Jump to content
3DCoat Forums

3D-Coat 3.7 updates thread


Recommended Posts

Hi, I ve just got 3dCoat( EDU version) So I am just finding my feet with it so to speak. What I am curious about at the moment ,is upgrades . According to the Help tab in the software you can check for updates...it says I am up to date (3.701b) which is fine. When I hover my cursor over the downloads at the head of this thread it says 3.702C, but I believe this is a beta and that I am running the "stable" version? What do folks do about this....do you download the latest betas? And does this just replace the version you are running or do you have then 2 versions running? And how often are the betas updated, should I come back often to check or would I be on some sort of Email list informing me.

Personally I run the very latest version, but I always keep the installer for the previous version just in case. Honestly only once have I found the need to use it. However it's up to you, have the very latest features or have the security of knowing its a tested, stable release.

Link to comment
Share on other sites

  • Reputable Contributor

LiveClay tablet bug has been fixed. The package 3D-Coat-V3-7-02C has been updated. Please download it again.

Downloaded new updated 3D-Coat-V3-7-02C beta Jan. 1, 2012 running 64 bit non-cuda version on linux ubuntu 10.04 LTS

Surface Mode, LiveClay brush

On my linux version the problem is fixed if you do not use spacing as the tablet pen (intuos 3) works correctly when spacing is not used.

Select spacing in the brush parameters and select 10% . The depth and radius problem returns.

Deselect spacing and the problem goes away.

Thank you Sergy for tracking down this nasty bug. I know you will squash the spacing issue too...

Picture included

Top 2 strokes without spacing selected (stroke is correct)

Bottom 2 strokes with spacing selected at 10% (stroke is incorrect at the start of the stroke)

Edit: Plus a big thanks for squashing the lag bug, I forgot to thank you for that one. :yahoo:

post-518-0-93783600-1325432517_thumb.jpg

Link to comment
Share on other sites

  • Advanced Member

For me it's not working,but I found a thing odd,if in the right panel ("Options") I change the "zero pressure radius" to 0, the first time I do the stroke, the bug is gone,the pressure and radius are modulated correctly,but the second(and the others) stroke have the dotting effect.

If I change again the parameter (putting again zero to "zero pressure radius"),the bug is gone again,but again only for first stroke.

Link to comment
Share on other sites

  • Advanced Member

Hi SERGYI, tried the new build(03 january 2012)

WIth spacing enabled is fixed,it seems working really well,thanks a lot.

Without spacing enabled,the behaviour is the same as previous build,so there is a circle(full brush radius but not full depth at the start of the liveclay brush stroke).

Thanks for working on this,as with spacing enabled it seems possible to have better quality stroke brush,I'm happy with this partial fix.

Link to comment
Share on other sites

  • Member

Current cloning code in 3D-Coat has to be rewritten from scratch to allow sub-pixel precise cloning to make the cloned data look exactly like the photo original. This is a must have, because current 3D-Coat clone tool blurs the pixels and clones with data loss,

in 3D-Coat 3.7:

ORIGINAL_AREA_pixel_data != CLONED_AREA_pixel_data

you do not get the original photo-data, but a compressed/blurred copy of the original photo now.

You could argue that onto a model that has bad stretched UV-s - plus painting farther parts not directly under the camera - you cannot project a sharp texture, because it will be blurred, stretched. We need to be able however to save a super-sharp cloned texture (new 3D projection engine) within the UV boundary and see this sharp texture in Photoshop. Blurring, stetching, viewing angle differences (user painting the sides of the model) must be eliminated by always making the brush to "paint" (rather project brushed-pixels) perpendicular to the poly surface. Write this data to the texture- memory. Handle paint brushes when they touch the models surface farther from the camera like they are always directly under the camera. Create a "multiple view-angle paint engine" so if user also paints the models edges, 3D-Coat internally "turns the projection camera" to set its facing right over the brush and projects perfect perpendicular-to-poly surface pixels onto the texture map. The user does not see this, he only sees, he is painting far, the sides of the model. So there will be NO stretched/blurred/cloned pixels because of not ideal camera-toward-surface angle projection.

If 3D-Coat could:

1. do excellent projection painting delivering exact original pixel data, where original photo area and cloned area is 100% the same

2. At least would be able to allow voxel painting/photo_source texture cloning on two resolution levels, going higher TEXture res if user wants: this only because of MARI and Mudbox do it as well. :-/

3. have a per-vertex deformable - not like the current! per-brush deformable - projection paint mesh

I think if you could make this a reality, 3D-Coat could become a market leader in texture painting, voxel/LC sculpting and retopo.

A set of sharpness-regulator buttons for the render room would be really good, since a lot of 3D-Coat users maybe will want to save the voxel-painted, rendered surface to be used as 2D texture. So it can be used in Photoshop.

This. THIS. THIS!

I cannot say enough how important and awesome such changes would be. A lot of texturing work relies on projection painting which always results in a blurry texture in 3d coat. This needs to be adressed!

Link to comment
Share on other sites

  • Reputable Contributor

Tablet pen spacing bug seems fixed on 3DCoat Version 3.7.02C January 3rd, 2012 running on Linux 10.04 LTS using a Intrus 3 tablet and pen. I'm using the 64 bit non-cuda version of 3DCoat.

So far so good as I go through testing all the brushes and strip brushes. I will report back if I have any problems.

With renderdemon still having problems maybe could it be in the difference of Linux versions?

Thanks Sergy for squashing the pen spacing bug so quickly...

Link to comment
Share on other sites

  • Advanced Member

I can confirm that liveclay tablet problem could be related to ubuntu 11.04.

I have tried with my old system(on ubuntu 10.10)and it works perfectly(I have tried with wacom intuos 4 but as previously I found the bug on this system I think tablet model doesn't matter).

Anyway,thanks again.

Link to comment
Share on other sites

I can confirm that liveclay tablet problem could be related to ubuntu 11.04.

Maybe it is driver related? Today I have installed Linux Ubuntu 11.04 and Linux Ubuntu 11.10. My tablet Wacom Graphire 4 with default drivers worked on both of them w/o the LiveClay bug. I have compiled test executable for you. Please download it from here:

http://www.3d-coat.com/~sergyi/3d-coat-64-liveclay-tablet-bug.tar.gz

Please run it. Select default voxel sphere. Go to surface mode. Select LiveClay brush and do just one stroke (with the bug) on your tablet. Close 3D-Coat, and send Log.txt to me.

Link to comment
Share on other sites

ORIGINAL_AREA_pixel_data != CLONED_AREA_pixel_data

you do not get the original photo-data, but a compressed/blurred copy of the original photo now.

I strongly agree with this statement. I have good idea how to improve clone & smudge tools without using rendered projection. In so way smudge/clone tool quality will not depend on screen resolution.

Link to comment
Share on other sites

  • Advanced Member

SERGYI, trying to execute the file you linked gives me a segmentation error(I have tried enabling execute on properties and also doing from the terminal "sudo chmod +x namefile",the file is in the 3d-Coat-V3-7 directory,where other executable files are)

Link to comment
Share on other sites

  • Member

Happy new year all,

I tried 3D-Coat ver 3.7.02C for mac.

LC-CutOff-brush causes a crash.

I guess that 3d-coat recursive-called itself and killed by stack-overflow when the application build an AABB tree up.

the debugger told me this:

Program received signal EXC_BAD_ACCESS, Could not access memory.

Reason: KERN_PROTECTION_FAILURE at address: 0x00007fff5f3fffec

0x0000000100725588 in bmAABBNode::Init ()

(gdb) bt

#0 0x0000000100725588 in bmAABBNode::Init ()

#1 0x00000001007259d8 in bmAABBNode::Init ()

#2 0x00000001007259fd in bmAABBNode::Init ()

#3 0x00000001007259fd in bmAABBNode::Init ()

** sniff **

#30792 0x00000001007259fd in bmAABBNode::Init ()

#30793 0x00000001007259d8 in bmAABBNode::Init ()

#30794 0x00000001007259d8 in bmAABBNode::Init ()

#30795 0x00000001007259d8 in bmAABBNode::Init ()

#30796 0x00000001007259d8 in bmAABBNode::Init ()

#30797 0x00000001007259d8 in bmAABBNode::Init ()

#30798 0x00000001007259d8 in bmAABBNode::Init ()

#30799 0x00000001007259d8 in bmAABBNode::Init ()

#30800 0x00000001007259d8 in bmAABBNode::Init ()

#30801 0x00000001007259d8 in bmAABBNode::Init ()

#30802 0x00000001007259d8 in bmAABBNode::Init ()

#30803 0x000000010072736c in bool_struct::do_boolean ()

#30804 0x000000010072f14e in BasicMesh::boolean ()

#30805 0x00000001006e6cf7 in VolumeObject::SurfaceRectCut2 ()

#30806 0x00000001006e5995 in DoRectCut ()

#30807 0x00000001006139d1 in MakeSymmOpVO ()

#30808 0x000000010070c43b in CutterExt::OnRectSelectionEnd ()

#30809 0x00000001006b5438 in VoxelSculptTool::OnRectSelectionEnd ()

#30810 0x00000001001f7799 in BasicTool::btOnLMB_Up ()

#30811 0x0000000100420d8b in SceneWidget::OnButtonUp ()

#30812 0x00000001000df67d in comms::cWidgets::HandleInputEvents ()

#30813 0x00000001005f9e12 in comms::cMain_OnRender ()

#30814 0x000000010007ad5a in cMacMain_OneIterationOfMessageLoop ()

#30815 0x000000010007a28f in main ()

(gdb) info reg

rax 0x1745afa20 6247086624

rbx 0x7852 30802

rcx 0x7852 30802

rdx 0x0 0

rsi 0x179281000 6327635968

rdi 0x1745afa58 6247086680

rbp 0x7fff5f4000e0 0x7fff5f4000e0

rsp 0x7fff5f3fffe0 0x7fff5f3fffe0

r8 0xb 11

r9 0x1eaa40 2009664

r10 0x0 0

r11 0x1e9760 2004832

r12 0xffffdc00 4294958080

r13 0x0 0

r14 0x1745afa90 6247086736

r15 0xf4240 1000000

rip 0x100725588 0x100725588 <bmAABBNode::Init(bmAB*, int, int)+20>

eflags 0x10202 66050

cs 0x2b 43

ss 0x0 0

ds 0x0 0

es 0x0 0

fs 0x23 35

gs 0xf 15

(gdb)

Link to comment
Share on other sites

  • Advanced Member

Please do me a favore and check the latest mac build.

1. Merge the 'creature' mesh as voxels at ~700 k

2. drop to surf mode

3. Ask for decimation

Full of spikes and holes here and there

Now try it under 3.7.01B

It works perfectly.

post-2454-0-12236900-1325756719_thumb.jp

Link to comment
Share on other sites

  • Advanced Member

Another one

Merge 12edron, without voxelizing. Merge a second and a third.

Move them to the same layer. Booleans. Great.

Export it and check to see.

Using 3.7.01b everything is OK. Booleans rocking

Using the last build, vertices not welded !!

Please check these simple tests and let me know if have to be reported as bugs.

My problem with mantis bug reporting is that bugs are fixing but they are coming back on next build. Sometimes.

Anyway the current Mac build isn't stable, I already posted it some days ago but no luck.

Link to comment
Share on other sites

  • Member

I strongly agree with this statement. I have good idea how to improve clone & smudge tools without using rendered projection. In so way smudge/clone tool quality will not depend on screen resolution.

Is there any hope of also improving "normal" projection painting? It basically suffers from the same problem.

Link to comment
Share on other sites

  • Member

I'm afraid I have also seen this happen many times when I turn off symmetry. All I've been able to do is select and delete the extra polys and then turn symmetry back on.

I turned symmetry tool off/on at once or ... twice.

Let's take a look this figure.

Some blue points that are used in 'Points/Faces' tool still leaved themselves on symmetry's another side.

This is something wrong, isn't that?

Link to comment
Share on other sites

Updated to 3.7.03 (Win)

Changes:

- painting over voxel models. Just switch to paint room. If you need to turn it off - use View->Show voxels in paint room

- better symmetry support in Paint room (there was problems with Points & faces)

- better colors correspondence when baking voxels->paint room

- new brushes in LC tools - build, extrude, updated general brush, flattern.

  • Like 1
Link to comment
Share on other sites

  • Advanced Member

Ok just tried this and its Very very good.Great speed ,Everything works as it should and fast too .

You've outdone yourself this time Andrew .This is A LOT better than zbrush's polypaint system by far .Zbrush's polypaint layer system is very buggy with a lot less features.

I cant believe how quickly you implemented this too.You are one ninja coder.

One suggestion ( although it may already be possible ) .I can increase the resolution using LC with a zero depth setting but the tesselation depends on brush size and trying to cover a large area with a small brush takes a long time and the tesselation is uneven .Is there any way to mask off an area and increase the resolution ?

Also did a quick test exporting the model and to try and bake it in Xnormal but it seems that the vertex colours dont get exported with the mesh .Is there some way I can export a vertex coloured obj mesh?

Thanks so much for all the hard work.

Link to comment
Share on other sites

Guest
This topic is now closed to further replies.
 Share

×
×
  • Create New...