanks a lot!

Image Attachments:
Bruno_sweep2.jpg 

---

Hi Bruno,

re:
> I re-built them and then the sweep was much more cleaner. That Is the reason why I am
> confused by the ''- Sweep does not use the control point structure of the rail curves directly"

Maybe the dense curves also had some slight oscillations which can be difficult to see, and the rebuilt ones are smoother.

Or it could also be a matter of what's called "parameterization" of the curve, which has to do with how evaluating the curve might have spots where marching along the curve in the curve's parameter space may be either stretched out or compressed together instead of producing evenly spaced points.

Anyway, just to demonstrate it try sweeping the 2 cases in the attached file and then look at the control points of the rails.

Dense control points do give more potential for a curve to possibly have little tiny wiggles in it though so it can be good to be suspicious of them but it's not necessarily automatically bad.

- Michael

Attachments:
SweepControlPointExample.3dm

---

Hello Michael.

Ah...ok...

No new things but anyway thanks a lot for this lighting options!

Yes...I already new that kind of things since the old days of the original "Majik Primer: MoI's Lighting Options and You!" forum thread :)
(at http://moi3d.com/forum/index.php?webtag=MOI&msg=5222.45)

One of my favorite plugins that I use daily.

Thanks. Have a nice day.

---

Hi Bruno,

> 1. How did you find the rial curves for the segments (attachment)

To set it up I drew in a line across using "Perp" snap like this:





Then I used that line to trim the profiles and used Rebuild (http://moi3d.com/3.0/docs/moi_command_reference10.htm#rebuild) to make those into single segment curves instead of being made up of 3 segments.


> 2. Is there any reason why you connect one of the 'radius' of the curve to the
> perpendicular part of the other one and not with the 'radius' of the other one ?

It's to give a more gradual transition to the straight parts. If you match radius to radius it will make for a very sudden transition in a stretched out zone here:



If you turn on control points for that you can see the straight segments have stations that are slanting and stretching a lot, especially with elongating into a kind of sharp spike at the tip:


That type of slanting will usually have some undesirable effect on the surface's shape in those areas, it just becomes more difficult to control the shape.

If you match it the other way and then look at the same section you can see a much more orderly arrangement of stations, the more that they resemble regular planar formations it tends to give more controlled surface shaping there as well.



This kind of thing comes into play when you're trying to connect together outlines that have zones of different length from each other. Just matching things by "small feature to small feature directly" can produce shearing and so can just an overall "match full length to other full length".

- Michael

Image Attachments:
bruno_alignment1.jpg  bruno_alignment2.jpg  bruno_alignment3.jpg  bruno_perp1.jpg  bruno_perp2.jpg 

---

Hello Bruno,

another way to get cleaner surfaces is to use a G2 or G3 blends for the radii of your cube and not a G1 (classical radius). Have a look in the fillet command. The resulting curves of the projection/slices will have also a better continuity. The resulting surfaces don't have seams. Have a look to my first post.

Have a nice day
Karsten

---

Another thing is to use more "natural" perimeters (not tortured lines ;)

---

Oh. I understand.

I managed to understand how you did it. The problem is that I did not rebuild them. So if you check my attachment you will see the control points look much more dense.



When I did that in Rhino (tried before), I mean to reconstruct as a single line the 3 segments...They told me that was not possible since of the discontinuity of G1 joint to straight line its no point to do that. Maybe I understood wrong ? I will check the rebuild option in MOI.

I guess I am getting closer. I am rebuilding the shape with your technique. I hope this translates good to Modo (polygons).

I would also like to ask you Michael if the screenshot looks good for you. I believe I followed correctly your instructions, but I forgot to rebuild the 3 segment lines.
Did you do the same with the 'arcs' on the other part of the 'triangle' ?


--------------

In order to get the rials for the segments (highlighted curves in the attachment) . Did you intersect the perp lines to with the 'old' sweep ? Did you use the project command ?



I am asking cause what I did was to extrude the perp straight lines and find the intersection with my old surface. What I am not sure if the direction of my extrusion. In my case I mirror the perp straight lines (you can see it at attachments) and then loft.

Let me know if I was not clear. I really want to render this soon and I get anxious.


Kind regards

Image Attachments:
Screen Shot 2017-12-14 at 10.56.22 AM.png  Screen Shot 2017-12-14 at 11.05.59 AM.png 

---

Checking your first post now. When you say there are not seams what do you mean ?

Can you attach a reference image?

Thanks so much.

---

I checked the file again and your post. It looks definitively better with G2/G3 radius on the cube.

In order not to do this model for the 12 time hahahah, I mean before proceeding to use that technique. Do you think by making the cube with different continuity on the radius I will solve it just doing sweeps (intersection as rials and two curves as profiles). If you believe this will reduce the amount of time of re-tweaking ...I will do it.

Just want to know your opinion.

---

Hi Bruno,

> So if you check my attachment you will see the control points look much more dense.

They look dense but they look well arranged without anything getting bunched up and crossing over each other. I wouldn't worry so much about the density just itself, the question is does it look better and is it shaped more like what you want or less like what you want.


> When I did that in Rhino (tried before), I mean to reconstruct as a single line the 3 segments...They
> told me that was not possible since of the discontinuity of G1 joint to straight line its no point to do
> that. Maybe I understood wrong ? I will check the rebuild option in MOI.

That's true that under normal circumstances you don't usually want to rebuild lines and arcs together into a single segment. But trying to control segmentation matching for a particular surface construction can be a special case where you do want to do it, if it produces a better looking surface result.

It's up to you to decide if it's better or not, if you were happy with the results of your old way you don't have to do it any differently.


> I would also like to ask you Michael if the screenshot looks good for you. I believe I followed correctly your instructions,
> but I forgot to rebuild the 3 segment lines.

Looks pretty good to me! You may test doing a mesh export to .obj file so you can see what that looks like.


> Did you do the same with the 'arcs' on the other part of the 'triangle' ?

Yes.


> In order to get the rials for the segments (highlighted curves in the attachment) . Did you intersect the
> perp lines to with the 'old' sweep ? Did you use the project command ?

I didn't make those rails, they are edges that were created from the sweep result. What I did was join together the triangles so there were 2 closed big triangle curves (with the same number of segments in each, so you will get segment-to-segment matching), and 2 small arc curves between them. Then I selected the triangle curves to use as profiles, then ran Construct > Sweep, and then picked the 2 small arcs as the rails. The curves that you are highlighting are then generated from the sweep, because the sweep will be split up at points where there are segment endpoints in the joined segmented curves.

Hope that makes sense, time for bed again over here!

- Michael

---

Hello Bruno,

<<<In order not to do this model for the 12 time hahahah, I mean before proceeding to use that technique. Do you think by making the cube with different continuity on the radius I will solve it just doing sweeps (intersection as rials and two curves as profiles).>>>



It's only one way - but in my opinion you don't Need more than one Profile.

Have a nice day
Karsten

p.s.: Quick and dirty Rendering with Blender/Luxrender

Image Attachments:
cube.png  LRTest.png 

---

I will try out. I did not think of making it with one profile curve.

Lets check thank you !

---

Michael. I checked the model (following your technique) in Rh...and it seems to have certain discontinuity. Is it possible the software shows the model different ?


There is a strange brake :




Are your surfaces also 'broken' in those edges in Moi ?

Thanks a lot

Image Attachments:
Screen Shot 2017-12-19 at 10.13.08 PM.png 

---

Hi Bruno, yeah I do see the same thing over here. maybe that technique is not so good after all...

Maybe you could minimize that by extending the corner segments of the small side a bit too like this:



It can be a pretty difficult task to work with tightly bent curves.

As sort of a general rule you usually want to keep things like that sharp and use filleting later on to put in the tightly rounded corner areas rather than having them baked directly into your surfaces. But I'm not sure that's really an option for your case here.

- Michael

Image Attachments:
bruno_corner.jpg 

---

Why this corner is so distorted ?
Does exist a rational reason of that ?
When a form is not "natural" it's normal that you obtain such difficulties! :)
Make simple and harmonious :)

---

>Why this corner is so distorted ?

I am not sure what do you mean with distorted. It is a poly-surface (filleted cube G1) and then intersected lines.

>Does exist a rational reason of that ?
I guess the reason why 'this' exist it is cause I exist. Which is the same reason why things (not talking about nature)
have a shape, someone existed. This is not a challenging answer, I actually wonderer why this shape exist.

>When a form is not "natural" it's normal that you obtain such difficulties! :)

Can you clarify what should be 'natural' . Again you are challenging my perspectives (which is good). There are thousands cars
I can see from my window now. Do you believe they are 'natural' ? Or they are more logical for a cad software than a cube with radius ?

I apologize if this sounds too much. The truth I am frustrated. I am wondering how is possible I saw so many crazy shapes on CAD software.
I ve seen an AUDI R8 in CAD, Its been weeks making a cube with G1 filleted edges and grooves :(

Thanks a lot for participating. I am looking forward for suggestions :)

---

Hi Bruno, well a couple other things you could try to smooth out the discontinuity would be to trim away a small slice around it and use Construct > Blend to build a blend surface through it. Also Rhino's MatchSrf command could be worth a try, if all the other shaping is looking good other than those discontinuities.

- Michael

---

Maybe something like that.?




The radii on the sidewalls are all equal. 8 Profiles 2 rails. In the corners the radius will be stretched. If you want avoid that the radius/diameter has to be big enough that he can fit to the distance of the rails in this area. But think about what happens then - the path of the grooves will moves up/down there.

Have a nice day
Karsten
p.s.: The curvature of isolines shows a little swinging

Attachments:
test2.3dm

Image Attachments:
crvtr.png  test2.png 

---

Show messages:  1-8  9-28  29-48  49-68  69