Polybevel SOP issues

Forums Houdini Lounge Polybevel SOP issues

8620 17 1


stu: Member; 246 posts; Joined: July 2005; Offline

June 11, 2014 9:56 a.m.

Next on my list, the Polybevel SOP .

RFE's on the SI topic notwithstanding, I keep coming across an issue that I'm going to call a bug, and one that has caused me a lot of grief recently.

When I use the Polybevel SOP with the “High Density Round” option and “Repetitions” set higher than 1 I get a strange result at the junctions. The result is a strange deformation in the position of the newly created junction point geometry that results in a bumpy, non-uniform result, and certainly isn't what's expected. I've attached a file and an image, below. I've run into cases where the phenomenon is far more exaggerated but I just quickly threw together an example.

Whenever I need to round the corners and use a “Repetitions” value higher than 1, I use “High Density Round” rather that “Low Density Round” because I'd argue that the “Low Density” result is even less intuitive.

Again, I consider this a bug., and has been submitted as such: 62183.

Attachments:
polybevelSOP.1.hip (60.1 KB)
polybevelSOP_issues.jpg (200.6 KB)


stu: Member; 246 posts; Joined: July 2005; Offline

June 11, 2014 12:34 p.m.

Two more issues that I have with the polybevel SOP are as follows (at least one of these was mentioned in the SI thread but I want to keep these together for organizational purposes):

1. The newly created bevel doesn't create the surfaces with an appropriate radius. The result is an unexpected bulge outwards in the beveled edge, beyond what can be represented with a circle with the same radius and point count (see image below). My expectation would be that a beveled edge would create a circular cross section, adjusted to accommodate the incoming tangents obviously.

2. The edge lengths in the newly created bevel geometry differ in length. I'd expect the edges to be the same length, as seen in a polygonal circle. Again, see below.

Bug or RFE? I'll submit accordingly.

Attachments:
polybevel_issues2.jpg (204.5 KB)


stu: Member; 246 posts; Joined: July 2005; Offline

June 11, 2014 2:18 p.m.

As a further illustration of the “bulge” and the segment length disparity, the issues become more pronounced as the incoming polygon tangency becomes more acute (see below).

Attachments:
polybevel_issue3.jpg (145.6 KB)


stu: Member; 246 posts; Joined: July 2005; Offline

June 11, 2014 5:20 p.m.

Not to beat a dead horse, but here's an animated example demonstrating the “bulge” increase and chord length inconsistency on angle change. The grey line represents the polybevel SOP solution, and the black line is mine (and represents my expectation). I'm not animating any of the polybevel SOP parameters here, I'm just animating the angle on the incoming polygons to open and close over 80 frames. You can see that my black line maintains a circular radius with uniform span lengths while the polybevel SOP does not (see attached).

Again, black=what I'd expect (consistently tangential circular radius under all circumstances with uniform chord lengths), grey= polybevel SOP in it's current state (inconsistent deformation with non-uniform chord lengths).

Thoughts?

Attachments:
polybevel.mov (895.4 KB)


Doudini: Member; 333 posts; Joined: Oct. 2012; Offline

June 11, 2014 10:13 p.m.

this really needs some fixes. the polybevel sop is just lame in general imo.


stu: Member; 246 posts; Joined: July 2005; Offline

June 12, 2014 9:56 a.m.

RFE submitted: 62207

One could argue that the issue with the inconsistent radius is counter to the expected behaviour and therefore a bug, but I've logged it as an RFE because, technically, it's not “broken”. If someone from SESI wants to make a change to either one of these classifications it's fine with me (whatever get's the SOP some attention

).

So, to recap:

1. Bug submitted for the bumpy geometry.
2. RFE submitted for the non-circular radius and non-uniform spans.

The Polybevel SOP is a staple in my modeling work and I'd love to see it live up to it's potential. Like I said earlier there are other things as well, but these are the first two big ones for me.


stu: Member; 246 posts; Joined: July 2005; Offline

June 15, 2014 12:41 a.m.

http://www.youtube.com/watch?v=K8E_zMLCRNg [youtube.com]


Doudini: Member; 333 posts; Joined: Oct. 2012; Offline

June 16, 2014 12:55 a.m.

stu
http://www.youtube.com/watch?v=K8E_zMLCRNg [youtube.com]

lol


anon_user_37409885: Member; 4189 posts; Joined: June 2012; Offline

June 16, 2014 9:31 p.m.

2.7 million views for the cricket!


Sadjad Rabiee: Member; 1391 posts; Joined: Dec. 2010; Offline

June 17, 2014 4:08 a.m.

I learn an interesting way for creating “Fake Rounded Corner Bevel” in the “Houdini Mograph Title Design” from cmiVFX , Maybe it can help you 8)

https://www.youtube.com/c/sadjadrabiee [www.youtube.com]
Rabiee.Sadjad@Gmail.Com


stu: Member; 246 posts; Joined: July 2005; Offline

June 17, 2014 9:23 a.m.

Joker386
I learn an interesting way for creating “Fake Rounded Corner Bevel” in the “Houdini Mograph Title Design” from cmiVFX , Maybe it can help you 8)

Maybe, but I'd also like to see the polybevel SOP performing as expected.


Sadjad Rabiee: Member; 1391 posts; Joined: Dec. 2010; Offline

June 17, 2014 12:17 p.m.

However check it , maybe it can temporary help you :?

Attachments:
JKBevel.hip (76.4 KB)

https://www.youtube.com/c/sadjadrabiee [www.youtube.com]
Rabiee.Sadjad@Gmail.Com


stu: Member; 246 posts; Joined: July 2005; Offline

June 18, 2014 9:39 a.m.

Thanks for that, that's an interesting approach.


Doudini: Member; 333 posts; Joined: Oct. 2012; Offline

June 18, 2014 10:44 a.m.

i use this approach too for small bevels it works great but its not a solution for doing real rounded corners.


Sadjad Rabiee: Member; 1391 posts; Joined: Dec. 2010; Offline

June 18, 2014 11:28 a.m.

Doudini
i use this approach too for small bevels it works great but its not a solution for doing real rounded corners.

Yes , Exactly , I usually using this for Texts 8)

https://www.youtube.com/c/sadjadrabiee [www.youtube.com]
Rabiee.Sadjad@Gmail.Com


stu: Member; 246 posts; Joined: July 2005; Offline

July 14, 2014 6:34 p.m.

I've just noticed another idiosyncrasy with the polybevel SOP - it's behaviour changes based on the size of the object. If I have a small object and bevel the edges using the “Absolute” setting at, say, an inset of .01 I get a different result than if I scale the object up 10x before the polybevel SOP and change the “Absolute” inset to be 10x the value that I used for the object at 1/10th the size. In theory, they should be identical, no? I get an unwanted result if the object and the “Absolute” inset are too small.


Siavash Tehrani: Member; 710 posts; Joined: July 2005; Offline

July 15, 2014 2:37 p.m.

I think this is a precision issue, similar to how the Cookie SOP can behave differently depending on scene scale.

Btw you have my full support in this crusade stu.


stu: Member; 246 posts; Joined: July 2005; Offline

July 15, 2014 8:38 p.m.

Thank you! And yes, this smells like precision to me too.

Quick Links

                    
                        Search links
                        Show recent posts
                        Show unanswered posts