Stories
Slash Boxes
Comments
Search

News for nerds, stuff that matters

Slashdot Log In

Log In

Create Account  |  Retrieve Password

Wine Tasting Via Computer

Posted by Zonk on Fri Dec 30, 2005 07:53 PM
from the snoots-replaced-by-technology dept.
Software Science
smooth wombat writes "What makes a good wine? Why do some wines have a smooth, almond-like bouquet while others have a sharper, more acidic bite to them? These questions and more have usually been answered by oenologists who can list the subtle nuances of a particular wine and tell you if it's good or not. However, vinters don't have the luxury of waiting until a wine is ready to be drunk to know if they have produced a good, drinkable product. Lorenz "Larry" Biegler, who teaches chemical engineering at Carnegie Mellon University, in conjunction with industry scientists in Chile, is working on mathematical formulas to automate the fermentation process, adjusting ingredients and conditions to ensure robust flavors and higher yields from grape harvests. The researchers have been collaborating for more than two years and are studying only white wines, since reds are more complex and contain solids that make them difficult to analyze."
+ -
story
This discussion has been archived. No new comments can be posted.
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.

Wine Tasting Via Computer 50 Comments More /

 Full
 Abbreviated
 Hidden
More
Loading... please wait.

  • Ouch (Score:2, Interesting)

    by Kelz (611260)
    Victory Wine anyone? I don't see any wine enthusiasts buying into this.
    • This will still be made with the same grapes, the same yeast, and the same assorted-other-stuff as regular wine. Why would adjusting the precise amounts/times/temperatures with the aid of a computer somehow render it undrinkable?

      • Re:Ouch (Score:5, Insightful)

        by Kelz (611260) on Friday December 30 2005, @08:13PM (#14368251)
        Not undrinkable, but creating a GOOD wine basically an art. You can't replace a Van Gogh with electronics. I CAN however see this being used to make low-quality cheap wines more consistant and a bit better while still keeping costs down.
        • You can't replace a Van Gogh with electronics.

          Yes, but a pretty picture and a Photoshop filter = Andy Warhol...

          • Re:Ouch (Score:2, Interesting)

            by Kelz (611260)
            But you don't always have the same grape crop, which makes it rather impossible to copy, and thats where the human element is needed to create a great wine with different grapes every year.
              • When it comes to rock, I like live music performances better than studio recordings, even if it means hearing a few minor mistakes. A good band never quite plays something the same way twice.

                For classical or other instumental works, I like recordings that are somewhat "dirty". I want to be able to hear a bit of a rasp from the bow on the cello strings, so that I can practically smell the rosin dust.

                I don't just want to hear sterile, perfect notes. If I did, I'd listen to MIDIs. I like studio chatter. I
    • IIRC, it was Victory Gin actually, but I see where you are coming from (Yes, I did have to flaunt my interest in Orwellian literature =D) (And I have lots of karma to burn)

  • Seeing if the wine is "Ready"? (Score:5, Interesting)

    by Ruff_ilb (769396) on Friday December 30 2005, @07:58PM (#14368179) Homepage
    Wine tasting, as I'm sure most experts will agree, is as much of an art as anything; I doubt that people will allow a computer to tell them if a wine is "good" or not, even if it's right most of the time.

    OTOH, if the computer only tells people if the wine is drinkable, or ready to be tasted, that's a different story. As long as the computer doesn't try to encroach on the "art" side of wine tasting and stays firmly on the "science" side, I think that it could be quite a useful invention - although to a tiny demographic.

    • Re:Seeing if the wine is "Ready"? (Score:5, Interesting)

      by shawb (16347) on Friday December 30 2005, @08:06PM (#14368217)
      Or if the computer could tell the vintner subtle variations in temperature or other factors during the fermentation process which would help improve the wine. Maybe evnetually the computer could help determine which woods would be the best for storing a certain batch. Computers wouldn't really be able to help a bad winemaker make good wine, but they could help a good winemaker make better wine.

    • Re:Seeing if the wine is "Ready"? (Score:5, Insightful)

      by Ruff_ilb (769396) on Friday December 30 2005, @08:08PM (#14368225) Homepage
      Another issue with their formula: It implies that "If we add more X to mixture Y, it'll taste better."

      When wine tasting, again, is such a subjective thing, even given the fact that they may be able to figure out what gives the wine more of a fruiter aftertaste, for example, they still don't know if they actually WANT a fruiter aftertaste.

      Again, if we get the experts deciding what would make the wine taste "better" and then working with the machine to decide what can be done to make the wine taste the way the experts want, we're still only doing something that can probably be done already (IANAWineGeek, BTW) without the aid of an expensive machine. And one expert's "better" might be another expert's "ruined"
      • Sure the computer knows if a fruity aftertaste is desired -- it's called a configuration parameter. And while the humans know the desired results, they may not have the observational ability to guide it there.

        A basic machine learning approach would go as follows:
        * prepare several batches in slightly different ways
        * measure everything imaginable at various stages of the growing and vinting process
        * when the wine is ready, ask experts to rate values like "fruity aftertaste" and "body strength"
        * throw all the
    • Wine tasting, as I'm sure most WINE SNOBS will agree, is as much of an art as anything; I doubt that WINE SNOBS will allow a computer to tell them if a wine is "good" or not, even if it's right most of the time.

      If I think the wine is good, I'll drink it. I don't care if someone or something agrees with me or not.
      • I doubt that WINE SNOBS will allow a computer to tell them if a wine is "good" or not, even if it's right most of the time. If I think the wine is good, I'll drink it. I don't care if someone or something agrees with me or not.

        Yeah, I hear ya, bro. Gotta love a fresh chilled flask of MD-20, those dudes know how to brew up a decent vin rosy!

    • Re:Seeing if the wine is "Ready"? (Score:5, Insightful)

      by blakestah (91866) <blakestah@gmail.com> on Friday December 30 2005, @08:28PM (#14368304) Homepage
      The company Enologix already uses a "formula" based on chemical composition to predict wine rating scores. It translates all of that winemaking into one number...one number that matters a lot.

      They contract clients who test wine at different times. It tells them how long to ferment, when to stop, if the batch will ferment faster or slower than usual, etc. And of course, in France, how much ethylene glycol to add at the end. They average a 5-6 rating point increase in the first year their clients use them.

      To take it even further, I'll use a coffee example. Illy did a LOT of scientific taste testing studies on its coffee (or in the US, espresso). They found the magical chemical formula. Then, they would test each batch, alter the chemical content to become perfect, and sell it.

      The Italians were OUTRAGED! It was as bad as cigarette makers adding nicotine to cigarettes!

      So what did Illy do? They stopped that process. Now, they draw several batches in parallel. Test all of them. And figure out how to combine them to achieve the magical formula. The end result is chemically the same, but the Italians are happy to know that Illy comes from 100% roasted Arabica beans from Brasil.

      Of course winemakers already try to do this with blended wines. But it would be pretty easy if each wine were independently chemically tested, and then the appropriate convex combination were defined to result in a 90+ Wine Spectator rating. And I'm sure many of them do this already.

    • I completely agree. Since the machine is only on white wine so far I see that they are having difficulties with reds. I'm going to consider port a step beyond red and hope I'm long dead and gone before a machine tells me what a good port tastes like.

      If they develop this, wine enthusiasts will just pay more for wine judged by humans.

    • Chemical analysis of wine flavors is (and has been for several years) used mostly to determine how to process the grape juices to get a certain flavor, not if the wine is drinkable. Manipulating grapes according to these computer-derived formulas has increased the ratings - and thus popularity - of a lot of California wines by allowing the creation of wines custom-tailored for the palettes of critics Robert Parker and James Laube. Because the company that does it keeps its list of customers confidential, th

  • My pedantic moment for the day... (Score:5, Funny)

    by KingSkippus (799657) * on Friday December 30 2005, @07:59PM (#14368182) Homepage Journal
    However, vinters don't have the luxury of waiting until a wine is ready to be drunk...

    For the record, the word is vintner [wiktionary.org], not vinter.

    In Soviet Russia, vinters are wery, wery cold.


  • Chemical anlaysis of wine [sciencenews.org] has been going on for some time for a variety of purposes [awri.com.au].

  • Yeasts (Score:3, Interesting)

    by TubeSteak (669689) on Friday December 30 2005, @08:05PM (#14368214) Journal
    http://www.distillery-yeast.com/turbo_yeast_functi on.htm [distillery-yeast.com]

    Yea for turbo yeasts.

    I'm still waiting for yeasts that convert both sucrose/glucose & xylose to be available to your average consumer.

    Wine yeasts give 14%~18% alcohol content.
    Distillers yeast gives up to 21%
    xylose converting yeast can up the yield significantly
  • There went my dream job in the 'quality control' dept!
  • ... a non-porn reason to lick my screen!

  • Stop it right now! (Score:5, Insightful)

    by rufusdufus (450462) on Friday December 30 2005, @08:10PM (#14368233)
    some wines have a smooth, almond-like bouquet

    No they dont. Nor do they taste like chocolate, raspberries or broccoli.
    Really wine has 5 basic flavorings: 1) rotten grape 2) alcohol 3) wooden barrel 4) cork 5) mold

    • It's true. You should have gotten some "funny" or "insightful" points, your choice. As long as the machines are testing for those, what's the controversy? It's just saving the vintner money. When they start telling us that a ten-year-old cabernet lacks the tang of a fresh cabernet, we'll know marketing has taken over the farm.
    • Well, yesh, for the kind of wine the average slosheddatter can afford you are abschooly correct - hic!
    • stop. next you'll tell me that my cigar doesn't taste like saddle leather, raw meat, vanilla beans, dirt or roasted corn. and yes, those are all descriptions lifted from Cigar Afficionado.
    • Hmm, well, wine is typically made in stainless steel vats nowadays (even most of the reasonably good wines) which probably impart little to no flavor... so you're down to four. :-)
    • Right. So, I assume that objects can't be orange unless they are made of oranges, and only violets can be violet? Of course, the same goes for beige, turquoise, indigo and pretty much any other color [wikipedia.org] beyond black, white, red, blue, green and yellow.

      Either that, or "chocolate", "vanilla", "burnt rubber", "red berries" etc are simply descriptions of flavors, just as "chartreuse", "lime" and "burnt umber" are descriptions of colors. Lacking a better way to classify and describe sensations, the only way to give

  • Bah. (Score:4, Funny)

    by Ruff_ilb (769396) on Friday December 30 2005, @08:10PM (#14368236) Homepage
    Old news. I've been able to debug my WINE install for ages now.

    And what's this talk about "grapes" and "yeast", are they new distros?

  • Under the influence (Score:3, Funny)

    by killmenow (184444) on Friday December 30 2005, @08:13PM (#14368255)
    Being under the influence of a goodly sum of wine as I type this, I can surely tell you no computer could possibly tell good wine from Mad Dog 20/20.

    Good wine ~ good art. I can't define it, but I know it when I drink it.

    Oh, and God Bless Oregon.

    w00t! ... (hic...)

  • Enologix (Score:4, Informative)

    by blakestah (91866) <blakestah@gmail.com> on Friday December 30 2005, @08:22PM (#14368281) Homepage
    There's a lot of money in this. Not surprisingly, a former vintner decided to make this into a company [enologix.com]. In the past they have accurately ranked wine involved in taste comparisons by experts. Unexpectedly, they use a one dimensional scale which works, suggesting the wine judges use a one dimensional scale too.

    The exact formula is a mystery/trade secret. But it is no secret that Enologix tests many of the top wines at various points in production, and they AVERAGE a 5-6 point rating increase for the first year they are contracted by their clients.

    I've been to their web site before when it was useful and worked...right now they appear to be hosed.

    • Wired had a big article on Enologix a few years ago. The wine industry is terrified that once wine is figured out, good wine will be cranked out on an industrial scale, by mixing ethyl alcohol, water, and flavoring.

      This has already happened in the hard liquor industry. They try to keep a low profile, but Frank-Lin Distillers [frank-lin.com] makes over 1000 different brands of liquor sold on the West Coast. They use only about 100 different formulas, though. It's all about branding. They're located near the railroad y

      • Because, most of the time, it will be better.

        No, it'll be cheaper. To make, if not to buy.

        American "blended whiskeys" are usually not different whiskeys blended together, but whiskey blended with neutral spirits. Which is why they tend to be awful. Skyy is the fanciest stuff that Frank-Lin produces, which is why they make sure it shows up in those pictures. What they don't tell you is that Skyy comes off a separate dedicated production line. Pretty much everything else they make is bottom-shelf crap:

  • Why? (Score:3, Interesting)

    by Harker (96598) on Friday December 30 2005, @08:28PM (#14368300)
    Where's the fun in that?

    H.

  • Homogenous Wine (Score:5, Insightful)

    by eander315 (448340) <tomatobasil.gmail@com> on Friday December 30 2005, @08:34PM (#14368333) Homepage
    This sounds like a great way to make every bottle taste the same, from vintage to vintage and vintner to vintner. Part of the fun of drinking wine is selecting a bottle that you like from the thousands available, not to mention tasting how your favorite wine changes from year to year as the growing conditions change.

    This might work well for jug wines that no one really drinks for the taste in the first place, but even cheap table wine has subtle (or not-so-subtle) nuances that might be erased by this process.

  • But... (Score:4, Funny)

    by BluhDeBluh (805090) on Friday December 30 2005, @08:44PM (#14368360)
    Wine Is Not Ethanol... Oh comeon. Someone had to do it.

  • You know you should read less slashdot... (Score:3, Funny)

    by Kjella (173770) on Friday December 30 2005, @08:56PM (#14368386) Homepage
    ...when the first thought is "How else are you going to test wine [winehq.com]?"

  • Red Wine (Score:3, Funny)

    by Fwonkas (11539) <joe AT flappingcrane DOT com> on Friday December 30 2005, @09:30PM (#14368483) Homepage

    I fear my reaction -- "Of course they could only analyze white wines." makes me a snob.

    Not that I drink much wine anyway. That there is the crazy sauce.

  • W.I.N.E. (Score:5, Funny)

    by Andrewkov (140579) on Friday December 30 2005, @09:36PM (#14368504)
    Wine Is Not an Emultar .. oh, wait a sec...
  • LCD (lowest common denominator) wine.
  • French wines have not yet surrendered. They are NP-problems and can not be solved in a reasonnable time-frame. RTFA, my mistake, they are talking about foreign beverages.
    BTW, a wine is as good as the meal it goes with. Chili beans any one?

  • Wine (Score:3, Informative)

    by Pompatus (642396) on Saturday December 31 2005, @03:05AM (#14369425) Journal
    Personally, my drink of choice is Crown and diet, but since I worked as a bartender for a wine bar for 2 years, I have a passing knowledge of wine. So here's my take. The statement "These questions and more have usually been answered by oenologists who can list the subtle nuances of a particular wine and tell you if it's good or not." is misleading. It doesn't matter what "oenologists" say about a wine. If you want to find a "good" wine, try out several different ones and decide which one YOU like. Then find the least expensive wine you can find that suits your taste. My 2 personal favorites are the 1999 Katheryn Kennedy Lateral, and (cant remember the year) J. Bookwalter Merlot. However, right now I'm drinking a 2003 Rosemont Estate Shiraz, because it is quite similar to those 2, but it is $9 at the local grocery. And lastly, the most important thing. After you learn enough to bullshit your way through a wine conversation (the last 2 sentances made me sound like I know what I'm talking about, huh?), you can talk about it and enjoy it while not appearing to be a drunk. Because wine FUCKS YOU UP :)

All trademarks and copyrights on this page are owned by their respective owners. Comments are owned by the Poster. The Rest © 1997-2009 Geeknet, Inc.