Implementing A New Fragrance Comparison System: Charting Similarities And Differences With Scent Prisms

Occasionally it becomes difficult to verbally convey my sense of similarities and differences between two fragrances, and as with all modes of communication, written language eventually reaches its limitations. Fortunately I have a good understanding of top-down fragrance structures, and can translate top, middle, and base notes into vertical bars, with each note assigned a horizontal color. Where this falls short is in the subjectivity of perception - for example, I may smell coconut in the top notes of something, while someone else smells it in the heart notes. Therefore my new Scent Prisms are strictly visual expressions of how I perceive a fragrance's structure, and cannot be standardized. Nevertheless, by using notes voted upon in forums like Fragrantica and basenotes, with the addition of notes that I subjectively perceive apart from public opinion, a fairly accurate Scent Prism can now be built for any fragrance. I've provided two examples of these prisms here, and will explain them.

I figured an easy way to learn how to use these prisms is to provide a comparison chart of two familiar fragrances, Green Irish Tweed and Cool Water. I've often compared the two, and indeed I find their fragrance structures compositionally similar. I have always acknowledged that the two fragrances are different, however. In the scheme of Cool Water comparisons, with dozens of masculines being compared to this aromatic fougère, there is little doubt that Green Irish Tweed is the closest anyone will get to wearing something directly similar to Cool Water, and vice-versa. Below are the two Scent Prisms for these classics - Please click on image to enlarge for detail:

As you can see, these fragrances are different in a few ways, but their general structures contain several similarities. As an aside, I should mention that I don't think there's a real lavender note in GIT, but there is linalool listed on its ingredients, and it is a fougère, so I suppose there could be a small dose, highly blended, in this fragrance. If we concede that there's a touch of lavender, that means both fragrances contain lavender in different proportions, and also citrus, green apple, violet, violet leaf, amber, and sandalwood. You may have noticed that the violet leaf in GIT is a wide band with a soft look, while the same note in CW is narrower and harder-edged. That's because GIT, being a simpler composition, relies heavily on octin esters for its violet/violet leaf effect, and that effect has a broad impact on how the fragrance smells - dense, a little fuzzy, and sweet. It comprises most of GIT's heart.

Cool Water's violet leaf is markedly less intense, but also sharper, crisper, and shorter-lived on skin. Where GIT gets about two hours of violet leaf, CW only gets around forty-five minutes. But CW is a more complex structure, boasting twelve detectable notes to Creed's eight. You can see the main similarity between the fragrances in their citrus-violet-violet leaf accords, which are sandwiched evenly between varying degrees of lavender and amber. GIT's amber is very full, while CW's is a "mini amber" that resides close to the base, and isn't alone in sweetening the structure. The additional notes of jasmine, tobacco, cedar, and musk conspire to make wearing CW a busier, multi-faceted experience. GIT is broad strokes, with fewer movements from top to bottom. Take into consideration that GIT's "muskiness" is largely attributable to ambergris, which is an animalic note comprising its base. This parallels CW's less exotic white musk base.

Because this chart compares two similar fragrances, I thought I'd throw in a comparison of two fragrances that are more different than they are similar. Below we have Scent Prisms for Dali's Laguna and Chanel's Egoïste:

Given that these two share strong oriental qualities, such as vanillic ambers and rich wood notes, there are some similarities between them. They both contain facsimiles of rosewood and sandalwood, and two variants of stone fruit notes. Laguna has separable notes of peach and plum, which are sweet. Egoïste has an inseparable "stewed fruit" accord that occurs fairly early on in the wearing experience with the most recent formulation. It isn't an overbearing accord, but it feels like a few fruit notes, perhaps cooked plum, apricot, and even dates, with a dense sweetness that snakes through the other notes. Laguna winds up smelling more floral, and much fresher thanks to galbanum and watery notes of grapefruit, pineapple, and coconut. The vanilla in Laguna smells more floral to me, so I have subjectively labeled it vanilla flower, while the vanilla in Egoïste is traditional and almost edible. Both fragrances have prominent sandalwood notes, but sandalwood comprises Egoïste's base, while a salty musk is in the bulk of Laguna's. Both have roughly the same degree of rosewood in them to my nose, although Laguna's is touched with tonka and smells a bit sweeter, and is not as definite as the rosewood in Egoïste.

What sets these two widely apart is the types of florals, the degree to which they're used, and Chanel's implementation of herbal notes. Laguna contains one prominent herbal note of patchouli. Linalool is loosely perceptible in its top accord, but it never settles into a true lavender note, and so lavender was excluded from Laguna's Scent Prism. Egoïste does have detectable lavender, along with thyme and several other herbs. Its herbal accord happens sharply, and settles itself in the heart of the fragrance. It is sandwiched between a massive rosy-carnation accord, and an equally-massive tobacco-vanilla-amber base. The top of Egoïste boasts a rich cinnamon note, tinged with lemon citrus, while the top of Laguna is very fruity and has no cinnamon.

I won't use these Scent Prisms very often, and people are bound to disagree with them. I figure they'll come in handy when a comparison is acutely difficult to make, and only a picture will do. When fragrances are more similar than they are dissimilar, you will see similar color progressions in their bars, but when fragrances are different, there will be fewer color consistencies between them. Hopefully these prisms will help people understand my comparisons, when the issue of making comparisons crops up here and elsewhere in the fragrance blogosphere.


  1. I love this, Bryan--how fantastic! I rarely find so many layers and subtleties in a perfume. I tend to get swept away by general impressions and then I start thinking about other things. Sometimes those other things are perfumes, but ... (ditto the above) ...


  2. This is a great idea, especially the soft/hard edges to denote drift, and the size of the section indicating strength. There's a guide in this!

    1. Thanks Stephen! Maybe there is a little guide to be made out of these - I'll be including them from now on.

  3. Hi, I also often think about scents in terms of colors and I very much enjoyed taking part of your "prism" way to explain scents. A question, do you have an automated way to create these prisms or do you create the images by hand after looking up notes and number of votes on Fragrantica? I'm just curious, it would be neat to have this done by automation :)

    1. Hi flavourfanatic, thanks for reading and welcome to my blog!

      The prisms are done by hand, but they are not really created via note statistics, and not strictly done by votes. I seriously consider votes on Fragrantica when creating them, but I'd say that's about 20% of the process. The remaining 80% comes directly from personal experience in repeatedly wearing the fragrances that I am making prisms of.

      There is also an aspect of realism that I try to inject into the process, and I don't know how much of a fraction of a percentage that takes up, but here's a couple of examples: GIT and Halston Z14. I haven't presented Z14's prism yet, and will soon. Many report smelling gardenia in Z14. I smell in an older formula a very shy and fleeting impression of one or more white flowers, but could never in a million years say with any confidence that it's actually gardenia. Also, because gardenia extracts are very low-yield and are almost never used in perfumery, the chances of there being a true gardenia material in Z14 are nil. Therefore, despite it being a talked-about note, gardenia is not something that I think should appear in a prism.

      Likewise with the supposed iris in GIT. I myself smell an analog of iris in that one, and so do many others. But because real iris butter (and any other true iris material) is extremely expensive and unlikely to appear in a mainstreamed niche oldie for men, iris is excluded from the prism. I chalk the iris effect up to the interplay between the violet notes and the octyn esters, which also create a very grey, eerie effect, akin to iris.

      I don't know how to do prisms via automation, but agree it would be interesting to try.

    2. I understand you put a huge amount of knowledge and experience into those prisms, and just generating them by data from Fragrantica would not be the same thing!

      Still, I work as a programmer so I am intrigued by doing thigs automatically. If nothing else, you'd have a picture to start working from. It can be done by downloading the page code of a certain scent from Fragrantica and then get the notes/votes by parsing that code. You would also need a database table to keep track of the color codes of the colors you use to represent each note. With that information a picture can be generated. I'm not so good with pictures so I'm not sure which way would be the best to do the generation or what image formats to use, but if I figure it out, I'll be sure to tell you about it :)

    3. That's interesting, I didn't know you could do that using existing internet pages. Do go for it and get back to me. The best way to assemble the result is to shoot for either a pyramid or a rectangular shape, like the ones in this article. Fragrances are top-down design, after all ;)


Thank you for your comment. It will be visible after approval by the moderator.