Default painting assets in GIMP currently nudge users toward specific styles without making that influence visible. I’m proposing a reorganisation of 3.4 based on the Bristle Atom (treating brushes as geometry, not visual outcomes). Seven overlapping categories are reduced to three (Basic, Artistic, Legacy). Texture moves from brush types to dynamics. GIH brushes can already support 3+ dimensions using existing features (no code changes required). Full article with rationale, performance data, and migration path below. Looking for feedback on the framework before I prepare a demonstration mini-set.
What this article is: A philosophical proposal plus a structural reorganisation for GIMP’s default painting assets, intended for the 3.4 release cycle.
What this article is not: A finalised technical specification, an official GIMP roadmap item, or a promise of features already merged into the codebase.
Part One: The Problem | Prêt-à-Porter Style
I am reviewing GIMP’s default assets in anticipation of the 3.4 release. This article offers both a philosophical and a structural proposal for improving GIMP for painting tasks, especially as more users seek sustainable alternatives to subscription-based software.
Two Ways of Using GIMP (With a Spectrum Between)
Over the years, I have glimpsed two main ways of using GIMP for painting, best understood as ends of a spectrum, not a strict binary.
1. A more “traditional” approach, Many graphic artists (painters, illustrators, concept artists, game illustrators) use conventional tools: paintbrush, pencil, airbrush, ink, smudge, eraser… often without tool presets. These artists rely on standard options (opacity, size, basic pressure response) to resolve their style and tasks. For them, GIMP is excellent precisely because it doesn’t require a large preset library. Their style is already well-formed.
2. A highly customizable approach, Other artists build deep custom workflows using painting dynamics, tool presets, and saved brushes. GIMP supports this, too.
The important point is that GIMP adapts well to both ends of this spectrum. But the default assets shipped with GIMP inevitably nudge users toward one side or the other. My argument is that they should nudge toward neutrality.
Style, Software, and the Hidden Influence of Presets
When I talk with digital artists, it’s evident that each painter has their own style (or unwittingly adopts one). Every painting application’s UI creates conditions, often unintentionally, that superimpose its “inner style” onto ours. This is most common in preset-heavy apps, where the application is viewed less as a tool than as a library of ready-made looks.
Here is a metaphor I find useful:
-
Every software UI is like a language — it shapes what we can express and how.
-
Every language is moulded by its environment and its assumptions.
-
Our artistic style is moulded, often unknowingly, by the UI we use daily.
-
Different UIs imprint different stylistic tendencies.
-
When an app’s capabilities are tied to a specific style that is not our own, everything we make becomes influenced by it.
When you use a tool preset made by someone else, you are borrowing not just a set of parameters, but a way of painting. That preset reflects the technique, habits, and aesthetic of its designer. Have you ever wondered why so much mainstream digital painting looks so similar? This is one reason.
Tool Presets as Prêt-à-Porter
Since GIMP introduced tool presets (they were already present in 2.6+), it has become possible to record and share our painting styles with others. That is powerful.
But the French term prêt-à-porter (ready-to-wear) from fashion is instructive. A ready-to-wear garment fits a generic body. It carries the designer’s silhouette, fabric choice, and cultural moment. You can alter it, but only so much. The same is true of tool presets. They arrive with embedded decisions about texture, opacity behaviour, blending modes, and pressure response, all of which reflect the preset designer’s stylistic bias.
Many users don’t realise this. Even fewer know how to customise presets deeply, because the interfaces for doing so are often complex or hidden.
This is why default presets in a painting application like GIMP should be as classical and basic as possible. They should be neutral triggers for the artist’s own experience — to borrow Brian Eno’s phrasing, they should be triggers for experiences, not finished objects carrying someone else’s signature.
What Neutral Defaults Mean
For GIMP 3.4, I propose four principles for default painting assets:
-
Behavioral neutrality – Default brushes should respond predictably: opacity → pressure, size → pressure. Avoid hardcoded texture, color dynamics, or wetness that mimic a specific medium (charcoal, wet oil, airbrush splatter). Those are wonderful options — but they belong in custom presets, not the default set.
-
Discoverable customisation – Right-clicking a preset should offer “Edit dynamics settings” in one click. No buried menus. The distance from “using a preset” to “making it your own” must be as short as possible.
-
Example-based documentation – Ship 2–3 classical presets (smooth round, flat, textured grain). Then provide a short tutorial showing how to derive custom presets from them. Teach fishing; don’t just give fish.
-
Minimal, not insufficient – A small set of well-chosen, neutral defaults is better than 50 presets that collectively push a specific contemporary style. Less can be more, but “less” must still be complete enough for basic painting tasks.
A Note on What Other Apps Do
Krita’s brush engines prioritise deep customisation; MyPaint’s defaults are famously minimal; Procreate ships with a small, carefully curated set that balances neutrality with character. GIMP can learn from all three without copying any. The Bristle Atom proposal below is GIMP-specific, drawing on the application’s existing strengths (GIH brushes, dynamics system) rather than importing paradigms from elsewhere.
This philosophical foundation raises a practical question: How do we reorganise GIMP’s brush engine to make such neutral defaults possible? The answer is the Bristle Atom.
Part Two: The Solution | The Bristle Atom
The fundamental unit of any digital painting tool is the Bristle Atom. Based on my analysis of the default brushes on GIMP, approximately 85–90% of GIMP’s brush library should be defined by internal spatial structure (geometry), not visual outcome (e.g., “Texture” or “Sketch”). (A full audit is available in the supplementary materials upon request.)
What Is a Bristle Atom?
Morphological Definition: A brush is a collection of bristles. Irregular or textured tips are simply specific spatial arrangements of bristles. There is no need for a separate ontological category.
Mathematical Limits: Geometric primitives (circle, square, ellipse) represent the densest, most uniform bristle clusters. All other brushes are variations (sparse, scattered, asymmetric).
Decal/Stamp Distinction: Tools that do not simulate fluid or pigment interaction (e.g., icons, complex shapes, decorative stamps) are Stamps/Decals. They belong outside the bristle simulation logic entirely. This is not a value judgment; it is a structural one.
Behaviour vs Geometry
Previous revisions of GIMP’s brush system lacked a clear distinction between what a brush is (its geometry) and how it acts (its behaviour).
Dual-State Functionality | Every Bristle Atom supports two modes:
-
Continuous (Stroke) | High-frequency sampling for smooth painting
-
Discrete (Pattern/Texture) | Low-frequency sampling for stamped effects
Universal Texture Dynamics | Texture should not be a brush type. Texture results from usage parameters: opacity jitter, size jitter, spacing, rotation, and pressure response. A dedicated “Texture” category is redundant. Replace it with dynamics that generate procedural grain across any brush.
For example, a charcoal grain can be achieved by combining: Size Jitter (small range, high frequency) + Opacity Jitter (pressure-linked) + a grain pattern mapped via Rotation dynamics. No separate “Charcoal Brush” category is needed.
The Three-Category Hierarchy
Replace the current seven overlapping categories with three clean ones:
| Category | Contents | Purpose |
|---|---|---|
| 1. Basic (Structural Primitives) | Parametric brushes (.vbr). Mathematical bristle density limits (circles, squares, simple ellipses). | Entry-level users, masking, clean line art. |
| 2. Artistic (Bristle Atoms) | Image brushes (.gbr) and smart containers (.gih / GIH2) defined by spatial bristle arrangements: fine, coarse, flat, round, angled. | All serious painting. |
| 3. Legacy (Historical & Community Data) | Essential older brushes for backward compatibility. “Fun” brushes and static stamps. | Preserve community work without cluttering the active set. |
What disappears as separate categories? “Media,” “Sketch,” “Splatters,” and “Texture” — all folded into either Artistic brushes (geometry) or Dynamics (behaviour).
Backward Compatibility
Existing GIMP 2.10.x and 3.0 users’ custom brushes will remain fully functional. Upon first launch of GIMP 3.4, a one-time migration assistant will classify existing .gbr and .gih brushes into the new three categories based on heuristic analysis (file size, dimensions, metadata). Users can override these classifications manually. No brushes are deleted; the Legacy category preserves everything that doesn’t fit cleanly into Basic or Artistic.
Simplification Targets
| Resource Type | Current (GIMP 3.2.x) | Proposed (GIMP 3.4) | Goal |
|---|---|---|---|
| Brushes | 55 | ≤40 | Move “Texture” logic to Dynamics, not brush tips |
| Patterns | 66 | ~20 | Focus on versatile, timeless patterns (tiling basics, grains, papers) |
| Dynamics | 18 | 10–15 | Prioritise tilt, velocity, opacity, size, rotation |
| Tool Presets | 39 | 5–10 | Onboarding modules (see note below) |
A note on those Tool Presets: While the article argues for neutral defaults, beginners do benefit from curated starting points. The proposed presets (Oil, Charcoal, Watercolour, Pencil, Airbrush) are therefore optional teaching tools, not default brushes. They live in a separate “Starter Presets” folder. The default brush set itself remains neutral. This resolves the apparent contradiction between philosophical purity and pedagogical usefulness.
GIH2: The Technical Carrier
Status note: The techniques described below use existing GIMP GIH capabilities in novel ways. No code changes are required; this is a creative reuse of current features. I have working examples available upon request.
This approach builds on research into GIMP brushes dating back to 2008. I studied how to improve the brush experience using multi-dimensional GIH brushes — the kind many artists already use, as seen in GIMP’s default set (e.g., the Acrylics series).
What does GIH2 mean here?
Normally, GIH brushes are used with only 1 dimension (e.g., random frame selection, as in the Acrylics series). My research shows it is possible to use at least 3 dimensions with selected behaviours (pressure, incremental, random, velocity, mainly) to better mimic real brushes.
By “3D logic gates” I mean conditional relationships between dimensions, for example: an array(2,2,2) where the 2 ranks for pressure, 2 ranks for incremental, and 2 ranks for random.
A GIH2 configuration would require:
-
8-dabs and 16-dabs configurations for multi-dimensional bristle simulation
-
3D logic gates to control bristle behaviour across dimensions
Performance considerations: Early tests on a 2018 laptop (8GB RAM, integrated graphics) showed no perceptible lag with 8-dab configurations at 256px. 16-dab configurations required reduction to 128px for smooth performance. Your mileage may vary.
Professional Standards and Onboarding
For high-profile fine artists – They need pigment and bristle physics, not software expertise. GIH2 hierarchical layouts ensure consistency.
For entry-level users – Tool Presets bundle Bristle Atom brushes with dynamics to mimic traditional media. Beginners get quality via presets. Professionals bypass presets and work with modular components directly.
Real-world testing – Artist friends Mozart Couto and Gustavo Deveze have tested these brushes and appreciated the methodology’s possibilities.¹ The advantages align well with the goal of a high-profile brush set for GIMP 3.4.
¹ Acknowledgements: Mozart Couto and Gustavo Deveze provided feedback on early prototypes. Their insights helped refine the Bristle Atom concept, though they bear no responsibility for any errors in this proposal.
Part Three: Summary — Current vs. Proposed
| Aspect | Current (GIMP 3.2.x) | Proposed (GIMP 3.4) |
|---|---|---|
| Philosophical basis | Implicit, mixed | Neutral defaults, no prêt-à-porter style |
| Organizing principle | Visual outcome (Texture, Sketch, Media) | Morphology (Bristle Atom geometry) |
| Number of categories | 7 (overlapping, redundant) | 3 (Basic, Artistic, Legacy) |
| Texture | Separate brush category | Generated via Dynamics (Universal Texture) |
| Stamps / Decals | Mixed with painting brushes | Isolated in Legacy or optional pack |
| GIH usage | 1 dimension, random behavior | 3+ dimensions with logic gates (GIH2) |
| Onboarding | Implicit, confusing | 5–10 Tool Presets as explicit optional modules |
| Default brush count | ~55 | ≤40 |
| Default dynamics | 18 | 10–15 (prioritized) |
| Default patterns | 66 | ~20 (timeless basics) |
| Default tool presets | 39 | 5–10 (optional modules) |
| Backward compatibility | N/A | One-time migration assistant, no data loss |
Conclusion: From Philosophy to Engine
GIMP is not a commercial app like Photoshop, but it could be a good alternative for painters and illustrators. That is its strength. In an era where SaaS subscriptions are a real problem for many users, GIMP offers a genuine alternative — if we are honest about what makes it different.
The philosophical argument: Default assets should not impose someone else’s style. They should be neutral triggers for the artist’s own experience. No prêt-à-porter.
The structural argument: The Bristle Atom reorganises the brush engine around geometry, not visual outcomes. Three categories. Texture as behaviour. GIH2 with 3+ dimensions.
These two arguments are not separate. The philosophy demands the structure. The structure serves the philosophy.
Next Steps
I propose we first discuss the big-picture framework, The Bristle Atom, as an organising principle. After reaching consensus, I can prepare a mini-set demonstrating:
-
Standalone Bristle Atom brushes (Basic + Artistic categories)
-
Dedicated Dynamics (10–15) for texture generation
-
Optional Tool Presets (5–10) for onboarding, stored separately from neutral defaults
This mini-set would help intermediate and entry-level users adopt the new resource set while giving advanced users a clean, modular foundation to build upon.
Art begins where technique ends. Let GIMP’s defaults be the technique. Let the artist bring the art.