07.04.2013 Views

Measuring colour appearance of red wines. - National Physical ...

Measuring colour appearance of red wines. - National Physical ...

Measuring colour appearance of red wines. - National Physical ...

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

NPL Optical Radiation Measurement Club<br />

Optical Technologies and Measurement Network<br />

Annual Meeting<br />

Digital Measurement<br />

<strong>of</strong> the <strong>appearance</strong> <strong>of</strong> <strong>red</strong><br />

wine – <strong>colour</strong> <strong>appearance</strong><br />

Wednesday 27th – Thursday 28th June 2007<br />

Globe Room, Bushy House, <strong>National</strong> <strong>Physical</strong> Laboratory


Martin, M.L.G.; Ji, W.; Luo, M.R.; Hutchings, J.; He<strong>red</strong>ia, F.J.<br />

(2007)<br />

<strong>Measuring</strong> <strong>colour</strong> <strong>appearance</strong> <strong>of</strong> <strong>red</strong> <strong>wines</strong>.<br />

Food Quality and Preference, 18, 862–871<br />

Department <strong>of</strong> Colour and Polymer Chemistry,<br />

University <strong>of</strong> Leeds, Leeds LS2 9JT, United Kingdom<br />

Laboratory <strong>of</strong> Food Colour and Quality, Department <strong>of</strong> Nutrition and<br />

Food Science,<br />

University <strong>of</strong> Seville, 41012 Sevilla, Spain


Aim:<br />

- To investigate the <strong>colour</strong> <strong>appearance</strong> <strong>of</strong> <strong>red</strong> <strong>wines</strong> in different<br />

depths,<br />

- To explore methods for quantifying the <strong>colour</strong> <strong>appearance</strong> <strong>of</strong> liquid<br />

foods,<br />

- To verify the Pridmore’s finding (2005) that perceived hue shifts<br />

from <strong>red</strong> at shallow to green at deep for some <strong>red</strong> <strong>wines</strong>.<br />

- To quantify the relationships between visual observations and<br />

physical measurements <strong>of</strong> <strong>colour</strong> for a series <strong>of</strong> <strong>wines</strong> with reference<br />

to the change <strong>of</strong> depth.


Four Wines: Table <strong>red</strong>, Oloroso,<br />

Tawny port and Rosé<br />

These four <strong>wines</strong> represent the <strong>colour</strong> range and<br />

characteristics <strong>of</strong> <strong>colour</strong> change <strong>of</strong> wine products.


Four Wines:<br />

These four <strong>wines</strong> were pou<strong>red</strong> into a Petri dish at<br />

different depths <strong>of</strong> 1.0, 2.5, 5.0, 7.5, 10.0 and 15.0 mm


Four Wines in :<br />

I II<br />

III<br />

1200 mm<br />

60°<br />

Φ 82 mm<br />

Φ7.5 mm<br />

870 mm<br />

1870 mm<br />

45°<br />

Φ7.5 mm<br />

The three regions <strong>of</strong> wine <strong>colour</strong> in the cocktail glass<br />

used for physical measurements and psychophysical<br />

experiments


<strong>Physical</strong> measurement<br />

Two instruments were used:<br />

- a Minolta CS1000 tele-spectroradiometer (TSR)<br />

and<br />

- a calibrated single-lens reflex digital camera (Nikon D1X)<br />

The measurement results were presented in CIECAM02<br />

lightness (J), <strong>colour</strong>fulness (M) and hue composition (H)<br />

(CIE, 2001).


• A VeriVide viewing cabinet with diffuse/0 geometry<br />

• A diffused D65 simulator<br />

• A GretagMacbeth ColourChecker DC chart was used<br />

for camera calibration.<br />

White backing<br />

Black backing<br />

Viewing windows<br />

Petri dish<br />

Camera


Visual assessment (magnitude estimation method)<br />

in terms <strong>of</strong> lightness, <strong>colour</strong>fulness and hue.<br />

•8 normal <strong>colour</strong> vision observers (Ishihara Test);<br />

•between 30 and 45 min for viewing;<br />

•3 females and 5 males;<br />

•aged between 24 and 41<br />

All observers had abundant experience in using magnitude estimation<br />

methodology but for most <strong>of</strong> them it was the first time estimating <strong>colour</strong><br />

<strong>appearance</strong> <strong>of</strong> wine samples.<br />

They were all research staff or research students either in the Department <strong>of</strong><br />

Colour Chemistry, University <strong>of</strong> Leeds (UK) or the Department <strong>of</strong> Food<br />

Science and Nutrition <strong>of</strong> University <strong>of</strong> Seville (Spain).<br />

The whole experiment was conducted at Leeds.


Wines<br />

(4 ×<br />

3 + 4 ×<br />

6 ) × 3 ×<br />

2 = 216<br />

×<br />

8 = 1728 observations


Lightness 100<br />

Colourfulness 40<br />

L* = 52<br />

a* = 31<br />

b* = 7<br />

Munsell<br />

four unique hues: <strong>red</strong>, yellow, green and blue<br />

10RP5/6


Results and discussion


Coefficient <strong>of</strong> Variation<br />

CV<br />

=<br />

Statistical Measures<br />

100×<br />

1<br />

N<br />

( Y − f × X )<br />

2 sets <strong>of</strong> data: CV =0% perfect agreement.<br />

Coefficient <strong>of</strong> determination<br />

N<br />

∑<br />

i=<br />

1<br />

CV = 30% 30% disagreement<br />

2<br />

n xy x y<br />

r<br />

[ ( ) ] [ ( ) ]<br />

2<br />

2 2<br />

2 ⎟ n x x n y y<br />

⎟<br />

⎛<br />

∑<br />

− ∑ ∑ ⎞<br />

= ⎜<br />

⎝ ∑ − ∑ ∑ − ∑ ⎠<br />

i<br />

Y<br />

i<br />

2<br />

i<br />

2


Mean (r2 / CV) Lightness Colourfulness Hue<br />

Observer<br />

accuracy<br />

0.88 / 17 0.67 / 37 0.99 / 10<br />

Observer<br />

Repeatability<br />

0.84 / 19 0.72 / 30 0.99 / 8<br />

• <strong>colour</strong>fulness is the worst performance;<br />

• changing in viewing and illumination for<br />

liquid sample;<br />

• repeatability > accuracy;<br />

• mean results were used.


Observer data<br />

b v<br />

10.0Y<br />

7.5Y<br />

100<br />

5.0Y<br />

80<br />

60<br />

40<br />

20<br />

1<br />

2.5Y<br />

2<br />

1<br />

3<br />

10.0YR<br />

2<br />

4<br />

5<br />

7.5YR<br />

5.0YR<br />

2.5YR<br />

10.0R<br />

5.0R<br />

2.5R<br />

10.0RP<br />

7.5RP<br />

0 1<br />

6 2 3 5<br />

5.0RP<br />

5 4 4 6<br />

-40 -20 0 20 40 60 80 100<br />

-20<br />

-40<br />

3<br />

2.5P<br />

6<br />

1<br />

3<br />

5.0P<br />

a v<br />

4<br />

2<br />

7.5P<br />

5<br />

6<br />

2.5RP<br />

10.0P<br />

7.5R<br />

Red<br />

Tawny<br />

Oloroso<br />

Rosé<br />

Lv<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

1<br />

1<br />

1<br />

2<br />

2<br />

2<br />

1<br />

3<br />

4 5<br />

3 3<br />

4<br />

5 5 6<br />

6<br />

4<br />

6<br />

2<br />

6<br />

5 4<br />

Constant hue loci for Munsell<br />

Table <strong>red</strong>: depth ↑, lightness ↓, <strong>colour</strong>fulness ↑ then ↓, hue ≈<br />

Others: depth ↑, lightness ↓, <strong>colour</strong>fulness ↑ ,<br />

hue oloroso<br />

≈ 10YR-5RP, Rose<br />

3<br />

0 20 40 60 80<br />

Cv<br />

Value5<br />

2.5RP;<br />

≈ 2.5RP-5RP;<br />

tawny ≈ 7.5R-2.5R visually significant.<br />

Red<br />

Tawny<br />

Oloroso<br />

Rosé


What <strong>colour</strong> are they ???<br />

I II<br />

III


There is some visual evidence for a brownish green <strong>appearance</strong> <strong>of</strong><br />

the shallow oloroso (Pridmore et al., 2005).<br />

Two out <strong>of</strong> eight observers reported a slight green <strong>appearance</strong> but<br />

these were masked out when the panel averages were calculated.<br />

The effect to the naive wine observer, however, can vary amongst<br />

individuals, i.e. some viewers see the effect and some do not.<br />

Members <strong>of</strong> the sensory panel used in the experiment are experts in<br />

describing the <strong>colour</strong> <strong>appearance</strong> <strong>of</strong> opaque surface <strong>colour</strong>s. They<br />

are wine <strong>colour</strong> naive, presumably not expecting to see the counter<br />

intuitive green tinge in <strong>red</strong> wine.<br />

This sensory phenomenon requires attention in future work.


TSR<br />

b cc<br />

10.0Y<br />

7.5Y<br />

1<br />

100<br />

5.0Y<br />

80<br />

60<br />

40<br />

20<br />

2<br />

2.5Y<br />

10.0YR<br />

3<br />

1 2<br />

1<br />

2<br />

7.5YR<br />

5.0YR<br />

2.5YR<br />

10.0R<br />

5.0R<br />

2.5R<br />

10.0RP<br />

-40 -20<br />

0<br />

60<br />

20 40 4<br />

5<br />

60<br />

7.5RP<br />

5.0RP<br />

80 100<br />

-20<br />

-40<br />

5 4<br />

1<br />

2.5P<br />

4<br />

5<br />

6<br />

3<br />

3<br />

3 6<br />

5.0P<br />

a cc<br />

2<br />

7.5P<br />

5 4<br />

6<br />

2.5RP<br />

10.0P<br />

7.5R<br />

Red<br />

Tawny<br />

Oloroso<br />

Rosé<br />

Jcc<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

1<br />

1<br />

1<br />

2<br />

6 5 4<br />

1<br />

2<br />

2<br />

3<br />

3<br />

6<br />

3<br />

3<br />

6<br />

5<br />

2<br />

4<br />

4<br />

4<br />

5<br />

0 20 40 60 80<br />

Mcc<br />

5<br />

6<br />

Red<br />

Tawny<br />

Oloroso<br />

Rosé<br />

similarity with observer results, except oloroso hue,<br />

a hue change from the greenish yellow Munsell 7.5GY (approximately hcc<br />

= 119, corresponding to Hcc = 132, i.e. 32% <strong>of</strong> yellow and 68% <strong>of</strong> green)<br />

to 7.5R as depth is increased. (see oloroso Point 1)<br />

This agrees with the results <strong>of</strong> the visual estimates <strong>of</strong> the smaller group <strong>of</strong><br />

panel members.<br />

The instrument detected the green hue effect.


cc<br />

TSR on cocktail glass at Regions I, II and III<br />

10.0Y<br />

100<br />

7.5Y 5.0Y<br />

80<br />

60<br />

40<br />

20<br />

2.5Y<br />

10.0YR<br />

7.5YR<br />

5.0YR<br />

2.5YR<br />

10.0R<br />

0<br />

7.5RP<br />

5.0RP<br />

-40 -20 0 20 40 60 80 100<br />

-20<br />

-40<br />

III<br />

I<br />

II<br />

III<br />

III<br />

2.5P<br />

II<br />

I<br />

I<br />

II<br />

5.0P<br />

acc<br />

I<br />

II III<br />

7.5P<br />

7.5R<br />

2.5RP<br />

10.0P<br />

5.0R<br />

2.5R<br />

10.0RP<br />

Red<br />

Tawny<br />

Oloroso<br />

Rosé<br />

Jcc<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

III<br />

II<br />

I<br />

III<br />

III<br />

I<br />

I<br />

II<br />

II<br />

0 20 40 60 80<br />

Mcc<br />

I<br />

II<br />

III<br />

Red<br />

Tawny<br />

Oloroso<br />

Rosé


Digital Camera<br />

b cd<br />

10.0Y<br />

7.5Y<br />

100<br />

5.0Y<br />

1<br />

80<br />

60<br />

40<br />

20<br />

2.5Y<br />

1<br />

2<br />

10.0YR<br />

7.5YR<br />

5.0YR<br />

2.5YR<br />

10.0R<br />

-40 -20<br />

0<br />

0 20<br />

3<br />

40 60<br />

7.5RP<br />

5.0RP<br />

80 100<br />

-20<br />

-40<br />

2<br />

6<br />

2<br />

54<br />

5<br />

4<br />

3<br />

2.5P<br />

1<br />

3<br />

3<br />

5.0P<br />

a cd<br />

6<br />

4<br />

4<br />

7.5P<br />

5<br />

6<br />

5<br />

2<br />

6<br />

2.5RP<br />

10.0P<br />

7.5R<br />

5.0R<br />

2.5R<br />

10.0RP<br />

Red<br />

Tawny<br />

Oloroso<br />

Rosé<br />

Good agreement between TSR and Digital Camera<br />

100<br />

80<br />

60<br />

40<br />

20<br />

0<br />

1<br />

6<br />

1<br />

2<br />

1<br />

5<br />

2<br />

1<br />

4<br />

2<br />

3<br />

3<br />

3<br />

3<br />

4<br />

0 10 20 30 40 50 60 70 80<br />

The digital imaging measurements also show the greenish-yellow<br />

<strong>appearance</strong> (at 2.5GY in Munsell hue, and around 116 Hc value in<br />

CIECAM02 hue scale) for the shallow oloroso.<br />

Jcd<br />

5<br />

6<br />

Mcd<br />

4<br />

2<br />

5<br />

4<br />

6<br />

5<br />

6<br />

Red<br />

Tawny<br />

Oloroso<br />

Rosé


Compare<br />

results<br />

between:<br />

Visual,<br />

CS1000,<br />

and<br />

DigiEye<br />

Lv<br />

Lv<br />

L_CS1000<br />

100<br />

80<br />

60<br />

40<br />

20<br />

100<br />

80<br />

60<br />

40<br />

20<br />

100<br />

0<br />

0 20 40 60 80 100<br />

0<br />

0 20 40 60 80 100<br />

80<br />

60<br />

40<br />

20<br />

J_CS1000<br />

J_DigiEye<br />

0<br />

0 20 40 60 80 100<br />

J_DigiEye<br />

Cv<br />

Cv<br />

C_CS1000<br />

80<br />

60<br />

40<br />

20<br />

0<br />

0 20 40 60 80<br />

80<br />

60<br />

40<br />

20<br />

0<br />

0 20 40 60 80<br />

80<br />

60<br />

40<br />

20<br />

C_CS1000<br />

C_DigiEye<br />

0<br />

0 20 40 60 80<br />

C_DigiEye<br />

Hv<br />

Hv<br />

H_CS1000<br />

360<br />

300<br />

240<br />

180<br />

120<br />

60<br />

0<br />

360<br />

300<br />

240<br />

180<br />

120<br />

60<br />

0<br />

360<br />

300<br />

240<br />

180<br />

120<br />

60<br />

0<br />

0 60 120 180 240 300 360<br />

H_CS1000<br />

0 60 120 180 240 300 360<br />

H_DigiEye<br />

0 60 120 180 240 300 360<br />

H_DigiEye


Compare results between: Visual, CS1000, and DigiEye<br />

Observer vs. CS1000<br />

Observer vs. DigiEye<br />

CS1000 vs. DigiEye<br />

r 2 0.83<br />

CV 21<br />

r 2 0.84<br />

CV 20<br />

r 2 0.99<br />

Lightness<br />

Colourfulness<br />

0.87<br />

21<br />

0.89<br />

21<br />

0.90<br />

CV 5 19 5<br />

0.9 Hc<br />

0.98<br />

11<br />

0.98<br />

11<br />

0.99


Conclusions:<br />

1. A study has been carried out with measurements <strong>of</strong><br />

the <strong>colour</strong> <strong>appearance</strong> <strong>of</strong> different wine samples;<br />

2. observer accuracy and repeatability were conside<strong>red</strong> to<br />

be acceptable;<br />

3. The mean visual results from all observers were used in<br />

the following comparisons with instrumental method;<br />

4. Tele-spectroradiometer and digital camera results agreed<br />

well;<br />

5. Observer and physical measurements show good<br />

agreements for lightness and <strong>colour</strong>fulness;


Conclusions (continuous):<br />

6. For hue there was some discrepancy between observer<br />

data and physical measurement: greenish-yellow hue was<br />

reported;<br />

7. Instruments detected this greenish-yellow hue;<br />

8. All assessment and measurement methods successfully<br />

depicted the changes <strong>of</strong> <strong>colour</strong> that occur with depth;<br />

9. The use <strong>of</strong> <strong>colour</strong> <strong>appearance</strong> methodology represents a<br />

significant step forward in the study <strong>of</strong> <strong>wines</strong>;<br />

Finally, the digital non-contact method introduced here<br />

has many advantages for quantifying <strong>colour</strong> <strong>appearance</strong> <strong>of</strong><br />

liquid food products;

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!