Visualizing How a Fragrance Evolves Perceptually: Cloud Plots

Most formula documentation shows you a list of materials and percentages. It doesn't show you how a fragrance actually behaves over time - which notes rise, which fade, and where the perceptual "shape" of a scent shifts across the wear arc. I've been building a set of Cloud Plots for formula research and development at Nodus Perfume House - visualizations that model the boundary between what's detectable and what isn't, across time. Each curve traces how a material's presence evolves from application through drydown.

An example from recent analytics is the cloud plot for the classic fragrance Seduction Homme, which was reconstructed by Wisemoor:

Leif Peterson 1 July 2026

In the cloud plot for Seduction above, the materials in the legend tabulate which constituent materials are perceptible over the wear arc. Imperceptible materials whose concentrations are too low for perception are listed in the footnote of the plot. Note: the materials listed include single materials in the formula used, as well as inferred constituent materials in essential oils, accords, resins, etc. In other words, the materials listed are not a direct display of raw input materials in the formula used.

The top notes commonly start cloud formation (evaporating) early, and can disappear in a matter of minutes to several hours. In point of fact, you can see limonene, linalyl acetate, and linalool clouds appearing early and then evaporating away. The "heart" or middle note materials typically form clouds near the middle of the wear arc, and then evaporate off within several hours after skin application. Florhydrol and citronellol elicit clouds toward the middle of the wear arc, and then disappear. At the top of the cloud plot, it's apparent that damascenone, Iso E Super, and muscenone form the largest "cloud" initially, since they form the top curves. A pattern observed is that many reconstructed commercial formulas use materials that are diffusive, airy, and long-lasting to provide a specific olfactive perception or experience. Such materials don't form rigid boundaries of the olfactive experience but rather a diffusive cloud-like veil over the wearer or skin location where it was applied. There are numerous fragrance materials employed for establishing these diffusive clouds, and some of the most notable are the large ring (carbon) macrocyclic musks, ambroxan, and hedione -- otherwise known as base materials. Other base materials are long-lasting resins, resinoids, or synthetics. Many base materials evaporate slowly and bocome embedded in the skin, and if used at too great a concencration can bind with lighter molecules and hinder their evaporation - so there are always tradeoffs.

First and foremost, these novel cloud plots are a different way of looking at perfumery - treating the wear experience as something that can be modeled and compared, not just described. For me, they are part of a long journey to decipher and disentangle a "latent semantic language of perfumery," which can be visible from pattern recognition involving manifold learning and latent dimension reduction. This involves a combination of algorithms ranging from molecular fingerprinting, cheminformatic n-gram analysis, text mining, etc. Thus far, a latent semantic langauge has emerged that indicates large perfume houses don't always view the scent of individual materials singly, but rather consider the joint olfactive response from multiple materials as an ensemble, i.e., as an accord. In other words, there may be less meaning to a statement like "florol smells like cut flower stems," qwhn in the presence of other materials may take on a wholly different olfactive experience.

More to come as the library is built out.

Cloud Plots - Wisemoor Free Formulas

Cloud Plots - Fraterworks Free Formulas