| 
|
-Another minimal test you can try yourself in the app header: the Transmission Distance (TD) test image.
+Another minimal test you can try yourself: the included Transmission Distance (TD) test image (`src/assets/tdTest.png`).
| TD Test |
|---|
@@ -44,8 +46,8 @@ Another minimal test you can try yourself in the app header: the Transmission Di
- 3D model export to binary STL or 3MF (Preview) suitable for multi-material slicers.
- Plain-text 3D print instructions that describe layer heights and exact layers where filament swaps are required.
- Copy-to-clipboard button for the print instructions (produces a clean, copyable plain-text plan).
-- **Filament calibration wizard** — Accurately determine Transmission Distance (TD) values through measured samples with confidence scoring.
-- **Advanced optimizer** — Simulated annealing and genetic algorithms find optimal filament ordering for complex multi-color prints.
+- **Filament calibration** — Accurately measure frontlit hiding distances from a camera-free printed wedge, with per-channel values and confidence scoring.
+- **Advanced optimizer** — Five deterministic effort tiers search for optimal filament ordering, from quick previews to exact base-order enumeration.
- **Region weighting** — Prioritize accuracy in specific image areas (center, edges) during auto-paint optimization.
## Notable implementation details
@@ -64,9 +66,9 @@ Another minimal test you can try yourself in the app header: the Transmission Di
1. Upload or drag an image into the preview area.
2. Switch to the **Auto-paint** tab in the 3D controls panel.
3. Click **Add Filament** and configure your filaments:
- - Use the **calibration wizard** (calibrate icon) to measure accurate TD values, or
- - Enter TD values manually.
-4. Enable **Enhanced color matching** for optimal results (auto-selects best algorithm).
+ - Use **Calibrate Filaments** to measure accurate hiding distances from a small printed wedge, or
+ - Enter hiding distances manually (the convert button accepts conventional backlit TD values).
+4. Enable **Enhanced color matching** for optimal results; start with the **Balanced** optimizer tier.
5. (Optional) Set **Region weighting** to Center or Edge to prioritize important areas.
6. Use the **layer-by-layer preview slider** to verify transitions.
7. Export via **Download STL** or **Download 3MF** and follow the print plan.
@@ -107,45 +109,51 @@ This is especially useful for complex multi-color prints where you need to under
## Auto-paint
-Auto-paint is an automated layer-generation mode that replaces the manual palette/swatch workflow. Instead of reducing an image to a fixed number of colors and manually tuning per-color slice heights, you define a set of **filaments** (each with a color and a Transmission Distance) and let the algorithm compute the optimal layer stack automatically.
+Auto-paint is an automated layer-generation mode that replaces the manual palette/swatch workflow. Instead of reducing an image to a fixed number of colors and manually tuning per-color slice heights, you define a set of **filaments** (each with a color and a hiding distance) and let the algorithm compute the optimal layer stack automatically.
### Core concepts
-- **Filaments**: Each filament has a hex color and a **Transmission Distance (TD)** value (in mm). TD describes how translucent the filament is — at a thickness equal to TD, only ~10% of light passes through. Dark/opaque filaments have low TD (e.g. 0.5 mm); light/translucent filaments have high TD (e.g. 6+ mm). When you add a filament without specifying a TD, Kromacut estimates one from the color's luminance and saturation.
-- **Beer-Lambert optical blending**: The algorithm simulates how light transmits through stacked filament layers using the Beer-Lambert law: `transmission = 10^(-thickness / TD)`. This physically models the color you see when printing thin semi-transparent layers on top of each other.
-- **Transition zones**: Each filament in the stack needs enough vertical space to visually transition from the color below it to its own pure color. The algorithm simulates adding layers one at a time until the blended color converges (DeltaE < 2.3 — the "just noticeable difference" threshold) or opacity exceeds 85%. The result is a set of transition zones, each with a start height, end height, and the filament used.
+- **Filaments**: Each filament has a hex color and a **Hiding Distance (HD)** value (in mm) — the depth of material at which the filament visually hides what's beneath it when viewed front-lit, the way finished prints are actually seen. Dark/opaque filaments hide quickly (e.g. 0.05–0.1 mm); light/translucent filaments need more depth (e.g. 0.5–0.7 mm). When you add a filament without specifying an HD, Kromacut estimates one from the color's luminance and saturation. If you have a conventional (backlit/lithophane) Transmission Distance from a spool sheet or TD test print, the convert button on the filament row turns it into a hiding distance (a conventional TD is roughly 10× the HD).
+- **Beer-Lambert optical blending**: The algorithm simulates how light transmits through stacked filament layers using the Beer-Lambert law: `transmission = 10^(-thickness / HD)`. It blends in linear-light sRGB before converting back for display and color matching, which better models the color you see from thin semi-transparent layers.
+- **Transition zones**: Each filament in the stack needs enough vertical space to visually transition from the color below it to its own pure color. The algorithm simulates adding layers one at a time until the blended color converges (DeltaE < 2.3 — the "just noticeable difference" threshold) or reaches the selected opacity endpoint. **Compact** uses 80% opacity, **Detailed** uses 90% (one HD), and **Maximum** uses 95% (the same endpoint as the foundation layer). Higher endpoints retain more printable intermediate colors but create taller stacks.
- **Luminance-to-height mapping**: Once the transition zones are computed, each pixel's brightness is mapped to a target height in the model. Dark pixels get the minimum height (base layer only), bright pixels get the full height (all layers), and mid-tones fall proportionally in between. This produces the characteristic lithophane-style relief where image brightness = model thickness.
### How it works (step by step)
-1. **Define filaments** — Add your filament colors and Transmission Distances in the Auto-paint tab. Use the color picker and TD input for each filament row.
+1. **Define filaments** — Add your filament colors and hiding distances in the Auto-paint tab. Use the color picker and HD input for each filament row (or convert a conventional TD value).
2. **Filament ordering** — By default filaments are sorted by luminance (darkest on the bottom, lightest on top). With **Enhanced color matching** enabled, the algorithm evaluates orderings to find the one that best reproduces the image's color palette.
-3. **Transition zone calculation** — For each consecutive pair of filaments, the algorithm simulates Beer-Lambert blending layer-by-layer until the color converges. The first (darkest) filament gets a foundation zone thick enough to be ~95% opaque (`TD × 1.3`).
+3. **Transition zone calculation** — For each consecutive pair of filaments, the algorithm simulates Beer-Lambert blending layer-by-layer until the color converges. The first (darkest) filament gets a foundation zone thick enough to be ~95% opaque (`HD × 1.3`).
4. **Height compression** — If the ideal height exceeds a user-set **Max Height**, all zones are uniformly compressed. The UI shows which zones are compressed and by how much.
5. **Virtual swatch generation** — The transition zones are sampled at each layer-height increment to produce a sequence of blended colors (virtual swatches). These drive the 3D preview and the height map, where each pixel maps to the layer whose blended color best matches.
### Filament Calibration
-Each filament row in the Auto-paint tab includes a **calibration wizard** to help you determine accurate Transmission Distance values:
+Kromacut measures hiding distances with a camera-free printed wedge — no light meter, no photos, no color picking:
-1. **Click the calibrate icon** next to any filament to open the calibration wizard.
-2. **Print test samples** — The wizard guides you through printing 3-5 samples of different thicknesses (e.g., 0.3mm, 0.6mm, 1.0mm, 1.5mm, 2.0mm).
-3. **Measure luminance** — Use a light meter or camera to measure how much light passes through each sample, or visually compare against reference colors.
-4. **Automatic TD calculation** — The wizard performs exponential regression on your measurements to compute the optimal TD value with a confidence score (High/Medium/Low/Very Low).
+1. **Click Calibrate Filaments** below the filament list and select the filaments to calibrate.
+2. **Pick bases** — Quick mode uses one base read per filament; Accurate mode repeats the read over two or three bases so Kromacut can measure per-channel (RGB) hiding distances directly.
+3. **Print the wedge** — Download an STL (any printer, one filament swap) or a multi-material 3MF (AMS-ready). Each filament prints as a row of patches from 1 layer up beside a fully opaque reference rail.
+4. **Read back one number** — For each print, report the first patch that looks identical to the reference rail. Kromacut converts the reads into measured hiding distances with a confidence score.
5. **Save profile** — Keep calibrated filaments in a reusable profile for future projects.
-Calibrated filaments display a confidence badge next to their TD value. Higher confidence = more accurate optical simulation = better print results.
+Calibrated filaments display a confidence badge next to their HD value. Higher confidence = more accurate optical simulation = better print results. Auto-paint uses calibrated red, green, and blue hiding distances for both preview blending and transition-zone thickness, so calibration can change the generated stack height and swap plan.
### Advanced Optimizer
-When **Enhanced color matching** is enabled, Kromacut uses advanced optimization algorithms to find the best filament order:
+When **Enhanced color matching** is enabled, choose an effort tier for the filament-order search:
-| Algorithm | When Used | Description |
+| Tier | Search | Best use |
|---|---|---|
-| **Exhaustive search** | 1-4 filaments | Evaluates all possible orderings to guarantee the optimal solution. Fast for small sets. |
-| **Simulated Annealing** | 5-8 filaments | Physics-inspired probabilistic search that escapes local minima via controlled randomness. Balances quality and speed. |
-| **Genetic Algorithm** | 9+ filaments | Evolution-inspired population-based search with crossover and mutation. Best for large filament sets where exhaustive search is prohibitive (8! = 40,320 permutations). |
-| **Auto** | Default | Automatically selects the best algorithm based on filament count and search space size. |
+| **Fast** | Narrow deterministic beam search. | A quick preview while tuning settings. |
+| **Balanced** | Full deterministic beam search. | Recommended default for most prints. |
+| **Thorough** | Full beam plus deeper multi-start refinement. | A stronger search without a long wait. |
+| **Deep** | Wider beam plus substantially more deterministic refinement. | Difficult images when Thorough does not improve the stack. |
+| **Exact base order** | Enumerates every no-repeat base order. | A guaranteed best base order; the candidate count grows very quickly with each filament. |
+
+Each higher tier retains the best result from the tier below for the same seed. Exact
+base order checks 109,600 sequences at eight filaments and 986,409 at nine, so use it
+deliberately on larger profiles. With extra repeated swaps enabled, its base order is exact
+but repeated occurrences are still refined heuristically.
The optimizer displays metadata after generation:
- **Algorithm used** — Which method was selected
@@ -154,7 +162,7 @@ The optimizer displays metadata after generation:
- **Converged** — Whether the algorithm reached a stable solution
- **Cache hit** — Whether results were retrieved from cache (instant)
-**Optimizer seed** — Set a random seed for reproducible results across runs. Useful for A/B testing different configurations.
+**Optimizer seed** — Leave blank for an automatic stable seed, or enter a number to use a specific repeatable seed for A/B comparisons.
### Region Weighting
@@ -164,7 +172,7 @@ Prioritize accuracy in specific image regions during optimization:
|---|---|
| **Uniform** | All pixels weighted equally (default). |
| **Center** | Gaussian falloff from center — faces and subjects in the middle get higher priority. |
-| **Edge** | Sobel edge detection — high-contrast boundaries prioritized over flat regions. |
+| **Edge** | Colors nearer the outer edges of the image get higher priority. |
Region weighting is most useful when filament budget is limited and you want the optimizer to focus on visually important areas.
@@ -173,14 +181,18 @@ Region weighting is most useful when filament budget is limited and you want the
| Option | Description |
|---|---|
| **Max Height** | Constrains the total model height (mm). When set below the auto-calculated ideal, zones are uniformly compressed. Leave blank or click `Auto` for the physics-derived default. |
-| **Enhanced color matching** | Optimizes filament ordering for best color reproduction rather than simple luminance sorting. Uses advanced algorithms (exhaustive, simulated annealing, genetic) automatically selected based on filament count. Scoring considers weighted DeltaE accuracy, height spread, layer count, and transition waste. |
-| **Allow repeated filament swaps** | (Requires Enhanced color matching) Allows a filament to appear more than once in the stack. This creates intermediate blended colors — for example, a thin white layer over red produces pink. The algorithm greedily inserts up to 4 extra swaps, each at the position that best improves the score. |
-| **Height dithering** | (Requires Enhanced color matching) Applies block-aware Floyd-Steinberg error diffusion to the quantized height map. Instead of sharp stair-steps between layer heights, dithering produces a stippled gradient that simulates intermediate heights, resulting in smoother tonal transitions in the print. Edge pixels between different heights are protected from dithering to avoid staircase artifacts. |
+| **Enhanced color matching** | Optimizes the printable filament sequence rather than simply sorting by luminance. It may omit filaments that do not improve the result and evaluates the actual preview color-to-height path. |
+| **Extra repeated swaps** | (Requires Enhanced color matching) Lets the optimizer use a filament more than once. Choose Off, 2, 4, 6, 8, or 12 additional occurrences; more repeats may create useful blend paths but substantially expand the search. |
+| **Preserve color separation** | (Requires Enhanced color matching) Treats the configurable ΔE value as a hard eligibility limit for distinct printable colors. The optimizer searches repeat tiers progressively and ranks candidates by preserved-color count, source coverage, preserved-match error, repeats, then stack length. Strict mode rejects a result unless every 2D color receives a unique in-limit match; non-strict mode drops unmatched colors by merging their regions into surviving preserved mappings. Mutually exclusive with Height dithering. |
+| **Height dithering** | (Requires Enhanced color matching) Applies block-aware Stucki error diffusion to the quantized height map. Instead of sharp stair-steps between layer heights, dithering produces a stippled gradient that simulates intermediate heights, resulting in smoother tonal transitions in the print. Edge pixels between different heights are protected from dithering to avoid staircase artifacts. Mutually exclusive with Preserve color separation. |
| **Flat Paint (flat face-down print)** | Builds a uniform-thickness slab printed image-side down instead of a stepped relief. Each pixel column's layer order is reversed so the artwork sits against the build plate (already mirrored — don't mirror in the slicer) under a transparent carrier layer, and the back is filled with the foundation filament so every layer has the full footprint. The result has a smooth, glass-flat face — great for bookmarks and coasters. Requires a multi-material printer (AMS/toolchanger); export as 3MF, which contains one object per filament plus the clear carrier object. Flat Paint and Smooth Meshing toggle each other off because flat prints always use the full-footprint slab layout. |
| **Dither line width** | (Requires Height dithering) Controls the minimum dot size for the dither pattern in mm. This should roughly match your printer's line/nozzle width so dither dots are actually printable. Default: `0.42 mm`. |
-| **Optimizer algorithm** | Choose which optimization algorithm to use: Auto (recommended), Exhaustive, Simulated Annealing, or Genetic. Auto selects the best algorithm based on search space size. |
-| **Optimizer seed** | Set a random seed for reproducible optimizer results. Leave blank for random behavior. Useful for testing and comparing configurations. |
-| **Region weighting** | Prioritize specific image regions: Uniform (equal), Center (Gaussian falloff), or Edge (Sobel detection). Helps focus quality budget on important areas. |
+| **Optimizer algorithm** | Choose Fast for a preview, Balanced for the recommended full beam, Thorough or Deep for more search effort, or Exact base order when you want every no-repeat base stack checked. |
+| **Transition detail** | Chooses where a color transition stops: Compact at 80% opacity, Detailed at 90%, or Maximum at 95%. Higher detail preserves a longer physical color ramp, increasing model height, swaps, and runtime. |
+| **Optimizer seed** | Leave blank for an automatic stable seed, or enter a number for a specific repeatable run. |
+| **Region weighting** | Prioritize specific image regions: Uniform (equal), Center (image middle), or Edge (outer image border). Helps focus quality budget on important areas. |
+
+Enhanced matching uses every color in the palette currently produced by 2D mode; it does not silently reduce that palette again. For detail-critical work, first prepare the image in 2D with K-means (for example, weight 128) and the Auto palette set to the number of colors you want to preserve. More 2D colors improve the target representation but make the optimizer proportionally slower.
### Filament profiles
@@ -190,7 +202,7 @@ Auto-paint configurations (filament lists) can be saved, loaded, and shared as *
- **Load** — Select a saved profile from the dropdown to restore its filaments.
- **Rename** — Change the selected profile name without changing its filaments.
- **Export** — Download the current filaments as a `.kfil` file (JSON) to share with others.
-- **Import** — Load a `.kfil`, legacy `.kapp`, or `.json` profile file from disk.
+- **Import** — Load a `.kfil`, legacy `.kapp`, `.json`, or HueForge spool `.csv`/`.tsv` profile file from disk.
- **Delete** — Remove a saved profile.
Profiles are stored in browser `localStorage` and persist across sessions.
@@ -214,7 +226,7 @@ The result card shows:
|---|---|
| **Hex** | Suggested filament color. |
| **Est. ΔE** | Estimated reduction in blend-aware average ΔE if this filament is added. A rough relative estimate — not a calibration confidence rating. |
-| **TD** | Recommended starting Transmission Distance, borrowed from the nearest existing filament by color distance. Adjust after printing a test patch. |
+| **HD** | Recommended starting hiding distance, borrowed from the nearest existing filament by color distance. Adjust after printing a test patch. |
| **Captures** | Percentage of image pixels whose blend-aware color error would improve with this filament. |
| **Isolation** | How far this color sits from existing filaments in perceptual color space (0–1). Higher means it fills a more distinct gap; lower means it overlaps territory already covered by blending. |
@@ -226,9 +238,9 @@ Click **Add to filaments** to insert the suggestion directly into the filament l
1. Load an image into Kromacut.
2. Switch to the **Auto-paint** tab (inside the 3D controls panel).
-3. Click **Add Filament** and configure each filament's color and TD to match your real filament stock.
- - **Tip:** Use the calibration wizard (calibrate icon on each row) to measure accurate TD values.
-4. (Optional) Enable **Enhanced color matching** for better results with complex images. The optimizer will automatically select the best algorithm (exhaustive, simulated annealing, or genetic) based on your filament count.
+3. Click **Add Filament** and configure each filament's color and hiding distance to match your real filament stock.
+ - **Tip:** Use **Calibrate Filaments** to measure accurate hiding distances from a printed wedge.
+4. (Optional) Enable **Enhanced color matching** for better results with complex images. Start with the **Balanced** optimizer tier, then use Thorough, Deep, or Exact base order when you want to spend more search effort.
5. (Optional) Set **Region weighting** to Center or Edge to prioritize accuracy in visually important areas.
6. The 3D preview updates automatically. Adjust **Max Height** if the model is too tall.
7. Use the **layer-by-layer preview slider** at the bottom of the 3D view to verify color transitions.
@@ -242,18 +254,20 @@ Kromacut now supports exporting directly to `.3mf` format. This file format pres
-## Transmission Distance (TD) — what it is and how to use it here
+## Hiding Distance (HD) and Transmission Distance (TD)
+
+Stacking thin translucent filament layers produces perceptual intermediate shades — the idea HueForge popularized with Transmission Distance: https://shop.thehueforge.com/blogs/news/what-is-hueforge
-Transmission Distance (TD) is the concept HueForge uses to produce perceptual intermediate shades by stacking translucent filament layers: instead of relying purely on opaque color pigments, TD models how light transmits through thin layers of filament and how stacking different colors (and varying thickness) produces new perceived colors. For a full conceptual description see the HueForge blog: https://shop.thehueforge.com/blogs/news/what-is-hueforge
+Kromacut models the same physics but stores a **Hiding Distance (HD)**: the depth at which a filament visually hides what's beneath it under front lighting, which is how finished prints are viewed. A conventional TD (measured with a backlit lithophane-style test print) is roughly **10× the hiding distance**; the convert button on each filament row accepts a conventional TD and converts it for you.
### Two workflow modes in Kromacut:
-**1. Auto-paint mode (Automatic TD processing)** — The recommended approach for TD-based prints:
+**1. Auto-paint mode (Automatic optical blending)** — The recommended approach for layered-color prints:
- **Automatic Beer-Lambert blending** — Kromacut simulates light transmission through stacked filament layers using physically accurate optical models.
-- **Filament-based workflow** — Define your actual filaments (color + TD values) and let the algorithm compute optimal layer stacks automatically.
-- **Calibration wizard** — Measure accurate TD values from physical test prints for each filament.
-- **Advanced optimizer** — Simulated annealing and genetic algorithms find the best filament ordering for complex images.
+- **Filament-based workflow** — Define your actual filaments (color + hiding distances) and let the algorithm compute optimal layer stacks automatically.
+- **Camera-free calibration** — Print a small wedge and read back the layer count where it turns opaque; Kromacut fits measured per-channel hiding distances from the reads.
+- **Advanced optimizer** — Deterministic effort tiers search for the best filament ordering for complex images, from quick beam previews to exact base-order enumeration.
- **Transition zones** — Automatically calculated vertical zones where each filament blends from the color below to its own pure color.
See the [Auto-paint section](#auto-paint) above for full details.
@@ -265,17 +279,17 @@ See the [Auto-paint section](#auto-paint) above for full details.
- Small adjustments to `base slice height`, per-color slice heights, and `layerHeight` change the produced layer numbers and perceived blends.
- Iterate with actual filament on small test prints — translucency and perceived mix depend heavily on filament brand, color, and print settings.
-**Recommendation:** Use Auto-paint mode for TD-based prints. It handles the complex Beer-Lambert physics and optimization automatically, producing better results than manual approximation.
+**Recommendation:** Use Auto-paint mode for layered-color prints. It handles the complex Beer-Lambert physics and optimization automatically, producing better results than manual approximation.
### TD test image (for manual mode experimentation)
- This repository includes a small test image `tdTest.png` (found in `src/assets/tdTest.png`).
-- The app header includes a `Load TD Test` button that will load that image into the preview so you can quickly experiment with manual layer ordering and per-color slice heights.
+- Load it into the preview with the toolbar upload button or drag-and-drop so you can quickly experiment with manual layer ordering and per-color slice heights.
- This is primarily useful for testing manual mode workflows or understanding how layer heights affect perceived color.
Quick manual workflow experiment:
-1. Click `Load TD Test` in the app header.
+1. Load `tdTest.png` (upload button or drag-and-drop).
2. Reduce colors to your desired number.
3. Switch to Manual mode (3D Controls → Manual tab).
4. Adjust per-color slice heights and ordering.
diff --git a/content/community/batmanga-finished.jpg b/content/community/batmanga-finished.jpg
new file mode 100644
index 00000000..4c1603c2
Binary files /dev/null and b/content/community/batmanga-finished.jpg differ
diff --git a/content/community/batmanga-kromacut-preview.png b/content/community/batmanga-kromacut-preview.png
new file mode 100644
index 00000000..3ca53ed2
Binary files /dev/null and b/content/community/batmanga-kromacut-preview.png differ
diff --git a/content/community/batmanga-slicer-preview.png b/content/community/batmanga-slicer-preview.png
new file mode 100644
index 00000000..9f3641a2
Binary files /dev/null and b/content/community/batmanga-slicer-preview.png differ
diff --git a/content/community/candidates/batmanga-2026-09-06/README.md b/content/community/candidates/batmanga-2026-09-06/README.md
new file mode 100644
index 00000000..f2c689e8
--- /dev/null
+++ b/content/community/candidates/batmanga-2026-09-06/README.md
@@ -0,0 +1,17 @@
+# Batmanga showcase candidate
+
+Saved at the user's request on 2026-09-06 before printing. The user subsequently supplied `20260906_214317.jpg` and approved adding the finished print with both screenshots to the showcase. The finished photo was then replaced with the user's cropped version from `D:/Downloads/20260906_214317.jpg`, copied without further edits. Exact copies are included at `content/community/batmanga-kromacut-preview.png`, `batmanga-slicer-preview.png`, and `batmanga-finished.jpg`. The archived screenshots remain unchanged; the uncropped photo remains at its original camera-file location.
+
+- `batmanga-kromacut-preview.png`: original Kromacut screenshot, copied without resizing or editing.
+- `batmanga-slicer-preview.png`: original Creality Print screenshot, copied without resizing or editing.
+
+Context visible in the screenshots:
+
+- Artwork title: Batman: The Jiro Kuwata Batmanga, volume 1. Credit: Jiro Kuwata / DC; [source image supplied by the user](https://m.media-amazon.com/images/I/81rOZq5ZgqL._AC_UF1000,1000_QL80_.jpg). This replaces the previously linked DC book page, which was not the user's image source. Print and screenshots by vycdev.
+- Kromacut model display: approximately 65.8 × 100.0 × 3.7 mm; transition summary shows 3.68 mm and 42 physical layers.
+- Deep optimizer, enhanced matching and Preserve color separation enabled; height dithering and Flat Paint disabled.
+- Creality Hi with a 0.4 mm nozzle; slicer layer height 0.08 mm and first-layer height 0.16 mm.
+- Slicer estimates: 3 h 12 min, 21.30 g total filament, eight filament changes. These are estimates, not measured print results.
+- Slicer screenshot is shown at layer 41 / Z 3.36 mm. This is the displayed preview position, not a confirmed total model height.
+
+Keep the two previews distinguishable from the eventual physical-print photograph. No color-accuracy conclusion has been drawn from these screenshots.
diff --git a/content/community/candidates/batmanga-2026-09-06/batmanga-kromacut-preview.png b/content/community/candidates/batmanga-2026-09-06/batmanga-kromacut-preview.png
new file mode 100644
index 00000000..3ca53ed2
Binary files /dev/null and b/content/community/candidates/batmanga-2026-09-06/batmanga-kromacut-preview.png differ
diff --git a/content/community/candidates/batmanga-2026-09-06/batmanga-slicer-preview.png b/content/community/candidates/batmanga-2026-09-06/batmanga-slicer-preview.png
new file mode 100644
index 00000000..9f3641a2
Binary files /dev/null and b/content/community/candidates/batmanga-2026-09-06/batmanga-slicer-preview.png differ
diff --git a/content/community/candidates/naruto-2026-09-06/README.md b/content/community/candidates/naruto-2026-09-06/README.md
new file mode 100644
index 00000000..562f5ac8
--- /dev/null
+++ b/content/community/candidates/naruto-2026-09-06/README.md
@@ -0,0 +1,18 @@
+# Naruto showcase candidate
+
+Saved at the user's request on 2026-09-06 before a planned overnight print. The user was unsure whether the calibration supported this image well or whether the prediction would be accurate. On 2026-09-07 the user supplied `D:/Downloads/20260907_080653.jpg` and approved adding all three views to the showcase. Exact copies are included at `content/community/naruto-kromacut-preview.png`, `naruto-slicer-preview.png`, and `naruto-finished.jpg`. All images were copied without editing; the archived previews remain unchanged.
+
+- `naruto-kromacut-preview.png`: original Kromacut screenshot, copied without resizing or editing.
+- `naruto-slicer-preview.png`: original Creality Print screenshot, copied without resizing or editing.
+
+Visible context:
+
+- Naruto looking upward against a blue sky in the Kromacut appearance preview. The user supplied this [Pinterest pin](https://in.pinterest.com/pin/169870217190172931/) as the artwork source; the original artist has not been verified.
+- Kromacut model display: approximately 110.4 × 190.2 × 4.1 mm; layer preview endpoint 4.08 mm.
+- Pixel size 0.15 mm, layer height 0.08 mm, first-layer height 0.40 mm, Smooth Meshing enabled.
+- Selected profile: 8 Colors Calibrated Frontlit. Profile ratings shown in the screenshot do not establish prediction accuracy for this particular stack.
+- Creality Hi with a 0.4 mm nozzle. Slicer layer height 0.08 mm and first-layer height 0.16 mm. The first-layer values differ between screenshots; this is recorded without diagnosing its effect.
+- Slicer estimates: 5 h 55 min, 72.10 g total filament, nine filament changes; seven listed filaments. These are estimates, not completed-print measurements.
+- Slicer preview is displayed at layer 50 / Z 4.08 mm.
+
+The showcase distinguishes the Kromacut appearance prediction, slicer's filament-color preview, and finished print without claiming measured color accuracy. Print and screenshots by vycdev. The footer links the user's Pinterest source without attributing the underlying artwork to the print maker or guessing the original artist.
diff --git a/content/community/candidates/naruto-2026-09-06/naruto-kromacut-preview.png b/content/community/candidates/naruto-2026-09-06/naruto-kromacut-preview.png
new file mode 100644
index 00000000..8060c37f
Binary files /dev/null and b/content/community/candidates/naruto-2026-09-06/naruto-kromacut-preview.png differ
diff --git a/content/community/candidates/naruto-2026-09-06/naruto-slicer-preview.png b/content/community/candidates/naruto-2026-09-06/naruto-slicer-preview.png
new file mode 100644
index 00000000..f45c3052
Binary files /dev/null and b/content/community/candidates/naruto-2026-09-06/naruto-slicer-preview.png differ
diff --git a/content/community/candidates/titan-2026-09-06/README.md b/content/community/candidates/titan-2026-09-06/README.md
new file mode 100644
index 00000000..e9bd83ca
--- /dev/null
+++ b/content/community/candidates/titan-2026-09-06/README.md
@@ -0,0 +1,20 @@
+# Titan showcase candidate
+
+Saved at the user's request on 2026-09-06. The user subsequently approved adding a Titan showcase with these screenshots and selected `20260906_172702.jpg` as the replacement finished-print photo. Exact copies are included as `content/community/titan-kromacut-preview.png`, `titan-slicer-preview.png`, and `titan-finished.jpg`; these archived screenshots remain unchanged. The previous photo `20260906_170622.jpg` remains at its original camera-file location.
+
+- `titan-kromacut-preview.png`: original Kromacut screenshot, copied without resizing or editing.
+- `titan-slicer-preview.png`: original Creality Print screenshot, copied without resizing or editing.
+
+Context visible in the screenshots:
+
+- Titan artwork; Kromacut model dimensions displayed as approximately 102.4 × 152.7 × 2.6 mm.
+- Regular layer height: 0.08 mm in both screenshots.
+- First-layer settings differ: Kromacut shows 0.40 mm; slicer shows 0.16 mm. Retained as observed, without interpreting the effect on the exported geometry.
+- Slicer estimates: 2 h 58 min, 28.90 g total filament, six filament changes. These are estimates, not completed-print measurements.
+
+Original artwork: [NASA/JPL — Titan, Visions of the Future](https://www.jpl.nasa.gov/images/titan-jpl-travel-poster/). Print and screenshots by vycdev. The landing page labels each view separately and makes no measured color-accuracy claim. The selected finished photograph is retained without image edits; the earlier washed-out phone photo is not used.
+
+Verified original-file SHA-256 hashes:
+
+- Kromacut: `735f5066a77e62a72687997c5790d6f8480430e30719a5385da444e5c287c48d`
+- Slicer: `23054e0ebddd9b0a37b84b87bf92a47db09e5ef479f5f8e52ce0d51635be52cd`
diff --git a/content/community/candidates/titan-2026-09-06/titan-kromacut-preview.png b/content/community/candidates/titan-2026-09-06/titan-kromacut-preview.png
new file mode 100644
index 00000000..d2b4f8a5
Binary files /dev/null and b/content/community/candidates/titan-2026-09-06/titan-kromacut-preview.png differ
diff --git a/content/community/candidates/titan-2026-09-06/titan-slicer-preview.png b/content/community/candidates/titan-2026-09-06/titan-slicer-preview.png
new file mode 100644
index 00000000..05f45e7a
Binary files /dev/null and b/content/community/candidates/titan-2026-09-06/titan-slicer-preview.png differ
diff --git a/content/community/hobbits-and-dragons-1.jpg b/content/community/hobbits-and-dragons-1.jpg
new file mode 100644
index 00000000..0b998412
Binary files /dev/null and b/content/community/hobbits-and-dragons-1.jpg differ
diff --git a/content/community/hobbits-and-dragons-2.jpg b/content/community/hobbits-and-dragons-2.jpg
new file mode 100644
index 00000000..b87d9b16
Binary files /dev/null and b/content/community/hobbits-and-dragons-2.jpg differ
diff --git a/content/community/hope-finished.jpg b/content/community/hope-finished.jpg
new file mode 100644
index 00000000..bfcfae35
Binary files /dev/null and b/content/community/hope-finished.jpg differ
diff --git a/content/community/hope-kromacut-preview.jpg b/content/community/hope-kromacut-preview.jpg
new file mode 100644
index 00000000..d2364728
Binary files /dev/null and b/content/community/hope-kromacut-preview.jpg differ
diff --git a/content/community/hope-slicer-preview.jpg b/content/community/hope-slicer-preview.jpg
new file mode 100644
index 00000000..dd4695aa
Binary files /dev/null and b/content/community/hope-slicer-preview.jpg differ
diff --git a/content/community/king-of-hearts-slicer.jpg b/content/community/king-of-hearts-slicer.jpg
new file mode 100644
index 00000000..5d615b34
Binary files /dev/null and b/content/community/king-of-hearts-slicer.jpg differ
diff --git a/content/community/king-of-hearts.jpg b/content/community/king-of-hearts.jpg
new file mode 100644
index 00000000..ef8aa1ed
Binary files /dev/null and b/content/community/king-of-hearts.jpg differ
diff --git a/content/community/naruto-finished.jpg b/content/community/naruto-finished.jpg
new file mode 100644
index 00000000..05daf552
Binary files /dev/null and b/content/community/naruto-finished.jpg differ
diff --git a/content/community/naruto-kromacut-preview.png b/content/community/naruto-kromacut-preview.png
new file mode 100644
index 00000000..8060c37f
Binary files /dev/null and b/content/community/naruto-kromacut-preview.png differ
diff --git a/content/community/naruto-slicer-preview.png b/content/community/naruto-slicer-preview.png
new file mode 100644
index 00000000..f45c3052
Binary files /dev/null and b/content/community/naruto-slicer-preview.png differ
diff --git a/content/community/titan-finished.jpg b/content/community/titan-finished.jpg
new file mode 100644
index 00000000..14355646
Binary files /dev/null and b/content/community/titan-finished.jpg differ
diff --git a/content/community/titan-kromacut-preview.png b/content/community/titan-kromacut-preview.png
new file mode 100644
index 00000000..d2b4f8a5
Binary files /dev/null and b/content/community/titan-kromacut-preview.png differ
diff --git a/content/community/titan-slicer-preview.png b/content/community/titan-slicer-preview.png
new file mode 100644
index 00000000..05f45e7a
Binary files /dev/null and b/content/community/titan-slicer-preview.png differ
diff --git a/index.html b/index.html
index 10f08e30..fe4e9bb4 100644
--- a/index.html
+++ b/index.html
@@ -59,6 +59,34 @@
Frozen diagnostic print map
Start with A · Orange, B · Cyan, C · White. One strip per print job. Read print instructions before slicing.
Notch at top-left. Pad IDs run left → right. Measure the pad centres, not the exposed gaps. Diagrams are NOT to scale. Predictions are hidden initially so you can record observations first.
+${strips + .map( + ( + s + ) => `${escape(s.design.id)} · ${escape(s.design.title)}
${escape(s.design.purpose)}
${s.dimensions.width} × ${s.dimensions.depth} × ${s.dimensions.height} mm · 3MF · ${s.layers.filter((l, i) => i > 0 && l.filamentId !== s.layers[i - 1].filamentId).length} filament changes
+ +Backing: ${s.backing.map((r) => `${escape(r.filamentColor)} ${r.thickness.toFixed(2)} mm`).join(' → ')}. Top: ${escape(s.design.foreground)}. ${s.patches + .filter((p) => p.repeatOf) + .map((p) => `${p.id} repeats ${p.repeatOf}.`) + .join(' ')}
+${s.warnings.map((w) => `${escape(w)}
`).join('')} +Reveal frozen predictions (estimates, not physical guarantees)
| Pad | Top layers | Predicted sRGB | Evidence method |
|---|---|---|---|
| ${p.id} | ${p.topLayers} | ${p.predictedHex} | ${escape(p.prediction.predictionConfidence.method)} |
- {
- invalidate();
- setImage(tdTestImg, true);
- setMode('2d');
- if (parseDocsLocation(window.location)) {
- window.history.pushState(null, '', '/');
- }
- setDocsOpen(false);
- }}
- />
- {docsOpen && (
-
- {/* Build warning dialog */}
-
-
- )}
-
+ <>
+
+
+
-
@@ -632,7 +842,15 @@ function App(): React.ReactElement | null {
onDragOver={dropzone.onDragOver}
onDragLeave={dropzone.onDragLeave}
>
- {mode === '2d' ? (
+ {mode === '3d' && !hasMountedThreeDControls && (
+ {
+ const url = URL.createObjectURL(blob);
+ invalidate();
+ setImage(url, true);
+ return url;
+ }}
+ onTouchUpPickColor={(hex) => {
+ setTouchUpColor(hex);
+ setTouchUpTool('brush');
+ }}
+ onTouchUpExit={() => setTouchUpTool(null)}
/>
{processingActive && (
imageSrc && setIsCropMode(true)}
+ onEnterCrop={() => {
+ if (!imageSrc) return;
+ setTouchUpTool(null);
+ setIsCropMode(true);
+ }}
onSaveCrop={async () => {
if (!canvasPreviewRef.current) return;
const blob =
@@ -733,7 +998,28 @@ function App(): React.ReactElement | null {
onExportStl={onExportStl}
onExport3MF={onExport3MF}
flatPaintModel={builtFlatPaint}
+ modelExportDisabled={!builtModelValid}
isOrtho={isOrtho}
+ previewRenderMode={previewRenderMode}
+ onPreviewRenderModeChange={(next) => {
+ setPreviewRenderMode(next);
+ savePreviewRenderMode(next);
+ }}
+ autoPaintEnabled={builtModelAutoPaint}
+ previewColorMode={previewColorMode}
+ onPreviewColorModeChange={(next) => {
+ setPreviewColorMode(next);
+ savePreviewColorMode(next);
+ }}
+ touchUpTool={touchUpTool}
+ onTouchUpToolChange={handleTouchUpToolChange}
+ touchUpColor={touchUpColor}
+ onTouchUpColorChange={setTouchUpColor}
+ touchUpBrushSize={touchUpBrushSize}
+ onTouchUpBrushSizeChange={setTouchUpBrushSize}
+ touchUpTextSize={touchUpTextSize}
+ onTouchUpTextSizeChange={setTouchUpTextSize}
+ paletteColors={touchUpPaletteColors}
onToggleCamera={() =>
setIsOrtho((v) => {
const next = !v;
@@ -745,38 +1031,42 @@ function App(): React.ReactElement | null {
-
+
+ )}
+
<>
)}
-
- Building the 3D model may be slow due to:
--
- {buildWarning?.warnings.map((w, i) => (
-
- {w} - ))} -
Do you want to continue?
-