Harmonious Color Pairings: Insights from Human Preference and Natural Hue Statistics

TL;DR

Quantifies color harmony using controlled hue palettes in the HSL space by constructing a score matrix $S$ from participant judgments and deriving a Combinability index $C(j)$; analyzes the symmetric component $S^{s}=\tfrac{1}{2}(S+S^{\top})$ with principal component analysis to reveal two complementary hue groups and their ecological alignment with natural hue statistics. The study finds hue-dependent harmony with a robust global contrast-preference region around $\Delta\theta\approx160^{\circ}$–$220^{\circ}$, and shows that the observed patterns align with hue distributions in natural landscapes, suggesting perceptual and ecological underpinnings. It provides a quantitative framework for color harmony, highlights limitations of hue-only analysis and the HSL color space, and outlines future work toward perceptually uniform color spaces and broader, more diverse participant samples.

Abstract

While color harmony has long been studied in art and design, a clear consensus remains elusive, as most models are grounded in qualitative insights or limited datasets. In this work, we present a quantitative, data-driven study of color pairing preferences using controlled hue-based palettes in the HSL color space. Participants evaluated combinations of thirteen distinct hues, enabling us to construct a preference matrix and define a combinability index for each color. Our results reveal that preferences are highly hue dependent, challenging the assumption of universal harmony rules proposed in the literature. Yet, when averaged over hues, statistically meaningful patterns of aesthetic preference emerge, with certain hue separations perceived as more harmonious. Strikingly, these patterns align with hue distributions found in natural landscapes, pointing to a statistical correspondence between human color preferences and the structure of color in nature. Finally, we analyze our color-pairing score matrix through principal component analysis, which uncovers two complementary hue groups whose interplay underlies the global structure of color-pairing preferences. Together, these findings offer a quantitative framework for studying color harmony and its potential perceptual and ecological underpinnings.

Figures (8)

• Figure 1: Score matrix S, computed from the survey results and Eq. \ref{['eq:score']}.
• Figure 2: Combinability index from Eq. \ref{['eq:combin']} (circular markers, solid line), absolute preferences from palmer2010human (triangular markers, dashed line), and hue distribution of 12,000 natural landscape images (background histogram), as a function of hue angular values $\theta$. The solid gray lines reflect the contribution of each color to both the best-only and worst-only indices (see main text).
• Figure 3: Average preferences for color pairs as function of their angular distance $\Delta\theta$ on the hue wheel (see also Fig. \ref{['fig:wheel_errorbar']}). The background gray histogram represents the distribution of angular distances between hues found in natural images. The maximum angular distance between hues is $\Delta\theta=180^\circ$.
• Figure 4: (a) Eigenvalues of the score matrix S on the complex plane. (b) Eigenvalues of the symmetrized score matrix S$^s= \frac{1}{2}(\text{S} + \text{S}^\top)$. (c) Outer products $\lambda_i^\text{s} \bm{\text{v}_i} \bm{\text{v}_i}^\top$ for the six eigenvalues $\lambda_i^{\text{s}}$ with the highest absolute values, where $\bm{\text{v}_i}$ denotes the eigenvector associated with the $i^{\text{th}}$ eigenvalue. To highlight the clustered structure of the outer products, an index rotation has been applied to start with orange instead of red for the first line/column.
• Figure 5: Representation of our 13 sampled hues in the CIE 1931 xy chromaticity diagram. The outer curved boundary represents the spectral locus and the light-gray triangle corresponds to the sRGB gamut anderson1996proposalsusstrunk1999standard. Hues belonging to group 1 and group 2 are plotted as squares and triangles respectively, and dotted lines project each color to its dominant wavelength on the spectral locus. The two separating lines are obtained using hard-margin Support Vector Machine (SVM) cortes1995support. The gray one is the standard hard-margin solution, which maximizes the distance to the nearest points of both groups, while the black dash-dotted line is obtained by assigning a weight to the two support vectors on each side that, intuitively, encodes how strongly a color belongs to its group and is defined as $\omega_i = ( \sum_{j:\, \mathrm S^{s}_{ij}<0} |\mathrm S^{s}_{ij}| )/ ( \sum_{p,q:\, \mathrm S^{s}_{pq}<0} |\mathrm S^{s}_{pq}| )$ .