Posted Tue Sep 7, 2010, 9:06am Subject: Effect of variables on coffee taste perception
Hello CoffeeGeeks
This is possibly one of the most coffeeiest, geekiest things I have ever done, and I need your help.
I am putting together a comprehensive list of variables relating to the whole area of coffee, in order to produce an 'espresso diagnosis' chart.
I have big plans, and hope to eventually collate the information, with references, into a web application for growers, home roasters, professional coffee roasters working on different blends, and those just enjoying the cup.
I'm inspired by the excellent book "Espresso Coffee: The Science of Quality" (link here). It's a fabulous read and I can recommend this book to those serious about the art and science of coffee.
With a view to this, I'd like to get suggestions from the community (aka here!) on the following areas and variables.
More to the point, if anyone likes the sound of this, and wants to get involved in the coming months, or can spare some time, I need to hear from you! Right now I need people to help me rationalise the ideas, lists, variables, etc. So - thinkers!
Please note, only the variables that MAY have a quantifiable and consistent effect on the quality or taste perception of espresso coffee are to be included. I say "MAY" because of course taste perception can be subjective. But this is an inclusive project. In some cases, the quantifiable factor may be something other than "effect on taste", for example, "effect on bean hardness", but hopefully taste perception can be derived from this.
The below is a draft!
Coffee growing -------------------------------- - Effect of altitude - Effect of amount / intensity of sunshine and rain per year - Effect of drying method (e.g. sun-dried) - Effect of drying time
Coffee storage and distribution -------------------------------- - Effect of average ambient storage temperature - Effect of humidity - Effect of total time stored before roasting
Coffee bean roasting -------------------------------- - Profiling - Effect of increasing or decreasing temperature at a different point in the roast cycle (complex) - Effect of temperature scale applied to profile (i.e. slow or fast heating) - Effect of heat application type (to be further specified) - drum or air - Effect of drum materials (e.g. does a given metal or surface produce a faster roast, produce a more mellow roast, etc.) - Effect of agitation of beans - Effect of smoke dispersal, heat dissapation, etc. - Effect of speed of cooling - Bean types - Effect of size (surface area to weight ratio) of green bean on time taken to first crack (assuming all other variables the same) - Effect of dryness of green bean on time taken to first crack (e.g. percent of bean that is water) - Effect of other quantifiable aspects of green bean on time to first crack
Grinding -------------------------------- - Effect of consistency of grind size upon taste and/or upon extraction volume and/or upon crema - Effect of volume to surface area ratio of grain upon above variables - Effect of uniformity of shape of grain upon above variables -
Milk -------------------------------- (rough notes) - Different types / viscosity of milk - Freshness of milk - Temperature of milk - Temperature ramp of frothing - ... and all the other variables that people inadvertently refer to when describing exactly how they froth milk for the perfect latte art...
Tamping -------------------------------- - The book states the difference on extraction between a hard tamp and a soft tamp is minimal compared to the difference between no tamp and a soft tamp. Effect of tamp pressure on volume of extraction for a given particle size. This must be measured for a given filter basket (pressure is not measured in kilograms or lbs!) - Effect of tamping surface shape (i.e. convex or flat) on taste and extraction rate / voume - Effect of using a tamper with incorrect diameter (i.e. effect of consistency of density throughout puck AND/OR effect of consistency of surface)
Espresso machine architecture -------------------------------- - Water - Effect of temperature on extraction - Pressure - Comparison of two different machines with different maximum pressure passing same coffee beans through each with 30lbs tamp, 25s extraction; the effect on volume of extraction (and taste...) - Comparison of two different machines with different maximum pressure passing same coffee beans through, where both have been calibrated to the 30lb/25s/50ml rule. - Troubleshooting - Effect of wear and tear of following parts on taste: - Gasket, etc. etc.
Espresso Preparation -------------------------------- - Effect of age of beans on taste variables - Effect of age of beans on extraction volume (assuming constant tamping, extraction time, etc.) - Effect of grind particle size AND uniformity on taste variables - Effect of grind particle size AND uniformity on extraction volumes - A TONNE more.
Cupping variables -------------------------------- The below list combines variables and quantifiables. It's the 'results' end. Needs formalising!
It would be amazing to be able to MAP some of the above onto the below aspects of cupping.
- Bitterness / sourness - Brightness / mellowness ... other olfactory / taste sense descriptions (there are so many different views on how to correctly describe coffee taste, I wish there was an open consensus on the subject - this is something to discuss!)
- Thickness of crema - Colour of crema - Pattern of crema (tiger pattern good, white spots bad) - Density of crema, including bubble size - Viscosity of crema - Volume of extraction - Viscosity of extraction
-------------------------------- -------------------------------- ... My head hurt after getting this far. Please help me out!
The web app could potentially benefit a tonne of people from home users to those working in marketing a product. Many of the variables would come as second nature to a barista, but not a grower, and vice versa. Most would benefit home users.
The most complex application would be to be able to plug in a question at the beginning, and get a sensible range of answers i.e. "I am a home roaster. My coffee tastes bitter" -> "check the gasket, check the age of the roast", OR "I blend beans, however my blend darkens too quickly" -> "verify your beans are at a dryness of at least x percent water" etc.
The above list is a draft. The ideas need formalising. And of course, as you can imagine, there's probably no WAY a single researcher could complete this project.
But it's only right to get the idea out in the open at this point :)
Posted Tue Sep 7, 2010, 6:27pm Subject: Re: Effect of variables on coffee taste perception
Whoa. This is going to be a huge undertaking! I don't envy you or those helping you as you try to organize that information. Get ready for lots of flow charts lol!
In all seriousness, though, be sure you're prepared with people who've done this sort of thing before. As a student, I was tasked once with creating a simple mobile phone mock-up, including only a handful of various states (make a call, receive a call, a maps application, a reader application, etc.). I used half a ream of printer paper just making all the flow charts for the interactions of what was really a fairly simple device. Our professor used the example of WebMD's Symptom Checker as the harsh reality of web development: they have to decipher something like "my stomach hurts," and give you real, possible medical causes for that perception. The statement alone could mean anything from indigestion or gas to ulcers or cancer, or even something that is only perceived, not real. Apply that scope to the entire body, and you have a product that took forever to realize, is in need of constant maintenance and revision, and, in the end, is often still inferior to a professional opinion. Espresso is going to be the same way, I would think. Now, it would definitely be nice to have a compendium of sorts, but as a "diagnosis chart?"
Bitter shot, standard dose, standard volume: See Overextraction, Water Temperature, Water Impurities, Blend Composition, Dark Roast, Supertasters, What You Ate Three Minutes Ago...
All of the above are real, possible causes for a bitter shot, and the last one is only partly in jest. There's sometimes more to the end product's taste than what goes into the liquid in the cup. Taste is very subjective, of course, and can be affected by an almost incalculable amount of variables. From a diagnosis standpoint, you're looking at heaps of possibilities in many cases, for common problems. Have you started looking at problems and possible causes, rather than total amount of typical variables?
Now, if the Symptom Checker can be done, I don't doubt that this could be either, but damn is it going to be a lot of work! Good luck!
bodum_fanatic Senior Member Joined: 19 Dec 2006 Posts: 478 Location: Missouri Expertise: I live coffee
Espresso: Won't become one of those... Grinder: Braun KMM30, Krups Fast... Vac Pot: Never had much luck there. Drip: Melitta Deluxe 132 made in... Roaster: West Bend Poppery II
Posted Wed Sep 8, 2010, 5:34am Subject: Re: Effect of variables on coffee taste perception
Zin has a very good point. The chemical compounds in coffee can be effected by many different variables. People's subjective taste can also be effected by different variables, including psychological. You could serve any number of people the same cup of coffee, telling them each something different that was supposedly done during the roast/brew/etc. (even though there was no change), and get different reactions merely based upon the perception that something was different. Serve twenty people the highest rated, freshest brewed coffee, and tell them it's Maxwell House, and see how many of them tell you that it's flavors are muddy and stale.
Posted Wed Sep 8, 2010, 7:05am Subject: Re: Effect of variables on coffee taste perception
SteveRhinehart Said:
Bitter shot, standard dose, standard volume: See Overextraction, Water Temperature, Water Impurities, Blend Composition, Dark Roast, Supertasters, What You Ate Three Minutes Ago...
I know it seems very complex the way I have presented it. I mean, even quantifying the effect of the core variables (tamp pressure, grind size, dose, volume) requires some crazy multidimensional mapping due to the dependences of one variable on another. But I think I would like to start at least *somewhere*. I would like to produce *something* like a model, even if it's wildly inaccurate to start with. As an analogy, computer games nowadays are able to beautifully model the movement of a human being walking in a straight line, the flexing and contracting of muscles during movement, the internal accelerations within various parts of the body, whereas 15 years ago we just had sprites on the screen that moved in a given direction, or fell down with the acceleration of gravity when the floor ran out. Although that old model of a living creature was inaccurate, the basics (gravity, falling, walking left and right) were there, and they provided us with something to build on.
SteveRhinehart Said:
Apply that scope to the entire body, and you have a product that took forever to realize, is in need of constant maintenance and revision, and, in the end, is often still inferior to a professional opinion
Indeed. I need to give more thought to the different ways of presenting the data. A diagnosis chart would only be suitable for basic troubleshooting as opposed to advanced tasting. WebMD may not be a useful tool for doctors and brain surgeons, but it could help patients to differentiate between something they should just take a headache pill for, and something that they really should go seek medical advice for. That in itself is valuable.
And yes I agree that the diagnosis application is a ridiculously big task to rationalise. A system (e.g. the body, coffee) can't easily be reduced to a set of variables.
On reflection, instead of starting from first principles, it may be better to start with a hit-list of common faults with espresso brewing, and work out from there, improving and quantifying the various models where possible. Then introducing more people from different fields as the project progresses.
bodum_fanatic Said:
People's subjective taste can also be effected by different variables, including psychological.
That's the very thing I would like to investigate, in order to come up with some kind of commonality.
On this subject, firstly, many of the experiments would be working with quantifiables, not taste perception. We already know how to describe the taste of an underextracted or overextracted espresso, and therefore we can derive the difference in taste between grinding too fine or too coarse, from the amount of under or overextraction. Therefore we can say "harder tamps produce slower extraction, which produces a more bitter taste". Or the like.
Secondly, of course, getting two independent tasters to investigate the benefits of tamping harder would be a futile task. However, if they got together, tasted a load of different extractions, and agreed upon a scoring system (e.g. "yes, I agree that's around 6/10 bitterness") then this would be more interesting.
Thirdly, regarding tests that work directly with taste perception, we already have common ways of discussing the taste of coffee:
clean cup acidity body brightness wet aroma mouthfeel aftertaste etc.
as well as the tools of describing taste likenesses:
sour sweet bitter cinnamon chocolate etc.
And I would be interested in seeing these develop.
Wine tasting, for example, has a generally accepted language of taste description.
I honestly don't know if the idea will take off and fly, but surely you have to agree it would be amazing to be able to diagnose espresso and pull from the vast corpus of knowledge on the subject available online and produce something that can be considered a benchmarking or reference point?
There will be very LITTLE commonality . . . and there will be a LOT. It is very dependent upon suggestions made (deliberately or unintentionally) by the researchers, such as the great coffee v. Maxwell House illustration above. This sort of thing has been done by universities for decades, both generally (think experimental psychology), and specifically -- it's also been the focus of research by the viticulture & oenology departments at universities around the world. This is precisely why, for one example, "black glass" tasting experiments were begun . . .
VA_Geek Senior Member Joined: 22 Dec 2007 Posts: 8 Location: Richmond, Virginia Expertise: I love coffee
Espresso: Isomac Zaffiro & Rancilio... Grinder: Nuova Simonelli MCI... Drip: Chemex and Clever Coffee... Roaster: Fresh Roast 8
Posted Thu Sep 16, 2010, 8:32pm Subject: Re: Effect of variables on coffee taste perception
Is it finished yet?? You've had a whole week! :-P
All kidding aside, this sounds like quite an adventure with legacy-producing results, but a phenomenally arduous process, as has already been pointed out. Like so many things, the subjectivity alone counteracts much of the ability to truly quantify the results without generous assumptions. On the matter of taste, my use of the word "bright" may be far different than someone else's use, never mind my potential (and likely) mis-use of the word. A 'foot' is 12-inches (in the US), everyone agrees, but a 'sour' shot has unknown frames of reference or standard to measure. As I've recently posted elsewhere, I'm not the Coffee Geek yet that I want to be, but I'm working toward it and want to continue learning. Helping on a project of this scope would certainly accelerate my quest.
I'm in the technical support/troubleshooting field of technology and from my 20 years of this developed knowledge, I've learned that I can take the leading symptom with a user and drill-down using follow-up questions, narrowing the possible solutions in my mind with each question until we've diagnosed and solved the problem -- nothing new here as this is the a core method of troubleshooting. I would think the software model of which you're speaking could also follow this pattern, however rather than immediately revealing the possible causes, continue to ask follow up questions that eliminate unlikely causes. I've always hated the troubleshooting grids in manuals that immediately provide 8 possible solutions to the generic "my printer won't print" problem because they require the user to then go through each solution hoping that theirs will be toward the top of the list. Of course with coffee, this too takes some assumption that the user with the problem knows enough to understand the questions. Would a common training/experience baseline need to exist in this study to ensure the quality of the subjective data? For us non-trade coffee geeks, does such an institution exist to provide this training or would it be limited to those with 20 years of developed knowledge?
So with all that said, can I help? Please please please???
a local roastery had an event earlier this summer that compared different roasting methods, Coffeestorks blog documented it (unfortunately i was out of town so couldn't go).
I would like to see just one variable being controlled and then do a blind taste test with 10-20 tasters. Same coffee, same roast, sme grind, same machine, different temperatures. from 195 to 204 differing by 1 degree. Of course you would need 10 machines of the sme type to try this, all with PIDS!!
Has anyone seen a study of this type (or similar)?
Symbols: = New Posts since your last visit = No New Posts since last visit = Newest post
Forum Rules: No profanity, illegal acts or personal attacks will be tolerated in these discussion boards. No commercial posting of any nature will be tolerated; only private sales by private individuals, in the "Buy and Sell" forum. No cross posting allowed - do not post your topic to more than one forum, nor repost a topic to the same forum. Who Can Read The Forum? Anyone can read posts in these discussion boards. Who Can Post New Topics? Any registered CoffeeGeek member can post new topics. Who Can Post Replies? Any registered CoffeeGeek member can post replies. Can Photos be posted? Anyone can post photos in their new topics or replies. Who can change or delete posts? Any CoffeeGeek member can edit their own posts. Only moderators can delete posts. Probationary Period: If you are a new signup for CoffeeGeek, you cannot promote, endorse, criticise or otherwise post an unsolicited endorsement for any company, product or service in your first five postings.
= No New Posts since last visit 