Has any body noticed or is this how houdini is supposed to work?
When I mirror a polybevelled surface the bevelled surfaces are not bevelled.
I am working on a mech spider and after I mirrored and consolidated seams.The subdivide is giving me some very weird results.
I have no idea how to upload an example but I will do so as soon as possible.
Instead somebody try this in h 7.0.192.
Create a box sop.
Apply a poly bevel to all surfaces of the box(using ctrl +a)
Mirror the resultant object in the y axis.
M<ove it till the seams almost match and then consolidate the seams.
Then I think you can see what I am talking about.

Since I am from 3ds max and have been using houdinifor very little time
(this is my second week) I realise that I could be doing something wrong.
Please tell me what happens or whether this is a hitch in Houdini,my own inssstallation or how I overcome this.



Thanks in Advance
Lanchka

How are you consolidating the seams? Everything seems to work fine for me. Are you just upping the consolidate seams tolerance in the mirror sop?

The mirror in my case (I mean with the box). misses out the bevelled surfaces and the subdivide produces some very dissimilar halves.

When dealing with the mech spider. the mirror seemed to work fine
but when I subdivided I got a very bad result.

I have tried using a fuse sop.
As to consolidating the seams I tried mirroring the object in the y axis
and then I move the mirror till the vertices lined up and then increased the tolerance level slightly till I got welded vertices

How do I upload a few files so you can see for yourself?

Cheers
Lanchka

Could you give me your mail id so I could mail the files over if you dont mind??

hey lanchka,

i'm getting the results that you are most likely seeing, but there's a good reason why it's happening.

the bevel is actually not causing the issue here. going by the example you have given, go back to that and bypass the bevel. you will still notice that a mirrored box offset with point's consilidated still gets screwed up at a subdivide.

the problem is occuring in the mirror. while you are consolidating the points, the two adjacent poly's are still staying there and are now co-planar. those co-planar poly's is what's causing all the grief.

while i can't seem to find a way to delete co-planar primitives (nor could find anything through searches), there is a way to get rid of them if you're model is being reflected around one of the axis.

based on your box->bevel->mirror example, go about like you normally would with your model and right after the mirror, append a delete sop. In the delete sop, enable ‘Bounding’ and the following parameters (size and center will have to be based on axis of reflection.. in this case, it was mirrored around the Y):

Bounding Type: “Bounding Box”
Size: 1 .01 1
Center: 0 .5 0

If you are viewing the delete in the Viewport, you will see a thin blue box that it is centered down the reflection line. Since it's getting everything in that, it's getting rid of all the co-planar poly's. if it's deleting other poly's that you don't wan't, change your size of the bounding box till it only get's the co-planar ones.

on another note, if anyone else wants to chime in here about getting rid of co-planar poly's, that would be wonderful.. because there's gotta be a way to do that and if not, it should definitely be an RFE…

hth,
dave

I found out what I was doing wrong.
When I ran a blast operation I happened to miss 3 or 4 very minute polys.
So the mirror seemed to work but the subdivide messed up due to the interior faces.
Houdinis procedural workflow saved the day(or rather 3 hours of modelling).I ran through the ENTIRE NETWORK and finally found the hitch.
Silly old me.
But still one thing is vague to me.
When I bevelled all faces of a box primitive and mirrored the result.
The final bevelled faces of each initial poly were missing.

Thanks for the help.I really appreciate it.
Love
Lanchka

Ok, I see what you mean now. If you turn on display of primitive normals, then you can see that there's actually two primitives sharing exactly the same points. You can just fix up the problem by deleting the polygon prior to mirroring. ie. With your display flag set on the PolyBevel, do Tab > Blast in the viewer. Select the primitives that are going to be inside the model after mirroring and then right-click. Now your mirror and subdivide should work.

Speaking of PolyBevel, do people also use a value of 1 for the Repetitions so that it produces quads at the corners instead of triangles?

Lanchka:

But still one thing is vague to me.
When I bevelled all faces of a box primitive and mirrored the result.
The final bevelled faces of each initial poly were missing.

I took a look at your file. The reason why you were missing those initial poly's is because your mirror sop was set to only mirror the group of primitives ‘6-25’.. So the the prim's not in that group didn't get mirrored… make sense?


edward:

If you turn on display of primitive normals, then you can see that there's actually two primitives sharing exactly the same points.

exactly.. this is how i found out what was going on. besides blasting them away before the mirror, what about other cases of co-planar polys (prims that share the same points) that may not be so easily taken care of? is there any way to get rid of them with expressions or vops or possibly a custom otl? any thoughts on that?

Speaking of PolyBevel, do people also use a value of 1 for the Repetitions so that it produces quads at the corners instead of triangles?

hadn't tried it until just now.. actually i never went through with getting 7 due to lazyness, but finally went ahead and just installed to see what you were talking about.. quite nice. especially with the bevel types (flat, low and high density roundness, cornered).. those will all come in handy. what i would like to see added to that is if low or high density roundness is selected, a ‘falloff’ parameter shows up allowing the user to create very tight or very broad fall off rates of the rounded bevels.. i.e. a value of 1 would equal what is there now, less than 1 would be a tight rounded edge with a fast fall off, greater than 1 would be a broad rounded edge with a slow fall off. basicially it would just space out the individual repetitive bevels exponentially at less than or greater than 1 and linearly at 1. does that make sense at all? if not, i'll create an image of what i mean..:wink:

WOAH :shock:

holy crap, look at all this great stuff.. man i knew i was being lazy, but i didn't know how plain stupid i was being for not trying out 7. this is amazing..

good work sesi. keep it up.






…man, it really is just unbelievable to see all this..

Yeah, definitely, I've asked for a falloff value for the rounded bevel too … I think of it like almost a “crease weight”.

no seriously.. i mean really.. are you freakin KIDDING me?

:the new help

:scatter sop

:network boxes

:new time slider with crazyness

:all the flip book stuff with block editing and exporting

:all the new channel editor stuff


and actually i just found a the whole list in help that explains everything new so i'm not even about to get in to it… plus you all probably think this is old news.. but good lord. blows me away.

What version were you using before?

Yeah, definitely, I've asked for a falloff value for the rounded bevel too … I think of it like almost a “crease weight”.

yea, that would be nice. it could even work with crease sop.. that way you can set different edges to different crease weights. and if there is no crease weight attribute at the time of the bevel, then it just has a standard fall off rate parameter with slider. kinda like the overide crease weights in subdivide.

What version were you using before?

6.1 ops: