News:

 

Topic: Bridge Tool that works with NGons with different amounts of vertices  (Read 8069 times)

0 Members and 1 Guest are viewing this topic.

  • No avatar
  • Posts: 28
  • Vertex
April 20, 2013, 07:48:59 am
I talked about this in another part of the forum, but I thought I'd make a proper feature request.

A bridge tool where you can select any two NGons (which might have a different amount of vertices), or any two sets of edges, and bridge them easily would be very nice. While the tool is active you could drag left/right to adjust the number of segments.

My reference is the modo bridge tool: http://www.luxology.com/training/video.aspx?id=36

  • No avatar
  • Posts: 2103
  • Polygon
April 20, 2013, 11:11:28 am
My reference is the modo bridge tool: http://www.luxology.com/training/video.aspx?id=36

I see the presenter hid the underlying triangulation/poly-flow created, first with sub-d, then at the end, where I thought I would see the triangulation/poly-flow due to the bridging of 2 basic solids, he disabled edge view. LOL

  • No avatar
  • Posts: 28
  • Vertex
April 20, 2013, 03:44:08 pm
Yes, but there has to be some triangulation if you're bridging between different amounts of vertices. I wouldn't expect that to be any difference in NVil. But then again, I'm not a programmer and if it's possible to somehow make it only quads (by adding coplanar vertices?), that would be fantastic.

I see the presenter hid the underlying triangulation/poly-flow created, first with sub-d, then at the end, where I thought I would see the triangulation/poly-flow due to the bridging of 2 basic solids, he disabled edge view. LOL

  • No avatar
  • Posts: 2103
  • Polygon
April 20, 2013, 04:58:44 pm
Yes, but there has to be some triangulation if you're bridging between different amounts of vertices. I wouldn't expect that to be any difference in NVil. But then again, I'm not a programmer and if it's possible to somehow make it only quads (by adding coplanar vertices?), that would be fantastic.
The 2 methods I have seen are, the triangulation of the bridge, or N-Gons produced on the side of the bridge with the least edges (the edges are broken (vertex added on the edges to increase edge-count)).
Both leave bad geometry/poly-flow.


  • No avatar
  • Posts: 28
  • Vertex
April 20, 2013, 10:19:23 pm
Well sure, I haven't seen a perfect method yet either. All I know is that the bridge tool in modo is pretty good and have saved me a lot of time and that it would make a very good addition to NVil. There are also many cases where perfect quad egde flow doesn't matter (or isn't possible). Do you mean that you would rather not have an improved bridge tool?

The 2 methods I have seen are, the triangulation of the bridge, or N-Gons produced on the side of the bridge with the least edges (the edges are broken (vertex added on the edges to increase edge-count)).
Both leave bad geometry/poly-flow.

  • No avatar
  • Posts: 2103
  • Polygon
April 20, 2013, 10:43:16 pm
Do you mean that you would rather not have an improved bridge tool?

I do not see how that would actually improve the bridge tool.

If I do find I have incorrect number of edges on one side of the bridge, I edit to correct, rather than relying on an underlying algorithm that will probably not give the result I want.

Anyway, it is not up to me if such an implementation/function is made. I am just putting forward my own opinion as you are.




« Last Edit: April 20, 2013, 11:01:45 pm by steve »

  • No avatar
  • Posts: 28
  • Vertex
April 21, 2013, 06:31:48 am
I guess we're modeling very different things. And maybe we're thinking of different things altogether. It's not about having an 'incorrect' number of vertices on one side, because of sloppiness, I'm talking about one shape on one side that needs to look a certain way bridging to another shape that needs to look a certain way. 'Editing' down a 7 vertice hole to a four vertice hole to match the other side alters the look of the model into something I don't want. I think the tools should help you model what you want, not be something you work around.

It's not like the old functionality would disappear. How is more functionality and flexibility not an improvement? You could still only make bridges between equal vertice amount holes if you wanted to.

I do not see how that would actually improve the bridge tool.

If I do find I have incorrect number of edges on one side of the bridge, I edit to correct, rather than relying on an underlying algorithm that will probably not give the result I want.

Anyway, it is not up to me if such an implementation/function is made. I am just putting forward my own opinion as you are.

  • No avatar
  • Posts: 232
  • Spline
April 21, 2013, 10:24:58 am
I agree with Steve. Such may have been a fancy feature in Modo some years ago but I believe the practical value is actually very low. To me such was just a sophisticated way to ruin the topology. Further that tech imo was made obsolete through Sculpting apps.
If one needs some really twisted connections one imo is far better off to quickly push the model over to any Sculpting App (also free ones) and get crazy here. One afterwards can quickly retopo just the bridged parts with just quads. No canned bridging process will give the same amount of control.


  • Posts: 546
  • Administrator
  • Polygon
April 21, 2013, 10:45:12 am
really is no predictable way to automatcally do that, and with good planning and modeling it is a situation you shouldn't hit that often, so just do it by hand.

  • No avatar
  • Posts: 3760
  • Developer
  • Administrator
  • Polygon
April 21, 2013, 11:46:26 am
Monkeybrother, you must be doing some models that are very different from others. I can imagine that some models actually do not need very tidy topology. For example, a tree trunk with a special shape changing from one section to another.

From the observation, I think the key to the success of this feature is how to match the sharp corners between the two shapes. I have to admit that this is a very tricky one and so far I haven't been able to figure out an algorithm.

  • No avatar
  • Posts: 28
  • Vertex
April 21, 2013, 11:47:45 am
It's a feature request, that is all. You might not see the point in what I'm suggesting, but in a best case scenario thousands of people will be using NVil, all of them with different workflows and needs. You work in a certain way, I (and many others) work in different ways.
What I immediately liked about NVil was that I could configure it the way I wanted and work very fast, with this one exception.

I'm sorry, I don't get why you get defensive and argue against new features. I thought the whole point was to be better than other modelers, which means flexible enough for every user, not just the ones that already bought it.

  • No avatar
  • Posts: 28
  • Vertex
April 21, 2013, 11:51:57 am
IStonia: You posted while I was typing my previous answer. Like I said, I'm not a programmer, if it isn't worth the time I understand. And like I said, it's only a feature request. Thanks for looking into it.


Monkeybrother, you must be doing some models that are very different from others. I can imagine that some models actually do not need very tidy topology. For example, a tree trunk with a special shape changing from one section to another.

From the observation, I think the key to the success of this feature is how to match the sharp corners between the two shapes. I have to admit that this is a very tricky one and so far I haven't been able to figure out an algorithm.

  • No avatar
  • Posts: 28
  • Vertex
April 21, 2013, 12:04:04 pm
I encounter situations where I would need this tool mostly when I'm doing complicated hard surface modeling, things that aren't going to be deformed or sculpted. There's no point in trying to keep a perfect quad workflow in those cases, it's usually a waste of time and polygons. But also things like trees, like you said.


Monkeybrother, you must be doing some models that are very different from others.

  • No avatar
  • Posts: 3760
  • Developer
  • Administrator
  • Polygon
April 30, 2013, 10:41:40 am
It is done. Sort of, at least it does something.