In 2019, FlippedNormals published a video called "What You Get Wrong About Smart Materials in Substance Painter". It is 13 minutes long. More than 200,000 people have watched it.
Their point: beginners lean too hard on smart materials and stop looking at reference.
I make smart materials for a living. There are 29 volumes of them in my shop. I have also read several thousand student files that used them badly.
So I should disagree with that video. I only disagree with half of it — and most people took away the wrong half.
They are right about the symptom
I see it every week. A student drops a smart material on a mesh. At 400% zoom it looks great, and the asset is still dead.
Everything on it is the same age. The edge wear runs along edges no hand ever touched. Dirt sits in corners a hand would have wiped clean.
Nothing has a history, because nobody decided what the history was. That part of the video is correct.
But they never said "stop using them"
Watch it again. The complaint is about relying on the tool, not about the tool.
That difference got lost in the comment section. What spread was "smart materials are for beginners". That version has cost people more time than the habit it was meant to fix.
A smart material is not a texture
A smart material is a saved layer stack. Its masks are driven by your mesh maps: curvature, ambient occlusion, thickness, position.
When you drop one on a model, you have not applied a surface. You have applied a guess about how this kind of surface ages.
That guess is written as generators, and you are meant to change it. Treating it as a finished texture is the whole mistake.
The setting that ruins it is scale
One setting fixes most student files that look like painted plastic. Not roughness, not generator opacity. Scale.
I author at 4096 × 4096. Every scratch and every grain is sized for the object I built it on.
Put that stack on a model with a different texel density and the detail is the wrong size. Too fine, and metal looks like cloth. Too coarse, and the prop looks like a toy.
The material is not fake. The material is the wrong size.
So do this first:
- Open the fill layer and fix the projection scale.
- Look at the asset from the distance your camera will really sit at.
- Stop judging it at 400% zoom.
That takes eleven seconds. It fixes a large share of "smart materials look fake".
The bake underneath matters just as much. I wrote about that in why mesh maps are not optional.
So here is my one real disagreement with the video: the order is wrong. Brushes and hand-made variation are the expensive fixes. Two cheap ones come first.
Where hand-painting everything falls apart
Say you follow the advice fully. You build every material by hand, for every asset.
That works for one hero prop in a portfolio. It does not work on a deadline.
My studio is four people. A prop set is not one steel pot. It is thirty objects that must look like they came from the same factory.
Hand-painting thirty surfaces gives you less consistency, not more. And it takes two weeks.
The answer is in the middle. Start from a base you did not have to build. Then spend hand-work on the ten per cent that carries the story.
Usually that is one custom mask, for the area the generators cannot reach. Paint it after the stack is down, not instead of it.
How to judge a pack before you buy it
That puts the pressure back on people like me. A smart material you cannot change is a bad product.
Four questions. Ask them about anything you download, including mine.
- Are the generators still there, or has everything been flattened into one layer?
- Is the aging a parameter you control? In 30 Copper Smart Materials VOL 16, oxidation is one number. A two-year roof and a forty-year roof should be your choice.
- Can you take it apart? Delete the top three layers. If nothing usable is left, it was never a stack.
- Does the documentation say what scale it was made at?
Number four is the one almost nobody answers. It is also the cheapest to fix — a line of text, not a feature.
The part of the video I would keep
None of this helps an artist who has not decided what the object is.
No stack will tell you that this rifle belonged to someone who cleaned it daily and dropped it once. That decision comes before any tool.
If you have not made it, no generator and no brush will save you. That is the part still true seven years later.
The part I would drop is the part they never said — that the tool is the problem. Thirty objects still have to agree with each other.
If you want the full workflow rather than the argument, that is what the Substance Painter Master Course covers.
