Ever wished you could have your broccoli taste like macaroni and cheese, watch movies in true Smell-O-Vision, and be able to quickly neutralize any bad smell you’ve ever come across? A new study by researchers at IBM and the University of Illinois could help translate this, and much more, into reality.

The information age has greatly advanced our ability to expertly manipulate signals in either visual or auditory form. But what about smells? Brothers Lav and Kush Varshney have set out to fill the gap, developing a framework that could be used to bring the sense of smell into everyday technology.

While the electromagnetic waves that reach our eyes and ears can be characterized by only three parameters (amplitude, phase and frequency), olfactory signals tend to be much more complex. Smells aren’t waves, but rather collections of molecules: an attempt to break them down to their fundamental components results in a huge, messy and hard-to-manage catalog of thousands upon thousands of different chemical compounds.

There is however some good news. Recent studies suggest that while the molecules involved in smell are many, the space of human olfactory perception is fairly limited and can be predicted from the chemical properties of the molecules. Also, other findings show that it is possible to create a neutral "olfactory white" (the smell equivalent of white light and white noise) for any given smell.

This means that it is theoretically possible to build a machine that analyzes the chemical composition of the surrounding air and then releases the appropriate cocktail of chemicals that neutralizes the detected smell, whatever it might be - a sort of noise-canceling headphones for your nose. Moreover, because our sense of smell is relatively unsophisticated, we could achieve this using only a small number of chemical compounds.

The perceived smells of four selected foods, before and after "active odor cancellation" (Image: L. and K. Varshney)

Making use of this knowledge, the Varshney brothers set out to map the olfactory perceptions of a number of different foods. They first had subjects describe on a scale from -5 to 10 how foods like onion or sauerkraut matched a series of 146 perceptual descriptors (ranging from "fragrant" and "cheesy" to "sickening" and "cat urine"). Then, using only edible chemicals, they calculated the perfect mixture that would neutralize their smell on each of those dimensions. The before and after results are shown above.

The scientists concluded that it is within our current technical and engineering capabilities to design a "white smell machine" by combining an electronic nose and an actuator like the oPhone that would release chemicals from a small set of cartridges to neutralize any smell.

Because smell strongly affects our sense of taste, the same techniques could be adapted to create an additive that hides the flavor of one unpleasant but healthy food (such as broccoli or cauliflower) inside the flavor of another, more pleasant food (say, lasagne) and therefore encourages people to eat healthier.

As a proof of concept, the scientists analyzed the chemical composition of broccoli and then used a comprehensive food database to design an additive that can completely mask its flavor. The idea here is that the combination of broccoli, the masking additive and any pleasant food of choice will be perceived as having only the pleasant food’s flavor. It turns out that if you want your kids to eat their broccoli, all you need to do is add a lot of Angelica oil and cumin seeds, barley, lobster, and a dash of brandy.

The purpose of the study was to lay the groundwork for future research and illustrate the range of possible applications  –  which, apart from the two explained here, could also include improving indoor air quality and creating a more interactive entertainment or virtual reality system.

It’s too early to tell how soon a practical "white smell machine" might become a reality, but we’ll be waiting.

Source: ArXiv (PDF)

View gallery - 4 images