dc.description.abstract | Sensory evaluation of wine color has not been extensively studied. There are few studies
that relate wine color sensory evaluation with wine color measurements and there is not a wme
color reference system, making wine color communication quite subjective. In this investigation a
wide spectrum of wines, 72 red wines and 39 white wines, from 25 varieties ranging in age from
1 to 31 years old, were evaluated for their sensory color attributes: lightness, saturation, and hue
(i.e., perceived color green, yellow, brown, red, purpie and blue). The wines were evaluated over
19 sessions by 22 panelists in order to study variations in the color attributes. The sensory data
were related to the wine color measurements of visible wine spectra and L*a*b* color
coordinates. At the same time, the wines were matched to Munseli color cards with the purpose of
developing a wine color reference system. The wine samples were evaluated in normal wine
glasses located in a light box under a "day light" illumination source. The sensory data were
analyzed by Canonical Vanate Analysis (CVA) and the relationships of sensory and instrumental
wine color measurements were evaluated by Partial Least Squares Regression (PLSR). The
matching data were analyzed by quantitative (Correspondence Analysis) and qualitative (color
card average for each wine) methods. The CVA showed that the biggest sensory variations in
color attributes were in the amount of perceived purple/brown color and perceived lightness for
red wines, and the amount of perceived lightness and saturation for white wines. Differences
within wine types were more associated with wine age than wine variety. The PLSR analysis for
red and white wines provided a correlation between spectrophotometric and L* a*b*
measurements with the visual color attributes determined by the panel. The visible spectra from
600 to 760 nm were important in red but not in white wines. L*a*b* color coordinates were well
correlated with the sensory data and can be useful for wine color discrimination. The matching
data analyzed by Correspondence Analysis did not give any useful results, but with the qualitative
method it was possible to identify 50 red color cards and 30 white color cards that described the wines evaluated. The color cards identified corresponded to Munseil hue pages 2.5Y, 5Y, 7.5Y,
1OY, and 2.5GY for white wines and 2.5RP, 5RP, 7.5RP, 1ORP, 2.5R, 5R, 7.5R, 1OR and 7.5YR
for red wines. The selected color cards then were related to their respective color nomenclature
using a color dictionary. | |