More Polys in Sculptris? YES, DOUBLE!

[RogersX]

Is there any way to get more polys out of Sculptris? I hit the 1mil -1.2 mil polys and it always crashes. Is there a way to allocate more memory or something to get the poly count in Sculptris higher?

(Yes, I know about the tips and ways to reduce polys when using Sculptris)

(Update, see post #6)

[Serek]

The only way to get more than 1.2M tris (if that's your limit) is to buy a stronger comp. Even so, you'd hardly get past 2 million tris...

[RogersX]

I'm running core 2 duo, 8 gigs of ram with Win 7 64 bit and a 512meg card. I thought it was a limit of the software for 32 bit machines that will not allow it to go over the standard 1gig off ram? I wish I could get it to use the rest of my ram. 32bit Photoshop had the same issue, 64 bit, not so much .

[Serek]

Sculptris is a 32bits native software but it doesn't make any difference if you run it under a 64bits system. It won't neither crash nor give a better performance.

Either way, it's not ZBrush which means no subtools, no subdivision management. So, besides the pure art practice, a heavy model made with Sculptris is almost unpractical for most uses such as texturing, rendering, animation, etc. Even trying a retopology in Blender may result annoyingly sluggish. IMO, I'd say that staying close to 1 million tris is fair enough to get very nice models. Some users may like even more, I prefer to stay around the 150-250k, with 500K as a max.

[RogersX]

Sculptris is a 32bits native software but it doesn't make any difference if you run it under a 64bits system. It won't neither crash nor give a better performance.

Yeah, that is kind of painfully obvious

Either way, it's not ZBrush which means no subtools, no subdivision management. So, besides the pure art practice, a heavy model made with Sculptris is almost unpractical for most uses such as texturing, rendering, animation, etc. Even trying a retopology in Blender may result annoyingly sluggish. IMO, I'd say that staying close to 1 million tris is fair enough to get very nice models. Some users may like even more, I prefer to stay around the 150-250k, with 500K as a max.

Obligatory but fair enough. I know, I am backwards etc etc. Sorry, I hate Zbrush with a passion Interface is ick... But then again, that is Off Topic.

[RogersX]

Okay, after doing more research with various Google queries I came across this little page:
http://maketecheasier.com/increase-m...-os/2011/08/13

Which uses an app made by this guy:
http://ntcore.com/4gb_patch.php

Basically what the app does is patch a 32bit app to take advantage of more memory in a 64 bit environment. I crossed my fingers, downloaded the app. Ran the patch and...

Perfect

Popped open a couple files that were pushing Sculptris to its limits. Anything else I would add to these files would instantly crash Sculptris. Typically the largest I could go was 1.2 to 1.3 mil tris. I opened two files, one at 1.2, the other at 1.3 mil tris, selected the entire model and copied and pasted it into the other one. I instantly got 2.4 million tris. The file saves fine, opens fine and runs fine though a bit slow. At first I got nervous as it would hang and the Windows blue loading circle would come up which would always happen before a crash. But it pulled through and didn't crash once (Unless I pushed past 2.6 mil tris). Working is slow but I think that is due to pushing so many tris on screen at once. A heavier processor and or graphic card would probably change that.

Sculptris now crashes at about 2.6 mil tris and beyond which is double what I got before so that is great. The new ceiling allows me to easily compose and compare different files side by side or swap compatible pieces in and out without having to decimate multiple meshes before doing so to meet the old requirement cap. New life, old app! = Sweet!

This was done with Sculptris Alpha 5, should work identical in Alpha 6.

[Serek]

o.O Never heard of it. Thanks, RogersX, I'll try it later with my AMD X6 1090t, 8gb RAM and MSI 560GTX-ti Hawk

[justadeletedguy]

Zbrush Million polygon models are useless outside of Zbrush! .. For a game model (humanoid, animated) 12-tris .... Cars and machines maybe 30-50 k tris..... Poser/DAZ models that render very realistic humans 80k Quads .... This is what all the Posts about Decimation and Retopo are about.

Try this, Build your model in pieces, start with the first sphere, sculpt until your at a good stopping point (100-200k), export that piece out and Decimate with Blender. Save it as "Blamodel-low20k.obj" ... Now import the low-poly, and a new sphere to work on in Sculptris. Dont touch the low-poly, just sculpt on the new sphere. When you reach a good stopping point with that, Decimate that one .. and keep going.

[michalis]

@rogersx
what serek tried to explain (I think so) is that a huge mesh may be possible but, what you gonna do with it?
You have to retopo it, you gonna export it for rendering etc. No app likes these huge meshes.
A good idea could be, export/save this mesh, then subdivide it as possible and use this for retopo work. Then replace it with the original and shrinkwrap it.
I also find a ~200 - 600 k enough in most cases. Get used to Y tool and see how to manage the mesh. A 1M dynamically tessellated mesh is the equivalent of a ~10 M zbrush dynamesh. Something like that.
BTW don't hate zb. Not that bad. I find it great as so many other apps. The zb UI is much stronger than people believe.

[Adamx]

@RogersX This is a great find. It does work form me, I was able to push it pass 3mill, but it is responding slow to brush movements. Also I checked the difference with hex editor and the edited file shows 5 bites changed. The file still registers as 32bit, but so what, all we care about is to work with more polygons.

[RogersX]

Zbrush Million polygon models are useless outside of Zbrush! .. For a game model (humanoid, animated) 12-tris .... Cars and machines maybe 30-50 k tris..... Poser/DAZ models that render very realistic humans 80k Quads .... This is what all the Posts about Decimation and Retopo are about.

Try this, Build your model in pieces, start with the first sphere, sculpt until your at a good stopping point (100-200k), export that piece out and Decimate with Blender. Save it as "Blamodel-low20k.obj" ... Now import the low-poly, and a new sphere to work on in Sculptris. Dont touch the low-poly, just sculpt on the new sphere. When you reach a good stopping point with that, Decimate that one .. and keep going.

I do do that. Thanks. Even covered it in my previous post in this thread (Decimator in old V. of ZB). Like I said, higher poly count would just be nice to pop open other files, copy/paste, compare side by side, swap pieces in and out etc without worry of crash. Now I have it

@rogersx
what serek tried to explain (I think so) is that a huge mesh may be possible but, what you gonna do with it?
You have to retopo it, you gonna export it for rendering etc. No app likes these huge meshes.
A good idea could be, export/save this mesh, then subdivide it as possible and use this for retopo work. Then replace it with the original and shrinkwrap it.
I also find a ~200 - 600 k enough in most cases. Get used to Y tool and see how to manage the mesh. A 1M dynamically tessellated mesh is the equivalent of a ~10 M zbrush dynamesh. Something like that.
BTW don't hate zb. Not that bad. I find it great as so many other apps. The zb UI is much stronger than people believe.

I know what he tried to explain. I understand what you are explaining. Just because I do things my own way doesn't mean I do not understand yours. I mean its cool, I understand normal work flow for 3d artists etc. I work a little messy and all over the place and do things slightly differently from the ground up. May not be optimal I understand. Thanks

[RogersX]

@RogersX This is a great find. It does work form me, I was able to push it pass 3mill, but it is responding slow to brush movements. Also I checked the difference with hex editor and the edited file shows 5 bites changed. The file still registers as 32bit, but so what, all we care about is to work with more polygons.

Aye, sweet! Yeah actual work with brushing is slow but having different pieces of the overall model in separate files is a good work-around to that if your project permits. For instance if making a sci-fi guy you can do helmet, arms, legs, torso, gun etc all in different files then bring into one file for composition. That way you just have to move and compose them in the master file, no brushing etc.

Pretty cool and exciting! Helps extend the life of the app, at least for me anyway

[michalis]

If you like to have great performance on a 6M dynamically tessellated mesh then try 3dcoat newer builds.
I mean, someone should rework the code of sculptris.
I had no problem working on a 5M voxels in 3dc, when a 32 bit OGL build was for OSX.
Now I can go up to 10-20 M of voxels (as 64 bit) and still good performance.
But the surface LiveClay mode (similar to Sculptris) can easily handle 6 M sc like meshes. Still lucks of the great Sc tools but this isn't the point we're talking about.

What I mean is that good performance is needed. And the limits of Sc are around ~500k-1M for the moment.
You may also noticed that braking symmetry kills performance.
You may noticed that stopping dynamic tessellation improves performance.

BTW Your methods may be different RogersX, but better visit other users threads first before start presenting your thoughts about their workflow.

[arch111]

While I agree with you michalis about better code is needed in sc, I allso don´t see the point of having huge meshes either.
Sculptris handles 1 million + in my system, but the mesh is useless. I can´t hope to edit it further once it´s that dense.

And for all who dislike zb, it´s a professional software in constant development. Dont forget that the plugins are what make the tools so useful. In my case I find Decimation Master invaluable, alongside retopology-tools.

And another thing, in order to use the model it has to become smaller, so as long as the workflow is inside sc, sure more polygons are welcome. That seems to be what pixo are aiming for with the new version of zb, less need for other apps and keep the workflow inside zb.
This is what I have been waiting for since version 3.

Mealea, I too think another version of sc will see the light of day. The more features are packed into zb, the more of the "basic" tools will be added to sc I think.

pixo have nothing to loose by doing this but it would benefit all to educate users before taking the leap into zb

[MentalFrog]

I didn't like Zbrush at first either, but I kept at it and now I really like it.

It's not end-all be-all, but it's another tool to implement into your pipeline. It's a great tool and one that shouldn't be ignored. If you understand its strengths and weaknesses you can get the most out of it and help streamline your workflow. Sculptris is also just another tool. You just need to learn what to use and when.