[3DCoat 2024 development thread]

On 4/30/2024 at 10:27 AM, Elemeno said:

 

 

Seems to be working as expected? The mask uses grayscale input. The green you selected had a value (lightness) of .76, and partially revealed the albedo layer once you began applying it. If you want to completely unmask the albedo layer, the value needs to be 1. Just like any other software.

If you paint a white mask with a value of .76 in substance painter the result is exactly the same. Apologies if I misunderstood the issue at hand. Let me know if you need assistance.

[guys any help with this ?]

4 hours ago, Elemeno said:

 

this is how they're being outputted 

Looks like inner cage is not deep enough. You must adjust outer and inner so that you capture the desired threshold. If this doesn'r work, please send project file and I can take a look.

[3DCoat 2024 development thread]

49 minutes ago, Elemeno said:

can you remove modelling room and retopo room and have all meshes inside sculpt room , and have modelling and retopology tools as a separate tool , with that the right click would be a pie menu with the selection type at the top like vertex,edge,face then below would be the tool you wish to use , the right side could have the current right click menu where you can relax,clear points etc and left side left clear for quickmenus maybe?

the mesh could be saved as P for polygroups ... and once youve modelled something you can then voxelize it or go into surface mode where SUBD would work nicely

having these in one room just saves the hassle and tidies up and simplifies things   

Yeah I'm definitely not a fan of the 4 different types of meshes. Problem is how can one organize the UI in a way where the tools are not scattered all over the place. Just looking in the modeling room, the left tool panel is filled to the brim. Combining that with sculpt tools may be problematic.

Maybe keep the separate rooms, but make it easier to manipulate meshes between rooms?

[3DCoat 2024 development thread]

22 hours ago, animk said:

It's just a generic_bust_male model from 3d coat, autoretopo it and bake 16k ppp.  Then bake either AO or cavity.  It exists as soon as open the bake option window.  I tried it on linux mint it still crashed.
It shows in the terminal
malloc(): unaligned tcache chunk detected
"Aborted (core dumped)

Interestingly, I also crash but get a different error:

malloc(): smallbin double linked list corrupted

What's funny is that this doesn't happen with any other specified texture resolution. I tried on my radeon card (mesa) and nvidia card and received crashes in both instances . It appears to be an issue with the code.

I also tried baking 16k occlusion on the sample robot, which also resulted in a crash. Seems no matter what you cannot bake 16k occlusion/cavity maps.

I did find somewhat of a workaround (more like a temporary bandaid) which is to set the texture resolution to 8k, bake whatever maps you need, and then switch back to 16k and continue texturing albedo etc.

I'm going to forward these posts to Sergii Kryzhanovskyi through email.

[3DCoat 2024 development thread]

4 hours ago, animk said:

Ubuntu 20.04.6 LTS

Baking 16k AO or cavity map crashes 3d coat.  Exit instantly as opening the baking window

my machine: Kubuntu, RTX 4090, 64 GB memory

Would you mind sending me the prpject file? Unable to recreate issue on my end. 20.04.6 Pop_OS, 3090

[3DCoat 2024 development thread]

Hello, Just wanted to express my gratitude because this is such a great update. Thank you sincerely to all of the devs. The new masking system is so much more intuitive, and I can safely tuck away Substance Painter forever now :)) Thank you guys! Keep up the great work.

[3DCOAT and Huion KAMVAS 13 Linux problem!]

Hi! Just solved the problem today with a buddy so I thought I would drop the solution here. You have to run it as a terminal and make sure your COAT_MOUSE is set to your Huion mouse, while COAT_ERASER is your actual physical mouse, and your COAT_PEN is your stylus. Mine looks something like this