Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.

Messages - Sergey Danchenko

Pages: 1 2 3 [4] 5 6 ... 10
Had the same issue just yesterday. You have to delete contents of Substance_UE4_Plugin_4.14.0.19\Plugins\Runtime\Substance\Binaries and rebuild the engine to make it work. I believe this step should be added to the Readme.

Ah, I see. Maybe the scripting feature would be expanded in the near feature to include "on-bake" event and texture packing could be achieved through some Python script.

Substance DesignerSubstance Designer - Feature Requests - Re: Weighting
 on: January 30, 2017, 07:33:26 am 
Hey michael_11, I'm pretty sure zurhoch meant something like this thread and not the game itself. Decal technique from Star Citizen is being discussed there from a technical standpoint.

It's on the wish list at the moment  ::)

You're welcome, @thephobos! Actually, it was 1 year and 7 months ago  ;D Time goes fast, huh.  ???

Substance community is indeed an incredible bunch of people. So many ideas, so many great works! And not to forget the Substance tech that is really empowering. I bet that many would be surprised if they knew that I started the SD around 2 months ago and approached the Pixel Processor node 2 weeks before this node came out.

Lately I've been looking inside graphs of nodes that come with SD and never seized to be amazed how they were actually built and how many ideas are tightly packed in them (not that I've complete understanding of what's going on in there, ha!  :P). Simply amazing.

Cool! Haven't thought yet of getting the mask from "solid" images using the edge detect. Will know now, thanks! Also it's great to hear that the node works well for you and I'm definetly happy to see some test cases. I've tried to organize the node in a way that will allow it to function correctly almost in any case, but there still can be some rough corners. If something doesn't work — just twiddle parameters or post here so I can take a look.

Tip of the day: if your're not happy with the luminance pattern created by the node, tune the random seed to get different result.


May be a separate node as well. With random luminance mask it should be possible to process the cells in a way that will get them gradients with a slope in arbitrary direction and employ some controls for the contrast/slope degree. It could be a separate node with an input in form of a luminance mask that will serve kinda like an ID map to help the node distniguish between separate cells.

That is great! The only thing I would miss from such a node is the ability to create sloped versions of the cells, i.e. cells with random gradients.  ::)

A particular example, if it is possible, would be a great way to troubleshoot it. Do you have any graph to share that will demonstrate your issues?

Would be great indeed  :) There is such a tool in Substance Painter already, but not in Designer. Would be a great new feature for SD 6, I believe.

One thing that you may have been missing is that Blend node by default is set to Relative to Input and will inherit output size of a node plugged into the Background slot. If you're using a lot of bitmap nodes in your graph, you have to be particularly careful in that regard. You may want to use a intermediary/dummy node like Levels between the Bitmap node and Blend node with output size mode set to Relative to Parent, so it will change dynamically in respect to your graph output size and will force the Blend node to do so as well.

Hey everyone,

I've finished the first iteration of Randomize Mask node and invite everyone interested to check it out and give it some tests (archive is attached to this post below or you can use this link to download it:;topic=5158.0;attach=23460). The graph itself is heavily commented, so don't hesitate to look inside and ask questions. ::)

Some usage notes:

1. In first version default values for input parameters are set in favour of processing masks in resolutions of 1024x1024 or higher. If you're testing the node in lower resolutions and experience issues, you may want to disable all Optimization input parameters (set them to 0).

2. There are two graphs in the package, use the main_assemble to instance it into your test graph.

3. Use a correct mask for the input. Something like the output of Tile Sampler node is a perfect example of a usable mask.

4. All input parameters are also commented, so hover your mouse cursor over their titles to see a tooltip that will explain how to use them.

5. Let me know what you think and feel free to ask questions, as this node is quite complex with a lot of steps inside and it sometimes it can be difficult to understand what's going on  ::).

6. If you experience some strange behaviour of the node, please report it here as I'm interested in improving it.

7. You can contact me directly about anything related to this node by using the icons below my avatar to the left.




I'm working on a package with detailed explanations about everything that happens inside the graph. It's quite a lot of comments, but it is almost finished. See picture attached for a sneak peak (it's about a half of what is inside, as there also a number of external functions that will be commented as well).

Also I've come up with an additional functionality for this graph - now it can output a Random Opacity Mask as well as Random Luminance Mask. It is currently dependent on the generated luminance mask, but this will be a subject of discussion once the graph is released. Picture of the Random Opacity Mask output is attached.


Awesome news! It's just today that I was going to inquire about normal map compression issues in UE4, and here it is - an updated integration!  :) Improved Normal map compression is the most welcome change.

Am I correct to assume that currently it's still not possible to output textures in UE4 without any kind of compression? 28 months ago it was in a to-do list — is there any chance to have a bump in priority for this?,1372.15.html#msg9241

Let me explain why I'm seeing this as an important feature. Substance as well as Unreal Engine 4 get some presence outside of video games industry. Unreal Engine 4 becomes an important tool in architectural visualization and product presentation, not naming some other applications.

Unlike games, where cutting-edge visuals most of the times had to be balanced against performance, architecture and interactive product presentations more likely to be pushed for the quality, not speed. Having an ability to use uncompressed resources is essential — when you're working on a small scene that would be showcased from a very close distance, any unnecessary artefact could become an issue very easily.

There could be some workarounds for this, but the point is to allow the use of pure Substances whenever the user wants them, including the high-end content. Higher adoption of Substance tech across all industries is a key for the success, isn't it?  :D

My point in short: if it is possible, let the user decide where his priorities are, be it quality or speed. If it's a "dangerous" option, hide it, name it experimental, but make it available for ones that really want it.

Pages: 1 2 3 [4] 5 6 ... 10