3D-Coat 3.5 updates thread

[Andrew Shpagin]

Usually "letter" updates ABCD... are only for win but some really critical cases because it will be too hard to maintain all just because of time that spent to make build. For example I am spenting less then minute of manual work to make Win build, so I am making Win builds frequently. But if someone have serious problem on mac or linux I may do update too.

[AbnRanger]

In the process of trying to make a quick video demonstrating the new Transform tool in the Retopo Room, I got some funky action going on there...UNDO not working properly. 3DC scaling a layer that wasn't even selected. Weird things, man....weird things.

http://www.screencast.com/users/dnashj33/folders/Default/media/14752c0f-17a5-4551-bbad-3638e50e8b4f

[philnolan3d]

I just did a quick test and noticed that you can use Transform on polys and points, but not edges. Would be nice to have the option for all selection types.

[b33nine]

Transform in retopo room is awesome. Basically gives us a very simple poly modeller that can hopefully be expanded upon.

[AbnRanger]

Transform in retopo room is awesome. Basically gives us a very simple poly modeller that can hopefully be expanded upon.

Yeah, if the soft (gradient) selection from the Sculpt room can be added, it opens up more possibilities for importing primitives (from the OBJ pens or elsewhere) and using them to block out initial shapes (box modeling). It also makes it easier to readjust major portions of a mesh if one brings a model in from another application and intend to use that as their initial retopo mesh.

[michalis]

I'm not sure if this the right topic for this.

I was playing testing the new mac build. I left the app for some hour and just quitted. But when opened today I found a test.obj.dp in recent items. This was in 3dcoat/temp/ as obj.

Just imported it in blender. Here it is. I was testing the curve tools. Interesting but someone should explain. How can I have lower dense meshes like this? etc etc . Something is cooking here...

[philnolan3d]

I believe the Curve tool is a lower res mesh. But it's just like using a primitive (also low res) and it's just used as a mold for the voxels, then it's discarded. You can see this low res mesh if you use the Curve tool and hold W for Wireframe before actually creating the voxels.

[michalis]

But I also found it as autosaved test.obj.dp.

Phil, the funnier is that we can export the curve as obj to use in other apps. We may like to use it as retopo mesh as well.

What is funny is when I opened it in blender or in 3dc. I realized that this mesh was subdivided 2 times. Imported in zbrush and reconstructed subdivisions. See captures.

Pressing W (wires) we see a triangulated mesh but its not the same when exporting.

[3dCoatWannabe]

I just had my first 3D Coat crash.

I'm using 3.5.14A CUDA DX64 on Win7.

I saved several revisions as I was working, and then I made some changes to 3D Coat's layout (moving the popup windows around).

I saved my workspace and the layouts to my Win7 documents area (at least the folder that contains 'my documents'). I made a new folder named 3D Coat Layouts (would be great if 3D Coat could default to someplace on Win7 that it can actually write to, as the Layout folder it first tried didnt' have permissions).

I closed 3D Coat, and when I opened it again, I tried to open my saved revisions for the last few hours work.

One after another, 3D Coat crashed on each revision.

Then, I noticed that my layout was back to the old layout.

I restored my workspace and layout from the ones I just saved.

Then, I could open my 3D Coat file revisions.

I believe this is related to my changing and saving the workspace and layouts, as I've never had a 3D Coat crash before this, and this was the first time I've changed and saved these.

I can see in the bug report I submitted, lines like this:

Can't load image file "Shaders/CustomSampler8.dds"

After restoring my layout and workspace, I can close 3d Coat and it works perfectly.

BTW - a thought occurred to me. How many are using ECC RAM on their boxes?

Without ECC RAM, there's no way of knowing if a 3D Coat crash is due to bad memory, or to 3D Coat.

I use ECC RAM, and even my RAID has ECC RAM.

Maybe part of 3D Coat's diagnostics could include a memory tester that could be run overnight (or one could be recommended), as I'm sure sporadic crashes are frustrating, and this would help to eliminate one possible cause to some degree?

----

I did a "New' and 'continue' to discard my changes. Now I'm back to crashing when I try to open my saved files (that opened just a few minutes ago, so I know they are ok).

Should I install the 14E version (this is 14A)?

Is there some config I should delete?

[philnolan3d]

Also don't forget to use autosaves (in the preferences). I can't count the number of times that has saved my butt (no pun intended).

[Andrew Shpagin]

I just had my first 3D Coat crash.

I'm using 3.5.14A CUDA DX64 on Win7.

I saved several revisions as I was working, and then I made some changes to 3D Coat's layout (moving the popup windows around).

I saved my workspace and the layouts to my Win7 documents area (at least the folder that contains 'my documents'). I made a new folder named 3D Coat Layouts (would be great if 3D Coat could default to someplace on Win7 that it can actually write to, as the Layout folder it first tried didnt' have permissions).

.......

I got several crashes related to UI layoit. I don't know if it was you.

This sort of crashes was fixed in the latest 14D revision.

[AbnRanger]

I got several crashes related to UI layoit. I don't know if it was you.

This sort of crashes was fied in the latest 14D revision.

It does seem like the memory issues are gone for the most part, in DirectX anyway. Seems like I can squeeze quite a bit more resolution in a scene, now.

[3dCoatWannabe]

I got several crashes related to UI layoit. I don't know if it was you.

This sort of crashes was fied in the latest 14D revision.

Yes, those crashes were me. I guess I'll try the latest 14E and kill off Options.xml.

BTW - any chance of being able to move windows outside 3D Coat's main window, like with Photoshop?

That would really free up a lot of space for me, to be able to put them on a 2nd monitor.

[Psmith]

Just a bit of news from Raul:

He has managed to solve the disconnection problem he was having with the first version of LiveClay for polygons. He mentioned that solving this issue will also help the following:

"it will allow not just dynamic subdivission but static local

refinement too or other derived surface tools."

I've asked him to explain this last part - so I'll let you know when he does so.

Greg Smith

[haikalle]

Just a bit of news from Raul:

He has managed to solve the disconnection problem he was having with the first version of LiveClay for polygons. He mentioned that solving this issue will also help the following:

"it will allow not just dynamic subdivission but static local

refinement too or other derived surface tools."

I've asked him to explain this last part - so I'll let you know when he does so.

Greg Smith

Thanks for sharing this with us. It's great news.

[brook]

I have been doing some tests in 3.5.14x. I use mostly the retopo and uvs at the moment. those tool sets are developed to the point where you can reliably and quickly get work done. autopo also is looking even better with new options.

the voxel sculpting work flow needs more refinement. the poly count is just about there. you can work with a 10mil poly model. that's a good benchmark. the memory usage is great and most of the brushes have a even level of speed . so on a high poly count model you can find the right brush size to get quick strokes and all the brushes work relatively at the same speed. the new resample in surface mode is a KEY development. surface mode is much faster at high poly counts then voxel mode. I find myself sculpting and subdividing in surface mode the then popping into voxel mode to remesh the surface and fix stretching polygon areas. then back to surface mode. it works because only the stretching polys are resurfaced and the rest are left untouched. over all as the poly count goes up more sculpting is done in surface mode. so at around 5mil polys 70% of the sculpting is done in surface mode.

the new brush engine:

the overall feel of the brushes in voxel room is quit nice now. the clay has a nice feel.

in surface mode things are REALLY messed up at the moment. I started a thread in the voxel bugs forum. about the this.

surface mode bugs thread with PIC

basically the brushes are not staying on the surface. as the surface changes the brushes totally miss the changes in concavity and convexity. it seems to effect almost all the brushes. when your sculpting you can not get a consistent result. and the strokes have artifacts. so you might get half a stroke under the surface and half on top and big creases in the middle. or no stroke at all. It is very difficult to work with these brushes like this. I had to stop because you just cant get clean strokes that stick to the surface. and I think it effects smooth also. so the basic foundation of the brushes have some bugs. it looks like the stroke is not orienting itself to the normal/tangent correctly and updating in new positions. and the lighter the depth the worse the effect becomes. so doing very light detail changes are almost impossible. such as using brush alphas for detailing etc.

so overall the new brush engine seems to work except for the fact that the brushes in surface mode have a problem conforming to the surface. if that is fixed I think surface mode and the brushes in general are looking quite usable.

[Gilded]

I have been doing some tests in 3.5.14x. I use mostly the retopo and uvs at the moment. those tool sets are developed to the point where you can reliably and quickly get work done. autopo also is looking even better with new options.

the voxel sculpting work flow needs more refinement. the poly count is just about there. you can work with a 10mil poly model. that's a good benchmark. the memory usage is great and most of the brushes have a even level of speed . so on a high poly count model you can find the right brush size to get quick strokes and all the brushes work relatively at the same speed. the new resample in surface mode is a KEY development. surface mode is much faster at high poly counts then voxel mode. I find myself sculpting and subdividing in surface mode the then popping into voxel mode to remesh the surface and fix stretching polygon areas. then back to surface mode. it works because only the stretching polys are resurfaced and the rest are left untouched. over all as the poly count goes up more sculpting is done in surface mode. so at around 5mil polys 70% of the sculpting is done in surface mode.

the new brush engine:

the overall feel of the brushes in voxel room is quit nice now. the clay has a nice feel.

in surface mode things are REALLY messed up at the moment. I started a thread in the voxel bugs forum. about the this.

surface mode bugs thread with PIC

basically the brushes are not staying on the surface. as the surface changes the brushes totally miss the changes in concavity and convexity. it seems to effect almost all the brushes. when your sculpting you can not get a consistent result. and the strokes have artifacts. so you might get half a stroke under the surface and half on top and big creases in the middle. or no stroke at all. It is very difficult to work with these brushes like this. I had to stop because you just cant get clean strokes that stick to the surface. and I think it effects smooth also. so the basic foundation of the brushes have some bugs. it looks like the stroke is not orienting itself to the normal/tangent correctly and updating in new positions. and the lighter the depth the worse the effect becomes. so doing very light detail changes are almost impossible. such as using brush alphas for detailing etc.

so overall the new brush engine seems to work except for the fact that the brushes in surface mode have a problem conforming to the surface. if that is fixed I think surface mode and the brushes in general are looking quite usable.

Are you hitting "enter" periodically while in surface mode to bake the surface details down and remesh?

[brook]

Are you hitting "enter" periodically while in surface mode to bake the surface details down and remesh?

unfortunately that does not seem to be enough. In addition to the stroke problem I am also getting little spikes in surface mode where the brush can not find some verts. so when you smooth etc it gives spikes. I have to go into voxel mode and do some cleanup work to get rid of spikes. I figure these may be due to the fact that the brushes can't find the surface correctly. so if the stroke problem is fixed the spikes might be fixed also.

[BeatKitano]

unfortunately that does not seem to be enough. In addition to the stroke problem I am also getting little spikes in surface mode where the brush can not find some verts. so when you smooth etc it gives spikes. I have to go into voxel mode and do some cleanup work to get rid of spikes. I figure these may be due to the fact that the brushes can't find the surface correctly. so if the stroke problem is fixed the spikes might be fixed also.

That is an annoying bug indeed. It's been there forever unfortunately and prevents me from using surface mode permanently.

[Laticis]

Been doing some small test lately with the current builds and auto UV Mapping.

So far 09 seems to produce the best results (expected results that is) after that things don't work as I believe they should.

Please excuse my following terminology:

A simple Roman type column imported into 3DC 09 will give me a very clean finished UV Map with an evenly sized/distributed chequer pattern. v10 clean chequers but missing half the column in the uv map...the latest version v14E makes a mess out of the chequer and a disproportioned UV map.

PC 64bit Win 7 Cuda GTX480 I7960 12Gig Ram (No Internet Connection)

I can redo test if required and post images but more so I wanted to know if this is a known issue and if anyone else is getting the same results.

So far build 09 is the one I keep going back too.

Cheers Ady

Laticis Imagery

[brook]

I have attached a details description explaining the bugs with surface mode brushes. and in general the attributes that could be changed in order to eliminate the artifact. everything is explained in the pic.

[artman]

I have attached a details description explaining the bugs with surface mode brushes. and in general the attributes that could be changed in order to eliminate the artifact. everything is explained in the pic.

Thanx for taking your time to expose the problem so clearly.

Did you try 3.5.03 ? It works really well.I hope Andrew can revert surface mode to how it was before.

[Andrew Shpagin]

I have attached a details description explaining the bugs with surface mode brushes. and in general the attributes that could be changed in order to eliminate the artifact. everything is explained in the pic.

Thanks for deep description. But look this screenshot from ZB - it is how default clay tubes works on obstacles. It does not (and should not) capture hills because it is intended to cover cavities. This is how Mod & Rapid works in 3DC.

[artman]

Thanks for deep description. But look this screenshot from ZB - it is how default clay tubes works on obstacles. It does not (and should not) capture hills because it is intended to cover cavities. This is how Mod & Rapid works in 3DC.

Here is what I got on my side in Zb.Clay tubes clearly capture hills like 50%

3DCoat effect is too radical.

Also try sculptris Clay mode of standard brush it has very good feeling,exacly like ancient 3DC SF.Clay.

[artman]

same thing in mudbox.

It does go through a little but not much. Otherwize you always see the previous strokesd and it gives very not nice results like the top right corner of Brook's image.

[artman]

Here is Sculptris clay mode,its exactly like 3.5.03 but 3DCoat Clay brush was even better because it had a little nice flattening effect.

When you are doing realife Clay sculpting you get that feeling too when you press your fingers when adding layers of clay.

That image is really the problem.Its not usable like that,it does not happen in ZB or MB.

Also 3DCoat's following effect leaves a lot of artifacts compared to MB/Zb.It is rotate along brush motion that is too jittery I think.

With a few small fixes I think 3DCoat brushes could really rock.

[BeatKitano]

Well.... if these sirs say it's broken, maybe it is, but please don't fix it, make it an option Andrew... I like to use it to even surfaces...

I agree about that last screenshot though, alphas are drawn with dots not curves that something not right (even with dot draw deactivated).

Ultimately those problems could be ancient history with a complete custom brush preset system...

[Digital777]

Well.... if these sirs say it's broken, maybe it is, but please don't fix it, make it an option Andrew... I like to use it to even surfaces...

I agree about that last screenshot though, alphas are drawn with dots not curves that something not right (even with dot draw deactivated).

Ultimately those problems could be ancient history with a complete custom brush preset system...

What i think should be done more is to simply have options to adjust the way tools work so you could use them in various different ways. I see no reason why this could not be done really as it just needs a drop-down list and maybe some sliders etc somewhere. The same could be done for other tools also to make them have multiple workflow uses.

With things like smoothness/jittery artifacts like shown in the comparison screens above although it would be better to have it default to a nice smooth cut there is times when maybe you would want a broken look. The MB result looks much better but roughness/artifacts could work well for some tasks like sculpting rocks or terrain etc so if it just had a smoothness slider and defaulted to the best result but still allowed a rough result if you set the value down that would be much better than just replacing it with a smooth result.

If this was done with enough variation options and also with a updated preset system to quickly re-use good settings again it would be really great.

[3dCoatWannabe]

Also 3DCoat's following effect leaves a lot of artifacts compared to MB/Zb.It is rotate along brush motion that is too jittery I think.

With a few small fixes I think 3DCoat brushes could really rock.

There's a ZB/MB thread on NewTek, where 3D Coat was mentioned favorably (apparently, ZB has a lot of nasty UV issues). http://www.newtek.com/forums/showthread.php?t=117522

If a ZB/MB user decided to try the 3D Coat demo after reading that thread, and got the horrible result above, he'd give up. The smooth version should be the default, trying to match what the ZB/MB folks actually do well. Having an option for the current 'rough' mode, sounds like a good idea also.

I also get a lot of artifacts when carving off the end of an object by drawing a rectangle and holding control - and when importing Lightwave objects with lots of curves.

I'd definitely vote for smoothness as the default.

[michalis]

ZB has a lot of nasty UV issues

If UVmaster is what we're talking about, this is an excellent tool and saved me from a lot of 3dcoat UV issues.

Most of my UVs are made with 3dcoat, but you never know, sometimes. Blender is an excellent UV editor BTW. Try to construct a shirt using a geometrical pattern as tex and we'll see what UV editor is the best.

Lets be reasonable, lot of cases, different solutions, I don't believe in stand alone apps anyway.