How Mixora turns a target color into a paint recipe

1. You define the target

Enter a HEX value, choose a color visually, or sample a photograph. Mixora converts that sRGB target to CIELAB under a defined reference condition. A digital target is useful because it is repeatable, but it does not contain information about gloss, texture, opacity, fluorescence, metallic reflection, or the original lighting.

2. You define the available palette

The search uses only paints enabled in your current scope: owned paints, selected collections, optional favorites, and an active project palette. This makes the answer practical. A theoretically close mixture is not useful when it depends on a tube you do not own or a paint type that is incompatible with the project.

3. Paints are represented spectrally

Each paint can carry reflectance information across 36 wavelength bands from 380 to 730 nm. When measured reflectance or absorption/scattering data is available, Mixora uses it. When a paint has only a reference HEX value, the engine reconstructs one smooth spectral metamer and labels the data as reconstructed.

Many different spectra can look like the same RGB color under one condition. A reconstruction is therefore an informed working estimate, not a recovered measurement of the real tube.

4. Candidate mixtures use Kubelka–Munk theory

For each candidate ratio, Mixora combines wavelength-dependent absorption K and scattering S values and predicts the reflectance of an optically thick paint layer. This is more appropriate for pigment mixing than averaging RGB channels because it models subtraction and scattering across the visible spectrum.

The simplified model still cannot know particle size, pigment concentration, binder, extender, film thickness, surface texture, contamination, aging, or the exact behavior of a proprietary paint formula.

5. Predictions are compared in CIELAB

The predicted spectrum is evaluated under D65 illumination with a standard observer and converted to CIELAB. Mixora uses CIEDE2000 Delta E to rank perceptual color difference. Lower is closer within the model and viewing assumptions. A low score does not guarantee an invisible physical difference across screens, lamps, surfaces, and finishes.

6. The result becomes a practical ratio

Depending on Easy, Precise, or Expert mode, the search balances closeness with a manageable number of paints and practical whole-part ratios. You can adjust individual parts in the live preview and see the predicted color and Delta E update before committing paint to the palette.

7. You verify and calibrate

Measure one consistent unit, mix a small batch, apply it over the intended ground at the intended thickness, and let it dry completely. Correct value first, then hue, then chroma. Save the successful physical ratio with brand, product line, surface, medium, coat, and lighting notes.

Start with the repeatable mixing workflow, or browse curated color recipes to see the pigment logic behind common targets.

Data quality labels

Measured means the calculation uses measured spectral input for the contributing paints. Hybrid means measured and reconstructed inputs appear together. Reconstructed means the spectral behavior is estimated from reference color data. These labels describe input confidence, not a guarantee about the final physical mixture.