Every brush pack listing leads with a number. 120 brushes. Over 300 textures. And every artist who has bought a few knows the number predicts almost nothing — you install 120 brushes, use four, and never open the folder again.
We build a brush engine, so we spend an unreasonable amount of time inside this. Here is what actually separates a pack you keep from a pack you delete.
1. Spacing, not shape
The single most common flaw in a bad brush is spacing set too wide.
A stamp brush works by stamping one image repeatedly along your stroke. Spacing is how far apart, as a fraction of the brush diameter. At 0.05 the stamps overlap so heavily the stroke reads as continuous ink. At 0.4 you can see the individual stamps, and the stroke reads as a dotted line pretending to be a stroke.
Wide spacing is tempting for the author because it renders faster and looks crisper in a thumbnail. It is the reason a brush looks beautiful in the preview and terrible at the size you actually draw.
How to test it in ten seconds: draw a slow, tight curve. Not a straight line — a curve, slowly. Wide spacing shows up on the inside of a bend, where the stamps splay apart into visible scallops. If it looks like a caterpillar, it is not an inking brush.
2. Grain that stays still
Texture in a brush comes from a second image riding under the stamp — the grain, or tooth.
The question that matters is what the grain is anchored to. Anchored to the canvas, it behaves like paper: the texture stays put and your stroke reveals it, exactly like dragging a pencil across a sheet. Anchored to the stamp, it travels with the brush, and repeated strokes over the same area produce a smeared, plasticky look that no amount of layering fixes.
Canvas-anchored is almost always what you want, and it is what a good pack ships. If you drag a textured brush back and forth over one spot and the texture appears to move, that is the wrong anchoring, and it is the reason some expensive packs feel cheap.
3. Tilt that does something
Tilt is the most under-used signal in digital drawing, and the packs that use it well are immediately obvious.
A pencil laid on its side should broaden and soften. A chisel nib should fan. A dry brush should catch more tooth as the angle drops. If a pack advertises “natural media” and every brush draws identically whether the pen is upright or flat, the naturalness is a texture, not a behaviour.
Worth knowing: this is per brush, not per app. In our engine it is one
parameter — tiltShade — and a good pack sets it thoughtfully per brush
rather than globally. A technical pen should have it at zero. A 6B should have
it high.
4. Stabilisation you can turn off
Smoothing makes lines cleaner. It also makes them deader.
Stroke stabilisation works by lagging the rendered line behind the cursor and averaging out the wobble. A little is invaluable for inking. A lot removes the evidence that a person made the mark — which is fine for a logo and quietly disastrous for a sketch.
The tell of a thoughtful pack is that stabilisation varies between brushes in the same pack. Inking brushes heavily stabilised; sketching brushes left raw and nervous. A pack where every brush has the same smoothing was configured once and copied.
5. A size range that matches the job
Every brush has a usable range, and a good author declares it.
A fineliner authored for 1–32 px does something sensible across that whole span. Push it to 300 and the stamp resolution gives out — the edge goes soft and pixelly, because you are scaling a small texture up past its detail. A wash brush authored for 20–300 does the opposite: at 3 px it collapses into a smudge.
Packs that ship every brush with an identical 2–300 range have not been thought about. It costs the author nothing to set it and saves you the discovery that brush 7 only works in one narrow band.
6. Fewer brushes, arranged
This is the one that runs against every listing on every marketplace.
A pack of twelve brushes you can hold in your head beats a pack of a hundred you have to hunt through. The cost of a large pack is not disk space, it is the decision you make every time you pick a tool — and a decision between a hundred near-identical options is a tax you pay on every stroke.
What good large packs do is shelve their brushes: inking, shading,
texture, effects. What bad large packs do is dump 120 entries in one flat list
with names like Brush_final_2.
The two-minute test
When you install a new pack, before you commit to it:
- Slow tight curve — checks spacing.
- Scribble over one area repeatedly — checks grain anchoring.
- Same stroke upright, then with the pen laid flat — checks tilt.
- One fast stroke, one slow — checks whether velocity or stabilisation is doing anything.
- Smallest and largest size the brush claims — checks the range is honest.
Two minutes, and you will know whether it is a keeper before you have built a habit around it.
What this means for buying
The useful signal in a listing is not the count. It is whether the author talks about behaviour — pressure curves, grain, tilt, spacing — or only about quantity and look. Someone who set twenty-two parameters per brush will tell you so. Someone who exported a folder of PNGs will show you a mood board.
Ours are described by what they do. The free packs are there precisely so you can run the two-minute test without paying anything first — and if a brush fails one of the five, we would rather you found out on a free pack.