A high note for bergamot, a low note for birch tar. In a 1997 study, people adjusted the pitch of a tone to match a smell, and systematic differences emerged between fragrance materials.
Interestingly, changing the first tone they heard shifted the pitches they selected considerably. Even so, there were similarities in which smells corresponded to relatively high or low notes. The researchers explored this correspondence as a way of putting numbers to odor quality, something that can be difficult to describe in words.
Matching odor intensity and sound loudness
The study used 20 fragrance materials spanning floral, citrus, woody, earthy, and other odor qualities. They included natural extracts such as essential oils as well as synthetic materials, ranging from familiar smells to substances whose names people rarely encounter.
In preliminary tests with a separate group of assessors, odor intensities were adjusted against a reference of 20% olibanum. Concentrations were changed and checked until the mean rating fell within 1 standard deviation of the reference mean. This was intended to minimize influences such as choosing a higher tone simply because an odor smelled stronger.
The sounds were simple sine-wave tones lasting 1 second, delivered through headphones. Because perceived loudness varies with frequency, the researchers used existing auditory data to make corrections intended to equalize loudness. The task focused on pitch, rather than on choosing a musical style or an instrument’s timbre.
32 people look for a tone that fits each smell
The first experiment involved 32 participants aged 19–40. They were recruited through newspaper advertisements and screened for their sense of smell and nasal condition. Professional musicians, hearing-aid users, and people reporting hearing loss were excluded.
After smelling a material, participants heard a tone. If it did not match the smell, they told the experimenter to make it higher or lower. Adjustments began in steps of 1 octave; whenever the requested direction reversed, the step became smaller, first half an octave, then a quarter of an octave, and so on. The point at which the participant judged the tone to fit was taken as the pitch corresponding to that smell.
Each person attended two sessions on separate days, lasting about 1 hour each. On one day, matching for every odor began at 200 Hz; on the other, it began at 1,000 Hz. Session order was counterbalanced across participants, and odor order was randomized. The sessions were an average of 7 days apart, with a minimum interval of 2 days.
Bergamot tends higher, birch tar lower
The selected pitches differed significantly between fragrance materials. Bergamot, for example, fell toward the relatively high end, and birch tar toward the low end. Rather than having to name a smell, participants could express differences using another sensory dimension: high and low pitch.
However, the initial tone also had a large effect. For all 20 materials, the mean pitch on the day starting at 1,000 Hz was higher than on the day starting at 200 Hz. The average difference across materials amounted to a frequency ratio of about 2.27, or nearly 1 octave.
Nevertheless, the correlation between the two conditions’ mean log-transformed frequencies for the 20 materials was r = 0.80; the rank correlation was 0.66. The central finding was that, although starting pitch affected absolute pitch, there was some consistency in how the odors were ordered from low to high.
These correlations compare mean values for each fragrance material. They do not measure how precisely an individual could select the same tone again. Because the same participants evaluated the same materials a second time, memory of their earlier judgments could also have contributed to the consistency. The authors noted this possibility in a footnote.
Can it all be explained by liking a smell more and choosing a higher note?
The second experiment examined whether pitch was simply a substitute for odor pleasantness. A separate group of 29 people aged 22–40 rated the pleasantness of the same 20 materials on a scale running from very unpleasant through neutral to very pleasant. They rated every material twice, and the mean values were analyzed.
The correlations between pleasantness ratings and the pitches selected in the first experiment were not significant. The correlation was 0.22 for the 200 Hz starting condition and 0.10 for the 1,000 Hz condition; rank-based analyses likewise showed no clear correspondence.
The researchers interpreted this as making it unlikely that participants had simply converted liking into pitch. With odor intensity brought closer together beforehand, and no simple association with pleasantness found, pitch matching becomes a candidate for capturing another aspect of odor quality. However, these comparisons used averages from separate groups, so they do not establish that pleasantness plays no part at all in an individual’s judgment.
A possible new yardstick for describing smells
The authors discussed whether pitch might correspond to as-yet poorly characterized qualities through which smells seem heavy or light, bright or dark, or hard or soft. This experiment did not identify which qualities actually produce the correspondence.
The study’s contribution was to show that people without specialist vocabulary could quantify differences between smells by comparing them with another sense. The authors also proposed exploring relationships with timbre, rather than pitch alone, as a next step.
The participants were not selected for having synesthesia in which smells spontaneously evoke sounds. They smelled materials while adjusting tones that were actually presented. Rather than assigning a universal frequency to each fragrance material, it is therefore more appropriate to read the finding as a shared tendency in relative pitch within this group and this task.
Summary
When participants matched tones to 20 smells, the selected pitches differed by material, with bergamot tending higher and birch tar lower. Starting with a higher tone shifted the responses upward overall, but the ordering of the materials by pitch was correlated across conditions.
Pleasantness scores from a separate group did not correspond significantly to pitch. Comparing smells with sounds may help explore qualities that intensity and liking alone do not readily capture. The study demonstrated both the potential of describing odors without words and a measurement characteristic that matters: the influence of the starting point.
Research context and disclosures: At the time of the study, all four authors were affiliated with Givaudan-Roure Fragrances. The correspondence details in the published paper also list Synesthetics, Inc. as Gilbert’s affiliation. The acknowledgments name people who helped with the experiments, but no separate funding or conflict-of-interest statement could be identified.
Reference
Belkin, K., Martin, R., Kemp, S. E., & Gilbert, A. N. (1997). Auditory Pitch as a Perceptual Analogue to Odor Quality. Psychological Science, 8(4), 340–342. Original paper / Author-hosted PDF
