Quality Testing for Tea, Electronic Noses and Such

Greetings! I work in IT and usually post about that--just not frequently--but I'm researching a subject related to my hobby, tea reviewing. A US tea producer mentioned that people work on developing quality testing by machines or chemical processes that eventually will take some of the place of taste testing by a human. Sounds horrible, doesn't it?

He was positive about it, for one main reason: it would be objective. As things stand in a developed industry like China's there really can be some consistency in agreement on quality level / grading of teas but there would always be some degree of looseness related to evaluations. Part would relate to people exaggerating claims about quality, and there being no way to really evaluate those. I'm talking about taste here, although issues related to contaminants do come up.

How could an "electronic nose" or lab tests even conceivably test the taste of tea? At a guess there are a few levels related to that as possible answers, beyond saying it wouldn't be possible. I'm asking for input about it here, not just explaining it, but I get it that's a longshot. There are a few main categories of compound types in teas, and it would relatively easy to test for concentrations of those as a set. The larger problem is that each general type is expressed as lots of individual chemical compounds (with caffeine as something of an exception--that's one thing), the distribution of which would be the main input to tea flavor. Testing for hundreds of individual compounds grouped as several main types would be problematic. Testing what is in a tea that shouldn't be, essentially doing what our tongues do, to taste it all in one go, wouldn't be so simple.

People really are working on this though. Any input for where to start looking, beyond Google search and Research Gate review (which I checked; there is content out there, most of which I can't get to without a subscription). In case anyone is really curious I'll pass on the best site reference I've turned up so far:

http://www.upasitearesearch.org/tea-quality-parameters/

I know this isn't new with regard to many foods, which includes drinks and such. I did a quick Google search (using key words electronic taste tests) and see there are wine tasting machines, and I know that I have heard about such electronic taste tests over the years. As far as I know, though, human taste testing will be around for quite a while.

Interesting topic, though.

I found this interesting: The Five Senses of Sensors - Taste - 2011 By Carolyn Mathas

That reference was interesting, thanks. It ties into the one complication of duplicating taste, that really flavor is mostly carried by sensation that occurs in the nose, not the tongue. Only basic flavor inputs / tastes are picked up there, the ones they mention developing a sensor for (sweet, salty, sour, bitter, umami / savory).

There is a really easy way to isolate which you are tasting in reference to a certain food, a taste sensation picked up by the tongue or a flavor sensation being identified by receptors in the nasal passages: plug your nose while you taste it. A basic flavor like saltiness or sweetness won't change at all but everything else completely drops out, like approximating having a cold instantly. Tea without sugar can seem sweet based on some compounds (amino acids, if I'm remembering right) approximating the effect of sugars but my understanding is that the actual sugar content is negligible.

I'm not looking forward to tasting being replaced by machines, just looking into the subject as one more thing to consider, something someone else brought up. Research is being done on making it happen, it's probably just like Artificial Intelligence research, small steps will happen for a long time. Again I'm here to ask for leads but I'll share another paper reference and abstract, which may be interesting (all available at this citation, but the paper itself is not):


http://www.sciencedirect.com/science/article/pii/S0925400503003678

Tea quality prediction using a tin oxide-based electronic nose: an artificial intelligence approach

In this paper, we have (analyzed using a metal oxide sensor (MOS)-based electronic nose (EN)) five tea samples with different qualities, namely, drier month, drier month again over-fired, well-fermented normal fired in oven, well-fermented over-fired in oven, and under-fermented normal fired in oven. The flavour of tea is determined mainly by its taste and smell, which are determined by hundreds of volatile organic compounds (VOC) and non-volatile organic compounds present in tea. Tea flavour is traditionally measured through the use of a combination of conventional analytical instrumentation and human organoleptic profiling panels. These methods are expensive in terms of for example time and labour. The methods are also inaccurate because of a lack of either sensitivity or quantitative information. In this paper an investigation has been made to determine the flavours of different tea samples using an EN and thus to explore the possibility of replacing existing analytical and profiling panel methods. The technique uses an array of four MOSs, each of, which has an electrical resistance that has partial sensitivity to the headspace of tea. The signals from the sensor array are then conditioned by suitable interface circuitry resulting in our tea data-set. The data were processed using principal component analysis (PCA), fuzzy C means (FCM) algorithm. The data were then analyzed following the neural network paradigms, following the self-organizing map (SOM) method along with radial basis function (RBF) network and probabilistic neural network (PNN) classifier. Using FCM and SOM feature extraction techniques along with RBF neural network, we achieved 100% correct classification for the five different tea samples, each of which have different qualities. These results prove that our EN is capable of discriminating between the flavours of teas manufactured under different processing conditions, viz. over-fermented, over-fired, under-fermented, etc.

I'm going to know a lot more about this subject by the time this article is finished but I wanted to pass on a reference I've found while still thinking about it:

Instrumental Assessment of Food Sensory Quality: A Practical Guide

Fascinating stuff, just a bit dry. I'm the opposite of an expert but I'll pass on an early first impression. It turns out that sensors can definitely smell or taste food already, they're just not quite as good at learning what the taste / smells mean. If it's as simple as testing for one compound that identifies a product is "off" then the equipment can work well. It's just that for checking if an apple pie tastes good it's more complicated than that. It's better for checking for spoilage, since that would give off a distinctive smell.

There are lots of issues related to process. We smell foods by giving it a sniff, but for tasting that isn't exactly how it works. We pick up taste on the tongue (sweetness, saltiness, bitterness, sourness, umami) but most flavor is picked up in sensors in the nasal passages. Mixing the food with saliva, grinding it up, warming it slightly, and allowing limited air exchange inside our nasal passages are all required. Machine sensors don't go through all that but different versions do work differently, and only some approximate "giving it a whiff." Others heat the test material quite a bit, or use some form of solution to convert the material to liquid and then evaporate it into gas (what we do).

It's not so simple to read what that machine tested, and there are lots of variations in possible results related to using different equipment, different process methodology, and data review. It might not ever be so simple to "machine taste" something and put it on a scale of very bad to very good. On the upside, if you could pin down an automated process for doing that for one certain food you could repeat it easily and consistently. I'm still not clear at all on how it works to taste for what shouldn't be there, or variations in secondary component levels, although it's hard enough to taste for what should be.

I write tea reviews, as a hobby, so it's not as if I'm looking forward to a day when a machine could do it better. It's just interesting since once you get far enough along this learning curve it starts to tell you more about what a person is doing. Much, much further along a machine can do part of it. Of course people are clever, so it won't be so long before people aren't tripping over looking for one specific compound to test for, or variations of one, or simple combinations, and they can get machines programmed to more or less take it all in. They can work backwards from having a machine and a person both test the same samples, and one interesting section in this book talks about that. The issue is that every flavor isn't either a good or bad input (present or not present value), or with a certain level that's ideal, too much, or not enough, it all has to balance.

I'll check back in once it gets further along.

Do you have any info on how they did "scratch and sniff" for things other than perfumes in magazines?

Absolutely no idea, really. I'm just researching this subject for a personal blog post, and I really work in IT, I just do tea tasting as part of a blog hobby.

I'll guess a little though; I have been reading about scents and mechanical noses and scent profile evaluation some lately. Part of that could relate to how they produce essential oils out of flowers and whatever else, I think more or less by using alcohol as a base to extract and retain the volatile component that is the scent. But I suppose it's more complicated than that.

It's an odd thought, isn't it, to soak a hamburger in alcohol to get a hamburger essential oil scent? How that works related to the part you scratch is also curious, how to get it to come off, is beyond what I might even guess about. Taking the ideas in a different direction, different compounds relate to different aromas or flavors, and I guess the right experienced person could mix chemicals to simulate a wide range of scents:

https://en.wikipedia.org/wiki/Aroma_compound