| Author | Message | ||
     Gustavo Figueira de Paula (gustavo) New member Username: gustavo Post Number: 4 Registered: 6-2008 |
Hi Pedro, I understand the confidential nature. But I�ll want to talk with you about things that you actually CAN talk about - could you provide a phone number or e-mail? My e-mail is [email protected]. I�ll like also to thanks again to all guys at this Forum, this is the most productive I�ve already joined on ten (or more) years. Hugs, Gustavo de Paula. | ||
| Pedro Issa (pedro_issa) New member Username: pedro_issa Post Number: 4 Registered: 2-2006 |
Hi Gustavo, The NIRS Taste Application in Coffee has been used in Brazil and Italy since 1997 until ours days. I have worked from 1993 until 1997 with this Italian Coffee Brand and their Coffee Grain buyer in Brazil to develop the Calibration Model to predict the taste. Unfortunately I�m not allowed to informed their name or further details of the calibration technique applied on it , but I can guarantee the NIRS Dispersive Grating it is able to do it. Rgds Pedro Issa | ||
| Pedro Issa (pedro_issa) New member Username: pedro_issa Post Number: 3 Registered: 2-2006 |
Hi Gustavo, The NIRS Taste Application in Coffee has been used in Brazil and Italy since 1997 until ours days. I have worked from 1993 until 1997 with this Italian Coffee Brand and their Coffee Grain buyer in Brazil to develop the Calibration Model to predict the taste. Unfortunately I�m not allowed to informed their name or further details of the calibration technique applied on it , but I can guarantee the NIRS Dispersive Grating it is able to do it. Rgds Pedro Issa | ||
| Gustavo Figueira de Paula (gustavo) New member Username: gustavo Post Number: 3 Registered: 6-2008 |
Hi Ana, I�ll contact Andres Villegas, thanks very much. About your work with olive oil, it�s available online? Surely it will be a good reference for similar works with coffee. Hugs, Gustavo. | ||
| Ana Garrido-Varo (garrido) New member Username: garrido Post Number: 1 Registered: 1-2006 |
Hi Gustavo, During NIR-99 held in Verona (Italy) we presented equation results for prediction of panel test scores in olive oils. The calibrations showed very good performance only limited by the well know intra en inter-panelists variations. I will recommend you to calibrate only with the data from some human panelists well reputed and recognized by coffee producers and industrials. At present, Mauricio Villegas from Cenicaf� (Colombia) is doing a training in the University of Cordoba on NIRS Coffee analysis. He has already calibration equations for coffee. You should contact him to learn more about his work ([email protected]) . He wants to use nIRS to distinguish between Brazilian and Colombian cofee!!!. With respect to your last post concerning if a DA-based spectrometer may be used I must say that from my view, what it is important is the scanning range. I will recomend to use full VIS + NIRS range for sensorial variables as you would like to calibrate with. Best regards to all the forum Ana Garrido-Varo | ||
| Gustavo Figueira de Paula (gustavo) New member Username: gustavo Post Number: 2 Registered: 6-2008 |
Hi boys, Thanks very much. All comments on this forum, and also by private e-mail were very useful. I believe that 'fruits' from these words will ripe soon. I really believe that work done with wine can be a good model on how to work with coffee - also the work done with chocolate and tea. I�ll ask my local library manager to gather these articles to an in-depth learn. Several are readily available to us in electronic way thanks to our governmental agencies that provide full access to more than 12,000 peer-reviewed journals, free of charge, to all public Brazilian research institutes. Venkynir cited tea: few publications deals with 'taste' of tea - sourness or acidity, etc. A lot deals with polyphenols. Since taste - a subjective attribute - is much harder to quantify than polyphenols, its not a surprise. Several researches on Brazil shows quantification of protein, fat, caffeine, etc, on coffe. But dealing with taste... there�s two recent works, one very exploratory (a master degree dissertation) and a on-going doctoral thesis. Very similar to tea�s case. Mark and David Donald cited work on wine done by AWRI, and I�m happy that Daniel Cozzolino contacted my (by private e-mail), because I believe that AWRI is a good model of academy-industry partnership that should be more used in several countries (including Brazil). I also agree with Mark that a "single magic tool" probably will never exists (at least in the timespan of my life), but talking strictly about coffee taste, NIR provides the best cost-to-performance ratio, since can determine taste (after training) while costing much less than a NMR, which is also capable to identify tastes (after training)... if someone has a better alternative, please, let me know!! I believe that pure electrochemical methods could do the job, but some problems, like electrode recovery, quickly arises; also, solid state measurements are very problematic... I tried impedance - hard work... The several other guys and places cited are also very interesting, I�m browsing web to find more information about them to contact directly. One point I like to ask for those who already worked with both dispersive and DA-based spectrometers: what would be the better choice for an application like taste "simulation"? [Oh no, this question again...] I have "a feeling" that a 512pixel, double TE-cooled InGaAs array have enough resolution and low noise to do such work, being less expensive. Some works reports that a given job can be done very well by NIRS, if properly calibrated, at 2, 4, 8, 16 and even 32 cm-1 resolution with little or no loss on prediction capability; so, intuitively, I understand that if a 2048-point spectra taken at 2 cm-1 interval can give almost the same qualitative/quantitative result that the same spectra at 8 cm-1 (which reduces the spetra to 512 points), the result with a 512-pixel DA will be almost the same. Supposing the noise level to be similar... It�s right? I�m new to spectroscopy, so forgive if I�m writing some non-sense things. Hugs, Gustavo. (Wow, it�s really a very long post... sorry!) | ||
| venkatarman (venkynir) Senior Member Username: venkynir Post Number: 67 Registered: 3-2004 |
Dear Gustavo; kindly look at the recent publicaiton of ASIAN spectrocopy . I have seen the tea taste through NIRS. It may helpful to you. | ||
| Mark Gishen (mark_gishen) New member Username: mark_gishen Post Number: 1 Registered: 2-2007 |
Hi Gustavo (and nice to hear from you again David!) I was formerly manager of the AWRI team and we did look at this type of approach. I've been out on my own as an independent consultant for a couple of years now so I might not be up to date with the most recent research. However, I do recall some work done with NIR on coffee taste in Italy and Spain I think some years back, but can't lay my hands on the references immedately. As for wine, there are a couple of articles that might be of interest. Firstly, using NIR for sensory of white wines (Analytica Chimica Acta 539: 341-348; 2005) where some promise was noted. Secondly, a combination of NIR and MS-ENose for sensory on white wines also (Analytica Chimica Acta. 563: 319 � 324; 2006) and this gave a better result. There are also some other publications from the team that will be of interest. It seems so far that there is no single technique that is a 'magic bullet' so a range of options should be considered. Hope this is of some help. Regards, Mark Gishen. South Australia. | ||
| David Donald (ddonald) New member Username: ddonald Post Number: 4 Registered: 11-2006 |
Dear Gustavo, The wine industry has a fair overlap with the kinds of questions you are asking. In Australia there is a organization called the AWRI (Australian Wine Research Institute) that has been working at correlating wine tasters results with NIR measurements. Here is a abstract link to a published manuscript from the AWRI: http://www.ncbi.nlm.nih.gov/pubmed/18389223 The guys from the AWRI are a friendly bunch of guys (as are all Australians) and the work they have published would be of interest to other industries wanting to look at qualitative measures with NIR. Best Regards, David Donald (sugar cane researcher - not wine). | ||
| Tony Davies (td) Moderator Username: td Post Number: 174 Registered: 1-2001 |
Hello and welcome Gustavo! I have done some work with coffee but mainly about analytical data. Many years ago we demonstrated that NIR could easily distinguish between regular and decaffinated coffee. However, what may be much more interesting to you is that I and colleagues at the Food Research Institute in Norwich UK did a very large study trying to relate the taste of chocolate (measured by a trained [NOT expert] taste panel) to the NIR spectra of raw cocoa beans. You will find it in Vibrational Spectroscopy 2 (1991) 161-172. Prediction of chocolate quality from near-infrared spectroscopic measurements of the raw cocoa beans, A.M.C. Davies et al. What we demonstrated was an ability to improve the final acceptance success of beans selected in the raw state. If you e-mail me ([email protected]) a postal address I can post you a copy. As far as I know it has never been used by the chocolate industry! Of course Chuck Miller is correct; Harald and Magni are the experts in this area. I recommend their book (available form NIR Publications) Multivariate Analysis of Quality, An Introduction. Good Luck and Best wishes, Tony | ||
| W. Fred McClure (mcclure) New member Username: mcclure Post Number: 43 Registered: 2-2001 |
The Council Offers a Bibliography for sale that has over 30,000 citations from the NIR literature. If you are interested in obtaining a copy please contact me: Fred McClure [email protected] | ||
| David Russell (russell) Senior Member Username: russell Post Number: 39 Registered: 2-2001 |
Steve Brown at the University of Delaware also did some work on coffee beans. www.udel.edu/chemo | ||
| Howard Mark (hlmark) Senior Member Username: hlmark Post Number: 191 Registered: 9-2001 |
Good thinking, Chuck! In fact, your message prompted me to recall that Heinz Siesler, of the University of Essen, has also reported on coffee analysis. My recollection is that the main thrust of the talk he gave was an examination of the effect of different spectral resolutions on the behavior of the calibration, but I suspect that was a side-issue in project whose main interest was about analysis of coffee. \o/ /_\ | ||
| Charles E. Miller (millerce) New member Username: millerce Post Number: 5 Registered: 10-2006 |
Greetings Gustavo: First of all- welcome to the Forum! I'm sure that there are several groups working with the kind of problems that you mention, but I'm most familiar with the extensive work done at the Norwegian Food Research Institute (Matforsk). If I recall correctly, Magni and Harald Martens did some very nice work in the early 80's using multivariate methods to relate sensory and instrumental (including NIR) data- and similar work has been ongoing there, with the above-named researchers, Tormod Naes, Tomas Isaksson, and others. I hope that this helps you get some searches started. I would be glad to provide more details if necessary. As I'm sure you know, sensory and consumer preference data tend to "behave" rather differently than instrumental data- and these folks have done a very nice job of discovering and documenting these behaviors. Good luck with your project! Best Regards, Chuck | ||
| Gustavo Figueira de Paula (gustavo) New member Username: gustavo Post Number: 1 Registered: 6-2008 |
Dear Guys, This is my first post to this forum, so I want to say that I�m very happy to join this community. I would like to ask about NIR application on Coffee Industry. Almost nothing has appeared in the past about coffee on this forum, and very few and scattered publications on this topic is available. I�m a Brazilian Materials Engineer, and on last four years I�m got involved on a scientifc research trying to correlate coffee taste with electrochemical measurements, using polymeric membranes as mediators. Despite some good results and a good acceptance of the idea by our local industry, this work is almost stalled, because some technicals challenges remains unsolved. I�m turned to NIR as a possible solution - NIR is fast, reproducible, not too much expensive (USD 20.000 is acceptable), easy to maintain and readlly available from dozens of sellers. So now I�m trying to setup a large calibration sample set to NIR evaluation. Multivariated methods are not a problem. The main parameters are those evaluated by professional tasters - sweetness, aroma, flavour, acidity (not pH), sourness, body, etc. The original idea is to measure powdered coffee (diffuse reflectance) wich is much more interesting that brewing the coffee to do the measurement, by several factors. So I would like to ask:: there�s similar works around the world? What about using people as primary reference (today there�s no instrumental method that evaluates taste as humans)? Why coffee industry around the world doesn�t use NIR as a tool to evaluate quality or to develop new products? Or uses, and I don�t know about it? There�s similar works with other foodstuffs - correlating sensorial attributes with NIR, at industrial level? Last, if someone has interest to keep in touch at more delailed level, including commercial aspects, feel free to contact me at [email protected]. Forgive possible mistakes in my english, I�m still evolving! Hugs, Gustavo de Paula. |