topic Re: Shell tool resolution in Modeling

Shell tool resolution

Anonymous — Tue, 23 May 2023 14:39:33 GMT

When creating a shell tool with a curve, I ended up have the magic wand setting at a low resolution, I think around 16. Is there any way to increase this existing shells resolution without replacing it with a new object?
Thanks!

Re: Shell tool resolution

Barry Kelly — Mon, 12 Dec 2011 03:13:59 GMT

The magic wand has no effect on the resolution of a shell.
In the attached image I had 2 different magic wand settings and the curved profile (was half of a long, flat ellipse) is identical.

The only resolution we can control is the resolution around the rotation axis of the revolved shells.

The resolution of the actual shell profile is all automatic based on what GS thinks we want. This was much worse than it is now during beta testing. We had a slight win to get the reolution improved but still it would be better if the user could set it.
Each curve in a profile has its own minimum and maximum number of segments based on a maximum deviation away from the true curve. I believe tha max number is set to 36. Hence small radius curves don't look to bad (look at the pointy end of the shell in this image) where as the flatter curves can look terrible and visually segmented as in this image.

There is a way of changing this but involves modifying registry values (Windows OS - there is an equivalent in Mac).

WARNING: modify registry values at your own risk.

These values can be changed and will affect what you see on that particular machine only.
I haven't played with them extensively but the tollerance value seems to be the most productive.
The lower the value the smoother the resolution of your profile.
i.e. the smaller the deviation from the true curve.
The existing shells will not alter straight away but will if you move or edit them in any way. Re-building the 3D view does nothing.
The image attached here has a tollerance value of 100 (which is the default).

The problem is as soon as the file is opened on another machine they will see the shell resolution set as with the default settings unless that machine registry has been modified as well.
So all machines that will open the file need to have the registry set the same - otherwise your model can change dramatically.
I stress I would not recommend doing this.
We need GS to implement an adjustable tollerence setting in the shell tool itself.

Barry.

Re: Shell tool resolution

Barry Kelly — Mon, 12 Dec 2011 03:16:12 GMT

The registry keys.
There is an equivalent in Mac but I am not familar with that.
Barry.

Re: Shell tool resolution

Barry Kelly — Mon, 12 Dec 2011 03:21:34 GMT

The same shells as in the previous image but with the registry value set to 10.

Of course this increases the polycount for the shell (from 4224 to 7486 in this case) which would probably be the reason why GS have set the limits as they have.

Barry.

Re: Shell tool resolution

Anonymous — Mon, 12 Dec 2011 15:17:36 GMT

That's great! Thanks, I will play with that.

Kind of a side question, is this also how I would have to change settings to increase resolution with curved section within a complex profile?

Re: Shell tool resolution

Barry Kelly — Tue, 13 Dec 2011 07:08:43 GMT

I haven't see a registry value for that.

The complex profiles seem to use the magic wand setting.
If set to "Best Match" it ignores all the settings and uses a maximum deviation of 1mm.
This will give the illution of a smooth profile i.e. no lines on the facets.

If you set it to "Linear" you can control it a little - the smallest deviation you can set is 1mm anyway so no difference there or you can set the number of segments in an arc/circle.
The problem with setting the magic wand to "Linear" is that your complex profiles will now show lines on all the facets.

Barry.

Re: Shell tool resolution

David Maudlin — Tue, 13 Dec 2011 13:26:05 GMT

mjkoarch wrote:
Kind of a side question, is this also how I would have to change settings to increase resolution with curved section within a complex profile?

One thread on this issue is here:
Custom Profile

David

Re: Shell tool resolution

Anonymous — Tue, 13 Dec 2011 15:26:39 GMT

Thanks for the thread link.
I was afraid that it was controlled by the magic wand tool as this thread suggests. It would be nice to be able to draw in section and 3d with actual smooth curves!

Re: Shell tool resolution

Mike96 — Wed, 03 Aug 2022 11:31:59 GMT

@Barry Kellycan you please explain where exactly we should change the registry entry?

I know you don't reommend it.

But I'll do it anyway.

Re: Shell tool resolution

Barry Kelly — Thu, 04 Aug 2022 10:17:51 GMT

That was so long ago I can't remember what the settings actually do.

But if you open your Registry Editor and look for these keys, I think these are the ones you are after.

But remember it will affect only your machine.

Open the file on another machine and the shells will be as they were before.

Barry.

Re: Shell tool resolution

Mike96 — Thu, 11 Aug 2022 08:44:24 GMT

Thank you!

My solution may be converting shell to morph, and then it should stay the same in other machines too.

Re: Shell tool resolution

Erwin Edel — Thu, 11 Aug 2022 10:13:00 GMT

Morph is pretty 'dumbed down' though. If you want composite shell roofs etc, morph won't do. Also editing a morph is a pain in the rear.

Just adding a simple resolution field, like most 3D objects have, would solve a lot. Also resolving a curve in equal segments, instead of ending up with an odd one out when making a circle fragment would be nice. Example: if you have curved segment, the middle part is generally not the top part, if all sides are equal height etc

I'd almost forgotten how long we've had this issue.

That said 90% of my shell roofs are actually zinc gutters or aluminum covering profiles, window sills and what not. Basically extrudable poly lines.