There is no consistency across export dialogues, Painter has presets for exporting to Unity and other software with the right material conversion (e.g. Metalness maps with smoothness in the alpha channel) - currently these export options are missing from Alchemist and Designer when you want to export to bitmaps or individual textures.

It would be useful to have these export options for when you are exporting simply tiling textures (like for a terrain or large surface in a game engine).

My current workaround is to have a template Unity 5 Substance #Painter# project that I import the archived Designer or Alchemist material into, apply it to a generic 2d Plane model and then export it with the Unity preset. Why should I need so many extra steps to get my textures right?

I've encountered a number of graphs that use a Transform node after a Clouds (1/2/3) node to "zoom in" the clouds to achieve specific effects. Unfortunately, this creates the side effect of breaking tiling and requires other nodes to try to fix that.

Instead, it would be very helpful if the cloud nodes could have additional lower scales - the default Scale of 1 (and the minimum) is still fairly "medium density" in terms of detail. I'd like to see the same cloud effects, except 2x and 4x "zoomed in". In other words, the equivalent of a Scale 0.5 and Scale 0.25.

Something that would really speed up workflow is immediate typing focus when selecting (clicking) nodes that have a single type-in control (especially function nodes, like constant values).

In addition to selecting, it should also happen immediately after creating these nodes. But I think newly created nodes are selected immediately anyway.

The only issue I can see this causing is interference with hot-keys while such a node is selected. But as long as hotkeys are typical (eg, ctrl + key), it shouldn't be a problem. No different than trying to use a hotkey immediately after typing something into a typical control. The user can just press enter to get back to the normal state in either situation..

It'd be great to have an option to export the displaced mesh exactly the way it looks in the 3D viewport (or with translation for iRay view as those always differ).

Use Cases: A quicker path to a 3D print.
Especially when you're designing real world materials it would be great to have a tangible example when getting a render signed off on. Or just to quickly create prop pieces or stamping tools for specialty costumers.

Substance Designer Path:
1. Mesh goes into Substance Designer
2. Deformation is dialled in
3. Export to OBJ

Traditional Path:
1. Export Height Map
2. Apply as displacement map
3. Tweak height map with levels and curves
4. Export height map again because levels and curves wasn't enough
5. Test Render
6. Adjust Displacement Depth
7. Bake displacement to Mesh
8. Cook fancy dinner while waiting
9. Export Mesh to OBJ
10. Consume said dinner
11. Import OBJ to printer software
12. Print

We should have the ability to create a single node that can support both grayscale and color, rather than needing to add "gray" vs "color" suffix or prefix onto each node that can support both.

This would be a simple internal detection process, and doesn't really require any advanced changes. We would still need to write both versions of our nodes, but with the difference that Designer would automatically employ whichever one is currently plugged in.

One method that would allow us to use a single graph would be a checkbox on input and output nodes that specify they are only to be active when grayscale is active, or only to be active when color is active. Or we could just create two different graphs in a single substance and specify similar options on the graph settings, to let Designer know which to use in the specific circumstance.

The current search algorithm seems like a basic substring search. Where your query must contain a single continuous substring of the name you want to match. For example "Flood" will match all the various Flood Fill nodes, but to match a specific node such as "Flood Fill to Position" you have to type "Flood Fill to P".

In many modern software applications a fuzzy searching algorithm is used. Where search results are sorted by the "distance" between two strings where the distance is determined by the number of characters that must be added/deleted/substituted to make a query string match a target string. For example "Flood P" requires the addition of 15 characters - "Fill to osition" - to match "Flood Fill to Position" perfectly.

There are many benefits to this approach. Since the distance is numerical a queries results can be sorted based on how closely they match a given string. This not only allows for even more fragmented queries to match targets more accurately, such as allowing a user to type initials "fftp" or short fragments "fltopos" to accurately match a desired target of "Flood Fill to Position". Depending on how completely an implementation attempts to transform the query string in to a target string these algorithms can even handle spelling errors - though in the real time context of search menu being this thorough can have performance issues.

Beyond the improved UX experience of being able to use shorthand queries. There's also how it results in a much more beginner friendly experience. You know you want a "flood fill" node but can't quite remember what it was called, was it position or displacement or warp? Just type enough to match "flood fill" such as "flood", read the query results and refine your query with the results as they update, e.g. "flood pos".

A popular MIT licensed implementation of the Levenshtein Distance search is FuzzyWuzzyMIT. Which has been ported to several languages.

Python: https://github.com/graingert/fuzzywuzzymit
C++: https://github.com/martinariel/fuzzywuzzy

I think there would be a huge benefit in having a node that links to other software, specifically ZBrush and Painter.

For ZBrush, the node would connect to a ZBrush session and allow for users to generate height information from the session. It can also send Designer's height data to allow for sculpting within ZBrush. This would be great for easily going back and forth between the program, while organizing the sessions for their various purposes like establishing the base forms, modifying shapes, or final touchups at the very end. I see many studios (Sony Santa Monica and Blizzard, as an example) either start with a sculpted height map or start in Designer and eventually go into ZBrush and back to Designer. A node like this would make it efficient and allow for the hand-work that Designer currently lacks, while using software that many professionals have.

For Painter, I imagine that each Designer link node would correspond to a folder in the linked Painter file, allowing a user to make a layer stack within Painter, while resulting in a single Designer node. This would make touchup painting for color, normal, roughness, etc maps very convenient.

Hello,
many tutorials show already how to create an individually designed surface, even with organic realistic themes like for example a meadow with mud and other details.
But if you use it in substance painter (or any other applicaton) to fill a large area, you can always see the pattern.
All the realism in detail is nothing worth if you see it from far distance (like a meadow in front of your feet that reaches to the top of a far mountain).
Because if you try to fill a large area, you need to stagger and replicate. But all your possibilities are currently limited then to positioning the tiles, or color or displacement instead of generating real variations of those tiles with access to creative parameters.

The stupidity is, that there ARE always randomization parameters available when published from SD, which could make each "tile" looking individual, but I did not see a function or application, that makes use of them, when distributing the tiles.

So I'm completely missing the ability, not only to create one identical tile as output which is replicated, but also to randomize the output parameters for each single individual tile within a customizable range of any wished parameter. Only color, displacement, size and position is not enough, it's still recognizable as the same tile.

This would give the chance, that there is no more artificial pattern visible, which makes pictures often looking so cheap generic.