Instrumental color measurement specifications and factors affecting measurement consistency in pork. NPB #

Transcription

1 Title: Instrumental color measurement specifications and factors affecting measurement consistency in pork. NPB # Invetigator: Institution: M.Susan Brewer University of Illinois, Urbana-Champaign, Illinois Date Received: 11/8/1999 ABSTRACT: To evaluate factors affecting selected measures of pork color, the gluteus medius (), longissimus lumborum et thoracic (), semimembranosus (), biceps femoris (), and triceps brachii (), from pale, soft, exudative (PSE), dark, firm, dry (DFD) and normal carcasses, were sliced and allowed to bloom for up to 30 min. L*, a*, and b* values, hue angle, and chroma were determined. Spectral reflectance was determined using HunterLab (illuminants A, C, D 65 and F) and Minolta (illuminants C and D65) Spectrocolorimeters. Eight judges evaluated pork color using the Japanese color standards. had the highest L* value and hue angle, and the lowest a* and b* values, and chroma. had the lowest L* and b* values, and hue angle, and the highest a* value. Bloom time had no effect on L* value; hue angle stabilized after 5 min, a* and b* values after 10 min and chroma after 20 min. ph was correlated with L* and a* values, hue angle and chroma of muscle (r = -0.82, 0.51, and -0.40, respectively). Using the Minolta/illuminant D65, pork color best correlated with b* and L* values (r = and ) of. Using the Hunter/illuminant F, visual pink color was best correlated with L* value of all muscles except and. Using the Hunter/illuminant C, visual pink color was well correlated with L* value of,, and (r > for each muscle). Using the Hunter/illuminant A, visual pink color was well correlated with L* value of all muscles (r > -0.86). Overall, the best instrumental measure for predicting visual pink color was L* value; the correlation coefficient was significant for all muscles except. Regression equations for interconverting color measures among the instrument/illuminant combinations are reported. Correlation coefficients between instrumental measures for individual muscles are also presented. OBJECTIVES: The objectives of this study were to evaluate factors affecting instrumental and visual measures of pork color including carcass ph, muscle location and bloom time, and to evaluate factors affecting instrumental color measures.

2 INTRODUCTION: During the Pork Chain Quality Audit, packers reported a 10% incidence of pale, soft, exudative (PSE) pork and a 4% incidence of dark, firm, dry (DFD) pork (Cannon et al., 1996). PSE can be observed in genetically susceptible carcasses that undergo rapid postmortem ph decline (from 7.2 to 5.8 in 45 min) and/or in carcasses with a low ultimate ph. Muscles most likely to develop PSE or DFD characteristics are the longissimus lumborum, semimembranosus, biceps femoris, gluteus medius, and lateral portion of the semitendinosus (Warner et al., 1993). These muscles are commonly evaluated (visually or instrumentally) as indicators of overall quality. The association between PSE and DFD quality defects and their respective "colors" has led the industry to assign visual color scores to carcasses. The use of instrumental color evaluation is of significant interest to the industry because of its speed, consistency of measures and potential for use as the basis for sorting. However, the transition to instrumental from human color evaluation is not without difficulties (NPPC, 1996). The use of appropriate color measurement systems can divide the color spaces into essentially equal increments across the spectrum. However, the relationship between wavelength, calculated color measures and human visual sensitivity is a non-linear one (Hunter and Harold, 1987), making direct comparisons difficult. MATERIALS AND METHODS: The PSE and DFD conditions were "created" as described by McCaw et al. (1997). Pigs (n=30) were slaughtered, carcasses were fabricated and color and ph determinations were made 24 h postmortem. The gluteus medius (), longissimus lumborum et thoracic at the 10 th rib (), semimembranosus (), biceps femoris () and triceps brachii () were removed and sliced (1.2 cm thick) perpendicular to the muscle fibers immediately prior to evaluating bloom time effects. Samples were placed on styrofoam meat trays and maintained at 6-8C during all evaluations. Instrumental color was evaluated at 0, 5, 10, 20 and 30 min immediately after cutting a fresh slice. Spectral reflectance was determined using a HunterLab MiniScan Spectrocolorimeter. L*, a* and b* were calculated based on illuminant C and the 10 standard observer (CIE, 1978). After 45 min, samples were subjected to instrumental evaluation as described above using illuminants A, C, D 65 and F, and using a Minolta CR300 (Minolta Camera Co., Ltd., Osaka, Japan) using illuminates C and D65, 2 o observer. Visual pink color intensity after 45 min bloom time was evaluated under cool white fluorescent light by a panel of 8 judges trained to use the Japanese Color Standards for pork (Nakai et al., 1973) Effects of muscle, instrument and illuminant on instrumental color measures were evaluated using Analysis of Variance (SAS, 1993). Regression analyses using the General Linear Model were used to evaluate the relationship between ph and various characteristics. The MAXR improvement procedure was used to develop prediction models for visual pink color of muscles using instrumental data. Prediction equations were developed to relate various instruments/illuminants and measurements to each other and to visual evaluations. RESUS AND DISCUSSION: Bloom time had the same effects on all muscles as far as 2

3 color characteristics using illuminant C wer concerned. Muscle affected all color characteristics measured (Table 1). had the highest L* value and hue angle, and the lowest a* value, b* value and chroma of the five muscles, indicating that it had the lightest overall color which diverged the most from the true red axis. had the lowest L* value, b* value and hue angle, and the highest a* value of the five muscles These data indicate that the selection of both a muscle and an instrumental color measure are important if color measures are to be compared from carcass to carcass, plant to plant, and/or study to study. Bloom time had no effect on L* value. However, hue angle increased for 5 min, a* and b* values increased for 10 min, and chroma increased for 20 min (Table 2). The bloom time effects indicate that instrumental color measures change to different degrees as cut meat surfaces "bloom"; some of the changes are complete in 5 min, others require 20 min before they stabilize. The absence of a muscle x bloom time effect indicates that while muscle differences exist and bloom time differences exist, each instrumental measure is affected similarly over bloom time regardless of muscle location, however the usefulness of an individual instrumental measure may be markedly influenced by the time (after cutting) when the measure was determined and by the muscle chosen for evaluation. The treatments used to produce the PSE and DFD conditions in this study resulted in 24 hr ph values () ranging from 5.13 to 7.15 and visual pink color intensity means ranging from 2.4 to 14.0 (on a 15 point scale) after 45 min bloom time (data not shown). As ph increased, visual pink color intensity increased regardless of muscle, however pink color intensity means were spread over a much wider range when sample ph was low. ph was most highly correlated with L* value, a* value, hue angle and chroma of, and most poorly correlated with L* and a* values and hue angle of. When all the muscles were pooled, the range of correlation coefficients (between visual pink color and ph) was smallest for L* value (-.67 to -.82) and largest for a* value (-.01 to.51). When visual data for all muscles was pooled, the correlation coefficient between visual pink color intensity and ph was 0.84 indicating that when a variety of muscles are evaluated, ph is a good indicator of pink color (and vice versa). The single instrumental measure that had the highest correlation (coefficient) with pink color (within muscle) is L* value; it ranged between and for all muscles except. The second highest correlation coefficient was hue angle (range = to -0.75) for all muscles except. The best single visual pink color intensity predictor was a 2-factor model using L* value and a* value generated from the (R 2 = 0.69) and a 3-factor model using L* value and hue angle generated from the (R 2 = 0.53). Instrument (Minolta vs Hunter using illuminant C or D 65 ) had significant effects on all color measures (Table 3). Illuminant (C vs D 65 within instrument) had significant effects on L* measured using the Hunter and on a*, b* and hue angle measured using the Minolta (Table 3). Differences existed due to illuminant between and within instrument. L* value was highest when determined using the Hunter / illuminant A, F and D 65. This is of practical significance because L* value has been widely reported to be one of the best indicators of pink (and red) color intensity of meat (Brewer et al., 1999; Brewer and Zhu, 1998). Extensive variation in L* value among various instruments and muscles, even when 3

4 illuminant is constant, poses potential problems. If it is necessary to choose a single color measure / muscle / instrument / illuminant combination, the correlation of these factors with visual pink intensity, based on color standards, becomes important. Based on the Japanese color standards used in this study, the instrumental measure most often and most highly correlated with visual pink color intensity, regardless of muscle and instrument / illuminant, was L* value (Table 3). The second best measure was hue angle. The poorest correlate with visual pink color was a* value. The best muscle / instrument / illuminant / measurement correlates with pink color were: 1. / Hunter / illuminanta / L* = / Minolta / illuminant D 65 / hue angle = / Hunter / illuminant F / Hue angle = / Hunter / illuninant C / L* = / Hunter / illuminant D 65 / L* = / Minolta / illuminant C / L* = Because individual industries and research laboratories are equipped with different types of instruments which may allow selection of a limited number of operating parameters, and because specific individual color measures may be of interest (eg. for "sorting"), there is a need for the capacity to make conversions (predictions) of specific color measures from one instrument / illuminant to another, if possible. The best individual models for converting individual instrumental measures determined using one instrument/illuminant into those using all other instrument / illuminant combinations were calculated. Many of the instrumental measures can be converted from one instrument / illuminant to another with a simple linear equation. L* values can be converted within instrument / illuminant combinations with resultant R 2 >0.84 and CV <5.0. Other instrumental measures are interconvertible only with complex equations. IMPLICATIONS: Bloom time affected all instrumental measures except L* value, but to varying degrees. Muscle also affected instrumental measures. As ph increased, visual pink color increased regardless of muscle, however some muscles were affected more than others. The relationship (correlation) between 24-hour ph and instrumental and visual measures of color was high. L* value was the best predictor of visual pink color intensity. had the highest correlation (coefficients) with instrumental and visual pink color measures; visual pink color could be predicted using L* and a* values of. Because L* value was unaffected by bloom time, correlated well with ph and visual pink color of most muscles, it is probably the best overall indicator the PSE and DFD conditions. The muscle, instrumental measure and bloom time must be specified prior to color determinations if subsequent comparisons are to be accurately made. Comparison of data generated using different instruments, illuminants and/or muscles may not be valid. Color evaluation should be optimized by preselecting the best combination of muscle (to be measured), color measure, and instrument/illuminant combination to be used for data collection. Standard use of specified parameters would make product sorting or comparisons from plant to plant more consistent and effective. 4

5 References Cannon, J.E., J.B. Morgan, F.K. McKeith, G.C. Smith, S. Sonka, J. Heavner, and D.L. Meeker Pork chain quality audit survey: Quantification of pork quality characteristics. J. Muscle Foods 7: CIE (Commission Internationale de l'eclairage) Recommendations on uniform color spaces--color difference equations, psychometric color terms. Supplement No. 2 to CIE Publication No. 15 (E-1.3.1) 1971/(TC-1-3). Paris. Hunter, R.S., and R. Harold The Measurement of Appearance. John Wiley and Sons, NY. McCaw, J., M. Ellis, M.S. Brewer, and F.K. McKeith Incubation temperature effects on physical characteristics of normal, DFD and halothane carrier pork longissimus. J. An. Sci. 75: Nakai, H., F. Saito, T. Ikeda, S. Ando, and A. Komatsu Standard models of pork colour. Bull. Nat. Inst. Anim. Indust., Chiba, Japan, 29:69. National Pork Producers Council, Color measurement on pork carcasses. NPPC Color Quality Meeting, Iowa State University, Ames, IA, August 8-9. SAS Users Guide SAS Institute, Inc., Cary, NC. Warner, R.D., R.G. Kauffman, and R.L. Russell Quality attributes of major porcine muscles: a comparison with the Longissimus Lumborum. Meat Sci. 33:

6 Table 1 - Muscle location effects on pork color characteristics Muscle 1 Characteristic 2 TR SEM L* value c bc a b d 0.53 a* value b 9.11 c 7.52 d b a 0.16 b* value a b c a b 0.14 Hue angle c b a c d 0.42 Chroma a c d a b = biceps femoris; = gluteus medius; = longissimus lumborum et thoracis (10th rib); = semimembranosus; TR = triceps brachii 2 HunterLab Spectrocolorimeter, illuminant C, 10 observer, 2.5 cm port 3 Hue angle = tan -1 (b*/a*) 4 Chroma = (a 2 + b 2 ) 1/2 acd Means with like superscripts are not different (p>0.05) Table 2 - Bloom time effects on pork color characteristics Bloom Time, min Characteristic SEM a* value 8.89 c b ab a a 0.16 b* value c b ab a a 0.14 Hue angle a b b b b 0.42 Chroma d c b a a HunterLab Spectrocolorimeter, illuminant C, 10 observer, 2.5 cm port. 2 Hue angle = tan -1 (b*/a*) 3 Chroma = (a 2 + b 2 ) 1/2 abcd Means with like superscripts are not different (p>0.05) 6

7 Table 3 - Instrument, illuminant and muscle effects on instrumental color measures Instrument Minolta Hunter LSD Ill-C Ill-D65 Ill-A Ill-C Ill-D65 Ill-F L* value ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 8.27 Mean b b a b a a a* value ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 1.42 Mean c b a d d d b* value ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 1.97 Mean c 7.20 a a ab b a Hue Angle ± 6.51 c ± 6.19 c ± 4.68 bc ± 6.29 c ± 7.07 b ± 6.63 a ± 6.98 ab ± 5.71 ab ± 7.04 ab ± 6.11 a ± 3.27 b ± bc ± 3.60 b ± bc ± 8.52 b ± 6.56 ef ± 6.92 e ± 3.62 d ± 8.27 e ± 3.91 cd ± 6.88 ef ± 2.91 e ± 3.18 de ± 2.35 de ± 4.68 cd ± 1.96 ef ± 6.41 e ± 2.54 de ± 2.83 e ± 5.16 cd Mean Chroma 2, ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±

8 Mean c c a b b b Mean ± standard deviation 2 No significant (p>0.05) muscle x instrument/illuminant interaction occurred for this characteristic. 3 Hue angle = Tan -1 (b*/a*) 4 Chroma = (a 2 + b 2 ) 1/2 ba,b,c,d,e,f,g,h,i,j Means with unlike superscripts are different (p<0.05). 8

9