Visualizing Perfume
Many of the perfumes in my collection are personal markers. They call to mind past moments or into imagined worlds. A few drops and I’m elsewhere. For years I’ve kept a perfume journal, trying to turn fleeting impressions into words, but language never quite caught the nuance. Images seemed more immediate, so I tried translating scent into visuals using Bing’s DALL·E3.
The problem: prompts still start as words, carrying the same limits. The results were mixed—unexpected, sometimes absurd, often fun. I learned how AI “reads” smell, with quirks like turning any spice into star anise. And some notes—oakmoss’s inkiness, galbanum’s bitter green, iris’s earthy root—just don’t land in pictures. They live in a vocabulary I’ve borrowed over years from other scent obsessives. For inspiration in this endeavor, I owe a large debt to Luca Turin and Tania Sanchez (perfume guides), Victoria Frolova (Bois de Jasmin), Katie Puckrick (KatiePuckrickSmells), Eugene Nizic (USmellsGood), Michael Edwards (Fragrances of the World) and Kafkaesque whose work have shaped the way I understand perfumes over the years, and whose influences have, directly or indirectly, made their way into these AI visualizations. I tried to credit any direct quotes or references. Each image below comes with a short caption about the perfume referenced. They are generally ordered by fragrance families (but some placed emotionally).
