3D-Coat 3.5 updates thread

[haikalle]

Best release yet, must say. One thing would like to see

it is better to avoid long stretched triangles or too big faces in comparison to existing in scene. Boolens will eat them without problems but it will create inconvenient geometry for further edit. Also in rare cases booleans may fail - in this case it will lag a bit more and do nothing after all.

Subdividing your objects helps in this situation. But For hard surface modelling it would be nice to have a option subdiving without smoothing. That way we could add the needed resolution quickly into

object we are merging without loosing hard edge details.

Did I already said, best release yet Really loving the boolean algorythm. Gives really nice results.

[Andrew Shpagin]

Best release yet, must say. One thing would like to see

Subdividing your objects helps in this situation. But For hard surface modelling it would be nice to have a option subdiving without smoothing. That way we could add the needed resolution quickly into

object we are merging without loosing hard edge details.

Did I already said, best release yet Really loving the boolean algorythm. Gives really nice results.

Yes, we will include hardsurface smart tesselation a bit later.

Booleans are a bit different from usual. They should be more stable than usual boolens in other apps because I am collapsing too long triangles and avoiding special cases that are hard to booleans. In so way they are not absolutely precise but it gives better stability. So even multiple sequential booleans should work well.

[Andrew Shpagin]

And of course in final release I will include updated manual.

[polyxo]

I am very sorry to say Andrew but on my machine do not see all the improvements people are raving about recently, really not.

What I experience is that a lot of essential Tools which just feel broken.

See Screenshots of the 3.7 RC showing how freshly created Primitives look (I just chose Sphere Primitive and hit Apply).

The second shows what the Sphere-Tool looks like with a large Brush.

Also I fail with getting Live-Clay Tools under control.

I see that they are not part of the official release, but still find them very raw.

Win7 64, Cuda

[polyxo]

I should add:

The Primitives in Surface Mode seem to work without errors and the result looks very clean.

[JimB]

I am very sorry to say Andrew but on my machine do not see all the improvements people are raving about recently, really not.

What I experience is that a lot of essential Tools which just feel broken.

See Screenshots of the 3.7 RC showing how freshly created Primitives look (I just chose Sphere Primitive and hit Apply).

The second shows what the Sphere-Tool looks like with a large Brush.

Also I fail with getting Live-Clay Tools under control.

I see that they are not part of the official release, but still find them very raw.

Win7 64, Cuda

I can confirm ,I am using a grey shader and this bug was not very apparent until zooming in.

[Andrew Shpagin]

I am very sorry to say Andrew but on my machine do not see all the improvements people are raving about recently, really not.

What I experience is that a lot of essential Tools which just feel broken.

See Screenshots of the 3.7 RC showing how freshly created Primitives look (I just chose Sphere Primitive and hit Apply).

The second shows what the Sphere-Tool looks like with a large Brush.

Also I fail with getting Live-Clay Tools under control.

I see that they are not part of the official release, but still find them very raw.

Win7 64, Cuda

Thanks for finding it.

Looks like it happens when sphere radius is >60 and CUDA enabled. I will fix it and include in final build.

[polyxo]

Thanks for finding it.

Looks like it happens when sphere radius is >60 and CUDA enabled. I will fix it and include in final build.

Good. Thank you!

[Phizter]

Hi Andrew,

Using Liveclay/cutoff in version 3.7 leads to crash. I am writing here since product version 3.7 is not listed in Mantis.

Well anyways, thanks a lot for the update

[LorenzoS]

Hi Andrew, I don't know if 'RemoveStretching' was ever included in the Mac builds. Ive never seen the tool.

Can anyone else confirm?

[The Candy-floss Kid]

Hi Andrew, I don't know if 'RemoveStretching' was ever included in the Mac builds. Ive never seen the tool.

Can anyone else confirm?

Lorenzo "RemoveStretching" is in the Mac version , top left hand corner of the interface's tool parameters.

I sometimes find that using an existing prefs directory from a previous version creates issues like this and therefore would recommend going to

"users" > and deleting the "3D-CoatV3" folder and starting again to see whether this resolves the issue. Keeping copies of this folder between releases is a good idea

as it contains the preferences for shortcuts and many other parameter settings. So by deleting the folder you will lose these.

Maybe somebody could pinpoint exactly what to trash from that folder so you don't lose all your existing prefs?

[Andrew Shpagin]

Linux build uploaded

Hi Andrew, I don't know if 'RemoveStretching' was ever included in the Mac builds. Ive never seen the tool.

Can anyone else confirm?

Edit->Preferences->Show beta tools

[Andrew Shpagin]

Andrew I still have issue with clean surface:

File: http://www.beatkitano.com/3dc_bugs/volumeprb.rar (file uploading, active in about 20 minutes, ~50mb)

Steps: activate symmetry on x, apply clean surface.

Seems Undo after Clean surface is really dangerous - it corrupts internal mesh structure. Probably it was corrupted in so way in one of previous stages. I will try to fix it now, but looks like you have to voxelize head (or use backup), othervice no LC tools will work over it... Try LC there - it produces really dirty stroke.

[Andrew Shpagin]

Hmmm this means I need to voxelize every volume... in all scene since liveclay introduction ? I'm happy i didn't use the betas for pro work now

Thanks for spotting the issue though. Is it the possible cause for symmcopy explosion too ? I was pressing enter when that happened before 3.7 to fix it (sometimes it was exploding again and needed to directly convert back to voxel and back to surface to fix it), but though it was a temp solution.

Btw this happened with cleanup memory too in past builds, I don't know about this one though (I need to go now).

Anyway, I will discover it deeper and report. I think problem appears only after cleanup mesh and undo, not averywhere in LC.

Now I am investigating it deeply.

OSX build uploaded.

[farsthary]

Perhaps pre LC saved surface models has no consistent data structure when loaded into a recent build and that may cause those corruption issues.....

[Andrew Shpagin]

Hmmm this means I need to voxelize every volume... in all scene since liveclay introduction ? I'm happy i didn't use the betas for pro work now

Thanks for spotting the issue though. Is it the possible cause for symmcopy explosion too ? I was pressing enter when that happened before 3.7 to fix it (sometimes it was exploding again and needed to directly convert back to voxel and back to surface to fix it), but though it was a temp solution.

Btw this happened with cleanup memory too in past builds, I don't know about this one though (I need to go now).

What version of 3dc was used for last modification of file?

[AbnRanger]

I got the same issue BeatKitno had, yesterday, working in 3.5.27A. I was getting some artifacts, tried to use cleanmesh, and got the explosion. I backed up, and checked "Remove strecthing" then hit ENTER and got the explosion as well.

As he said, don't use these tools yet on production work. I like to use CreaseClay and InflateClay often, so I hope LC gets stabilized soon.

[LorenzoS]

OK, I've always had beta tools on but didn't see 'RemoveStretching' in the corner there.

Thanks guys.

[Andrew Shpagin]

Andrew: as stated in private to Farsthary: this file is a part of a scene that may (i don't really remember) predate LC introduction. It's a testing file for 3dcoat advancement. If this doesn't happen on a newly created scene I'm willing to start again (it's not that much work as this was a slow side project I've been working on on and off) but would obj export/reimport work ?

You may just voxelize the volume. It works more-less correctly over this mesh.

Question - was this scene saved with 3.5.27 or 3.7 ?

It looks like file is missing one important chunk, so if it was saved recently - it is problem to me.

[digman]

The mesh explosion happens even if is a default voxel sphere or imported model, merged etc. I think Andrew will just have to find the root of it...

version 3.5.27A Linux 64 bit non cuda, windows 32 bit non cuda.

The default sphere was changed using the move tool and I'm sure I posted some screenies of it already..\

Waiting to test the new linux build when it's uploaded

[Andrew Shpagin]

The mesh explosion happens even if is a default voxel sphere or imported model, merged etc. I think Andrew will just have to find the root of it...

version 3.5.27A Linux 64 bit non cuda, windows 32 bit non cuda.

The default sphere was changed using the move tool and I'm sure I posted some screenies of it already..\

Waiting to test the new linux build when it's uploaded

It is uploaded. And if possible, please link me to the post you referring.

[digman]

Andrew, here is the link to the posting...

http://3d-coat.com/forum/index.php?showtopic=6634&st=2360 (page 119, down toward the bottom of the page...)

That doodle was started from a default sphere.

Here is another screen cap I just took a few minutes ago of a slighty different problem.

Same version as my other post linux 64bit non-cuda 3.5.27A

default voxel sphere.

Move tool, LC inflate brush,

Rapid2 and clay brushes used.

I would use the move tool some, then cleansurface and cleanupmemory

use move tool again, then inflate brush set a 1.00

Rapid2 and clay brushes plus using cleansurface and cleanupmemory.

Holes start appearing in mesh when using the clay brush and the yellow cursor disappears over some holes...

[digman]

Digman, I let you the honor

I hope you find a solution to those issues because liveclay is very dangerous in the current state. I'll continue to use only removestretching, avoid clean surface/cleanup memory and symmcopy only after voxelization.

Ah, I remember seeing that one... I hooked up with Mantis a few days ago so now those darn bugs can get eaten by the bug-eating thingy Andrew has there...

[digman]

I was able on some test to get things to work normally again by voxelisation then returning to surface mode but the bug would strike again at random times. I would recommend saving quite often and for sure before running any clean surface or memory cleanup. Once this critical bug is squashed, that will be a good day for the LC/ remove stretching beta...

[Tser]

Best Linux build yet! good work guys...

Any chance of multi threading the bucket fill? as it now takes forever using only one thread, its quicker to use a brush and to fill by hand.

Also, the abrupt slow down after performing a lot of brush strokes is still there, but if I go to the Voxel Room and toggle the Proxy mode on, then off, the slow down vanishes and I can work at full speed again, even though I am working in the Paint Room.

T.

[AbnRanger]

Best Linux build yet! good work guys...

Any chance of multi threading the bucket fill? as it now takes forever using only one thread, its quicker to use a brush and to fill by hand.

Also, the abrupt slow down after performing a lot of brush strokes is still there, but if I go to the Voxel Room and toggle the Proxy mode on, then off, the slow down vanishes and I can work at full speed again, even though I am working in the Paint Room.

T.

+1. Bucket Fill is slow as frozen moloasses, indeed, but what I do to get around it is to click on "FILL LAYER" in the fill tool options panel. If I click on the object instead, it takes a long time. Clicking on the individual island that you want to fill, in the Texture editor is another way to get near immediate response. But I agree, it is otherwise painfully slow.

I really think Andrew needs to revisit the Paint Room once he gets all the Voxel/LiveClay work under wraps. It is starting to fall behind tools that wasn't even available a couple years ago. Mudbox didn't even get texture painting until 2009, and in only 2 years it has surpassed 3D Coat's painting capability, and of course it has REAL ambient occlusion. I watched an Autodesk webinar yesterday for Mudbox 2012 and was extremely impressed.

3D Coat still doesn't have Layer MASKS (not to be confused with brush alphas by the same name), and that is a major limitation in my opinion. I use layer masks in Photoshop like a baker uses flour, and am leaning toward using Mudbox to texture paint from now on,for this reason alone, not to mention the performance (although 3D Coat's painting has come a long way in that regard, when Andrew made it multi-threaded).

[digman]

I deleted my other post because, I realised that Andrew had stated that closed models work best, I did have a problem with a non-closed model as he stated when merging a model in surface mode.

This version Linux non-cuda 64bit 3.7.00RC is much more robust.

With some limited testing I have not produced any errors in LC, remove stretching, surface cleanup and cleanupmemory.

More testing tommorrow...

[digman]

After this I will post at Mantis, I have signed up...

Linux version 64bit non-cuda 3.7.00RC

I thought I give a quick test before going to bed, as I had done all my testing with symmetry turned off.

Default voxel sphere

I turned symmetry on across the x axis.

Use the move tool twice with remove stretching turned on and got bad stretching on one side of the axis.

Next used cleansurface.

The mesh had the explosion..

This is still the basic error I been getting in the past few version of both windows and linux.

[simmsimaging]

1320981469[/url]' post='69076']

+1. Bucket Fill is slow as frozen moloasses, indeed, but what I do to get around it is to click on "FILL LAYER" in the fill tool options panel. If I click on the object instead, it takes a long time. Clicking on the individual island that you want to fill, in the Texture editor is another way to get near immediate response. But I agree, it is otherwise painfully slow.

I really think Andrew needs to revisit the Paint Room once he gets all the Voxel/LiveClay work under wraps. It is starting to fall behind tools that wasn't even available a couple years ago. Mudbox didn't even get texture painting until 2009, and in only 2 years it has surpassed 3D Coat's painting capability, and of course it has REAL ambient occlusion. I watched an Autodesk webinar yesterday for Mudbox 2012 and was extremely impressed.

3D Coat still doesn't have Layer MASKS (not to be confused with brush alphas by the same name), and that is a major limitation in my opinion. I use layer masks in Photoshop like a baker uses flour, and am leaning toward using Mudbox to texture paint from now on,for this reason alone, not to mention the performance (although 3D Coat's painting has come a long way in that regard, when Andrew made it multi-threaded).

Fully agreed. LC and voxels are cool, but I bought 3D Coat to, well, coat stuff. I gave up on the paint room updates and bought Mudbox 2012 just a few weeks ago. There are still some things that 3DC does better but the speed is great and the simplicity of the interface is refreshing. I'm not totally happy with it, and am not sold on the long term just yet, but I feel your frustrations anyway.

b

[Andrew Shpagin]

the export part (head) was saved with 3.7 and the original file it's coming from was saved in 3.7 I think, if not: in 3.5.27A

But the missing chunk part is interesting: could it be those random messages i got sometime stating it couldn't save an undo file ?

I don't have a screen shot but basicaly it's a warning message from 3d saying it couldn't save an undo file it's a 00000000charactershere.dat file.

I remember seing someone mention it too in this thread I think.

EDIT: THERE

Seems I found the root.

"Save volume as 3B" saves LC info incorrectly.

That was the reason of problems with this mesh.

Fixing it now.