Imagine a handheld device that can scan your food to deliver accurate, real-time data about its chemical and nutritional content—including calories and potential allergens.
That's the vision cooked up by TellSpec, a Canadian startup that claims to be developing the world’s first “food scanner.”
Using Raman spectroscopy, the TellSpec fires a laser into a given food item, measures the reflected light, and then sends the data to TellSpec’s cloud servers. There, the information is measured against reference data and run through a series of algorithms. In the final step, the results are sent back to the user and displayed via a mobile app. The whole process takes about three seconds.
Last week, TellSpec announced it has integrated a Texas Instruments DLP (digital light processing) device into its latest prototype. This follows last fall’s Indiegogo campaign, which saw the company raise more than $386,000 in funding.
But is it real?
If you’re still skeptical, good job: It means you have your thinking cap on. There’s a lot to question here, and thus far the online response to the TellSpec has been one of resounding doubt. Some commenters have gone so far as to claim it’s an outright scam.
Perhaps the biggest head-scratcher is how you can determine the chemical makeup of a food item by simply measuring the light it emits. Most Raman spectrometers—which observe vibrational, rotational, and other low-frequency light emissions—are used for testing product purity, since they provide what are essentially fingerprints of compounds and molecules. Without context or reference data to measure those fingerprints against, you really have no idea what you're observing.
This is where TellSpec’s cloud servers and algorithms come into play. They provide the “reference spectra” against which the data is measured—sort of like a police unit's fingerprint database.
I asked Reviewed.com’s own Dr. Timur Senguen to elaborate on this point. He explained that Raman spectroscopy is indeed an “incredibly powerful, low-energy kind of spectroscopy that is really good at determining the identity of organics.”
As long as the compounds that the TellSpec analyzes have been previously classified, he adds, “the scanner will be able to tell you what is in it. So qualitatively, this method should work great in theory.”
Okay, so the basic concept seems to check out. But what about the reliability of the data it provides? Not all foods have a consistent chemical or nutritional makeup—in fact, most prepared foods don't. Logic suggests that the TellSpec could only analyze a precise segment of a given food—the tomato in your sandwich, as opposed to the sandwich as a whole. Dr. Senguen agrees:
“Light from the laser will not penetrate the object, so all you're getting are readings from the surface. That's not a problem when you're dealing with homogenous food like oatmeal, but in the case of a birthday cake, nine-layer bean dip, or burrito, I doubt you could as easily just point the scanner at the food and get a detailed breakdown. You might have to dig into the food and mush it up, or scan each layer individually.”
Indeed, TellSpec's demonstrations and product photos all depict homogenous foods like bread, chips, and crackers.
There’s also the price: The TellSpec is currently on pre-order for a very attainable $320. Sure, that's a decent chunk of change, but Raman spectrometers currently on the market are much bigger and often cost tens of thousands of dollars, making the TellSpec's price seem suspiciously low. It’s not unthinkable that an innovative company could bring the cost down that much—especially if the data is analyzed by remote—but it would certainly be quite a feat.
Adding to the skepticism is a series of questionable marketing moves by TellSpec. Their Indiegogo campaign last fall, for example, raised funds through the “flexible funding” option. That option means a given project gets to keep all funds raised, regardless of whether the stated funding goal is reached. In an interview with Shoplocket, CEO and co-founder Isabel Hoffman explained that this was a mistake.
“Everybody thought we were a scam—that we just wanted to take the money and run with it. We could never change that. You really shouldn't make your funding goal flexible.”
Promotional videos of the TellSpec in action have been similarly misleading: They depict unfinished prototypes in a manner that appears real. In fact, the video was merely a simulation of the intended use.
PR mistakes aside, TellSpec appears to be a legitimate project. The folks behind the it are real, with real credentials that legitimize their ambitions. The underlying science seems fairly solid, and the price is shockingly low—perhaps due to the centralization of TellSpec’s data analysis. The only question left to be answered is whether the team can pull it off as a commercial product.
“Raman is great, and the fact that they've managed to turn a cabinet-sized, multi-thousand-dollar instrument into a handheld consumer-level device is truly amazing,” Senguen says. “This might be the real deal, as long as they can successfully instruct the user in the proper use.”