Menu Icon


Looking for homebrewing gift ideas? Check out our previous gift guides here or here!
Also, if you enjoy BrewUnited, please consider doing your Amazon shopping via our affiliate link!



You are here: Home --> Forum Home --> Brewing Forum --> Brewing Discussion --> on tasting

Jump to: [First Page] [Prev] 1 2   
ingoogni
nl
314 Posts


The problem is not getting different beers, or doted beers, the problem is getting different kind of taste/test panels. One brew, two fermentations, one batch of each for the pro's one for the home brewers.

On the other hand it is interesting to see at what levels of a added substance the result of a triangle test gets "clearer" and how that compares to average taste thresholds, following Marshall's standard procedure. Now all of this may well have been done by Meilgaard & Co. I'll have to look up the original literature on his work someday.

Use maggi-sauce for autolysis taste




Posted 34 days ago.
Edited 34 days ago by ingoogni

uberg33k
Charter Member
The Internet
314 Posts


So, just so I'm clear on what you're proposing, you plan to administer an off flavor panel to every Xbeerment participant in order to ascertain what they can and cannot taste.  Then future participants would be picked based on their ability to detect whatever variable is being tested?  That would be quite the epic undertaking.  I like it, but I'm not even sure if that's manageable.  I mean, you'd have to rerun the sensory panel on every participant before every Xbeerment to verify the results are still valid.  Changes in medication, health, or living situations could all change your ability to taste things theoretically.  Heck, even taking a week vacation to the beach could make you have a reduced sensitivity to DMS or be completely blind to it.  That's a whole lot more $$$/experiment and a whole lot more time to do them, but it would definitely make them more meaningful.  It would also radically change the scope of experiments being done.  For instance, you couldn't run an experiment like that last warm lager experiment and ask "does warm fermentation make a difference?".  You'd have to run it like "does warm fermentation of this single strain of yeast make a difference in perception of sulphur compounds?", run the sensory board to confirm your participants can sense minimal expected threshold changes in sulphur, pause so there's no palate fatigue, and then run the experimental tasting.



Posted 34 days ago.

brulosopher
Charter Member
Fresno, CA
167 Posts


???




Posted 34 days ago.

mchrispen
Bastrop, TX
485 Posts


Joke, didn't land inline with the other forum posts... move along.



Posted 34 days ago.

mchrispen
Bastrop, TX
485 Posts


How is a null hypothesis effective in measuring the testers? It is an unfair misapplication of the test, assuming you intend to draft inference of the tester's ability... and not that x-amount of vinegar in a given beer will result in difference. You KNOW the difference, and are therefore presenting an unfair and biased test.

Triangle tests are to determine statistical deltas blind to the noise of randomness (or the ability to determine and factor randomness) specific to the testing hypothesis. It intentionally introduces noise and eliminates calibration. A null hypothesis simply determines that there is no difference between X and Y, but in this case the difference is artificial - hence testing the test taker.

Here you are measuring the instrument (in this case the ability of a testing group to detect a specific off flavor) and its slope of detection. This is better accomplished through tester specific designed tests with very large sampling rate, or empirical qualification/calibration before testing.




Posted 34 days ago.

ingoogni
nl
314 Posts


@Uberg,

no, that would be impossible. All I'd like to do is compare the result of a pro panel tot the result of a home brewers panel. The home brewers in the standard XBMT way, the pro's in their own (analytical) way.




Posted 34 days ago.

uberg33k
Charter Member
The Internet
314 Posts


@mchrispen

I don't understand your objection.  Exactly what are you proposing?

@ingoogni

Would that be apples to apples though?  If some research institute in Germany ran flavor analysis on changing a single variable in a beer (say fermentation temperature) but the batch size was 10bbl as opposed to a Xbeerment run at 10gal, then the results might be different.




Posted 34 days ago.

ingoogni
nl
314 Posts


Apples can taste different too :)

Especially the fermentation temperature related XBMT's don't match at all with my experience. They don't match with literature, they don't match with the experience of thousands of brewers they don't match with what "the industry" does. Brewing is a process that scales amazingly well. Now why are there a few specific tests that don't give the expected result. All the other test do line up with other research, only brewers apparently didn't know. Also I get the impression that on may tests it is a very close match. That is always a frightening result as it doesn't say there is a clear difference, nor does it say there is no difference (and sadly many seem tot interpret it that way, "it does not matter"). In that regard a test with manipulated beer is interesting, at what level of manipulation do you get a clear difference of at least 80:20.

There actually is a school of statisticians that say that a result of less that 80:20 or 20:80 is no result and the set up of the test was wrong. I don't know.

In all the tests I see two mayor points that could be of influence, with the "way of tasting" as the one that stands out. That is what I would like to see tested in some way. I expect a difference there, a clearer differentiation with "the pro's" versus a less outspoken result from the amateurs.

Oh, and yes, I'm heavily biased against triangle tests or AB's or A-B-X. Another example in art history and painting/image analyses they go through great lengths to not compare two images next to each other. For example IR images of the under-painting versus the top painting. Switching your eyes from the image on the right to the image on the left takes enough time to make you forget essential details. They now make high res images and use "curtains " in software that are dragged over the image, one over another, so you can focus one one point and see that flipping from one to another situation. Similar in comparing satellite images, the old trick is to superimpose the images with mirrors and optics over each other to greatly reduce the chance that enemy missiles stations are missed. In all these methods the computers assist but it is still the human interpretation that is needed on a level that our untrained eyes can't achieve.

My bias is not important, the experiments are important, I and many others have gone through several, they should stimulate (young/fresh) brewers to do it. That is why I advise brewers that post their first own recipe to brew it, regardless of what all the others say, and after that come back and compare the results with the given comments and adapt and brew again. From many comments on the blogs and on diverse fora I get the impression that the results on the tests have an other effect "it does not matter, it is all the same regardless" and that saddens me.






Posted 34 days ago.

Jump to: [First Page] [Prev] 1 2