## 1. Introduction

Odors have various characteristics in terms of the types of sources and emission, with a low threshold limit, allowing odors to be identified at extremely low concentrations (Nagata and Yoshio, 2003). Odor analytic methods can be categorized into two different approaches: instrumental and sensory. The instrumental method conducts a microanalysis on odorous substances that represent sulfurs, aldehydes, ammonia, amines, and VOCs. The sensory method can be further divided into a direct sensory method and air dilution olfactory method. In the direct sensory method, measurements are taken on-site, while in the air dilution olfactory method, samples are taken on-site and evaluated through the human olfactory (Brattoli et al., 2011;Gebicki et al., 2016;Klarenbeek et al., 2014;Nicell, 2009).

Suprathreshold olfactometry (Odor Intensity Referencing Scale, OIRS) and threshold olfactometry (Dilution-to- Threshold, D/T's) are two categories of dilution-related measurement techniques (Gostelow et al., 2001). Suprathreshold olfactometry uses an OIRS made up of the standard odorant, usually n-Butanol, to quantify odor intensity. Threshold olfactometry utilizes a field olfactometer, which dynamically dilutes the ambient air with carbon-filtered air in distinct dilution ratios until it can just be detected. Because of its ease of handling and low cost of set-up compared to an olfactometer device, Static-Scale suprathreshold olfactometry has also been incorporated as a standard of practice by a number of odor laboratories. In standard practice, OIRS is defined and used in ASTM E544 which can have 5, 8, 10 or 12 points of intensity according to n-Butanol concentrations, and observation-matching approach is adopted (Segura and Feddes, 2010). Panel screening is done by repeatedly paring the odor intensity of a known concentration of 1- butanol vapor to the OIRS. A modified 5-point n- Butanol scale evaluation procedure was established as Korea Standard Industrial Odor Test Method (Hong et al., 2019;Park and Han, 2009;Yang, 2002). The panel selection test is performed with 3 odorants and 2 odorless test paper sheets by the 5-3 method. The panelists smell the n-Butanol solutions at five concentrations, attempting to recognize the degree of odor intensity.

Odor concentration is objective and can be quantitatively measured by using the instrumental method. Moreover, odor intensity is usually proportional to odor concentration, which reflects the objective sensation of scents being experienced by a different person. It relates to the degree of strength and magnitude of the observed sample. Some researchers used sensory methods to categorize odor judges by age, gender, smoking habits, etc (Bliss et al., 1996;Stevens and Cain, 1987;Vent et al., 2004). The results showed that women, people in their 20s, and non-smokers were more sensitive to scents than men, people in their 40s, and smokers, respectively. In addition to the characteristics of odor judges, there were vast differences between individuals (Cameron and Doty, 2013;Celesia et al., 1987;Covington et al., 1999;James Evans et al., 1995). However, previous studies only confirmed the need of olfactory function for selecting odor judges. They neither made any corrections for objectifying the intensity nor determined the correction effects.

Therefore, this study highlighted the importance of odor intensity and the need for minimizing the subjective differences among judges by correcting with odor reference. It also confirmed the results of odor intensity perceived by judges according to the n-Butanol concentration, while correcting the odor intensity using n- Butanol. The correction effect was also verified by using three types of test solutions for selecting judges.

## 2. Materials and methods

### 2.1 Reagents

The reference odorant in this study is n-Butanol, with a purity of 99.5% and free of strong odorous impurities. Four odorants are acetic acid (AA, vinegar smell), trimethylamine (TMA, rotten fishy smell), methylcyclopentenolone (MCP, caramel smell), and β-Phenylethylalcohol (PEA, rose smell). Distilled water is an odor-free diluent for n-Butanol, AA, TMA. Liquid paraffin is an odor-free diluent for MCP and PEA. Odor intensities and odorant concentrations are listed in Table 1. The n- Butanol 5-point ORIS is adopted from Korea Standard Industrial Odor Test Method and the other three odorants scale is defined in this work.

### 2.2 Pre-selection of panelist

The examiner asks the volunteers to detect the n- Butanol odor with an intensity of level 1, those who could detect any smell will pass the pre-selection and become judge candidates.

### 2.3 Panelist selection

Judges candidates are trained to recognize the n- Butanol odor with the intensity of level from 1 to 5. Then the selection test is performed by the 5-3 method. Four types of odorants are AA (1.0 wt.%, ORIS = 4), TMA (0.1 wt. %, ORIS = 3), MCP (0.32 wt.%, ORIS = 4), and PEA (1.0 wt.%), with distilled water and liquid paraffin as odorless reagents. Five sheets of test paper (a length of 14 cm, a width of 7 mm) are prepared. One centimeter of any edge of the test paper is dipped into the three different odorant solutions for three sheets of odor paper, and similarly dipped into odorless distilled water and flow paraffin (vertically for 1 minute). Each candidate receives a set of five sheets of above test paper, three of which contains three types of odor, and two containing distilled water and liquid paraffin. The candidates smell and choose three of the five sheets with odors. Those who are able to correctly match the three types of test substances to the type of scent and the corresponding paper, and detect their odor intensities as level 3 or 4, are selected as the panelist. In the present work, a panel is set up with nine male and nine female, inexperienced, non-smokers in their early to mid-20s.

### 2.4 Judgment procedure

Odor papers (140 × 7 mm, unscented paper) are prepared by applying 20 μl of the odor solution on them. All panelists smell them for 20 seconds, with a distance of 5 cm between the nose and the odor paper, and score their odor intensity scales. To minimize the influence of each scent evaluation substance, a local ventilation system is used when applying the scent evaluation substances (HYScience, AH-75, FB-50, Korea). The panelists will take at least a 15-minute break between two evaluations.

Evaluations without n-Butanol odor correction involve determining the odor intensities of the three test solutions at 5 concentrations, after the odor judgment criteria (0-5 stages) is explained based on the Korean Odor Process Test Standard, as shown in Table 2. Evaluations with n- Butanol odor correction involve determining the odor intensities referencing n-Butanol ORISs of the three test solutions at 5 concentrations, after training excise on n- Butanol ORISs.

## 3. Results and discussion

### 3.1 Perceived n-Butanol odor intensity

Figure 1 shows the averaging n-Butanol odor intensity judgment results made by the male subgroup (n=9, 45 trials), the female subgroup (n=9, 45 trials) and the full panel, with regard to the log of five n-Butanol concentrations and the Korea standard ORISs (see detailed data in Table 3). Not only the two subgroup’s perceived odor intensity curves but also the full panel’s curve exhibit a reverse “S” shape. All inflection points locate at between n-Butanol concentration level of 2 and 3 because male subgroup presents a very low odor intensity score (1.89) for n-Butanol concentration of level 3, even less than that of level 2 (2.11). The female subgroup also gives small difference for those two levels (2.11 vs 2.22). As for the full panel, the assessed intensity score of level 3 (2.06) is still less than that of level 2 (2.11) even though the concentration of level 3 is about four times as level 2’s. This means that all panelists underestimate the odor intensities for the odors above n- Butanol concentration of level 3. The reverse “S” shape of the curves and much difference between intensity results for bottom concentrations and top concentrations imply some kind of systematic error in the evaluations.

Another three auxiliary curves are helpful to understand this underestimation. One is a line plot of y = x (Curve I), representing the relationship between theoretical perceived odor intensity and standard ORISs. Another (Curve II, y = 0.47x + 0.96, r = 0.94) is a fitting curve for the full panel perceived odor intensity vs standard ORISs, which should be the same as curve I, theoretically. The third (Curve III, y = 1.61x − 2.19, r = 0.99) stands for the relationship between theoretical perceived odor intensity and n-Butanol concentration. As we can see from Fig. 1, the slope of Curve II is 0.47, less than one, the slope of Curve I, and the intersection point of these two curves near the ORIS level 2. The first two points of Curve II locate around the Curve I, but the latter three points locate far below the Curve I. Similar observation is also obtained when comparing the Curve III to Curve IV (y = 0.77x − 0.09, r = 0.95), which is the plot of the full panel perceived results vs n-Butanol concentration. The slope of Curve IV is 0.77, much less than that of Curve III (1.61), and the intersection point of these two curves also near the ORIS level 2. The first two points of Curve IV locate around the Curve III, but the latter three points locate far below the Curve III. Therefore, the full panel really underestimates odor intensities for the top three concentrations when n-Butanol correction is not performed. It is interesting to find that, a fitted straight line (y = 1.54x − 1.89, r = 0.99) for perceived intensities vs top three n-Butanol concentrations, is almost parallel to the theoretical Curve III (slope of 1.54 vs 1.61), even though the difference in intercept is a little great (-1.89 vs -2.19). If some improvements are made in the odor intensity evaluation, the panel’s perceived results would be closer to the real or theoretical ORISs.

The t-tests on the theoretical odor intensity and the perceived intensity of the male or female subgroup, show that all of the p-values for n-Butanol concentration above 1,500 ppm are less than 0.05 (95% confidence level). Thus, the differences between the theoretical and the perceived intensities are statistically significant. Namely, the perceived intensities are not the reflections of the real odor intensities. The previous study confirmed that judge training with known odor intensity significantly reduced the odor judge errors for most odorants (Nachnani et al., 2005). In the case of a simple on-site evaluation other than in a laboratory, when odor intensity training is conducted at the concentration of level 1 and 5 to make the judged intensity identical to the theoretical intensity, to the utmost extent, it is possible to reduce above evaluation errors for both the bottom and top concentrations.

The slope of a fitting equation on judged intensity vs n-Butanol concentration is 0.78 for male and 0.75 for the female subgroup, indicates that there is no significant difference between the genders (Fig. 1, Table 3). Similarly, the slope of a fitting equation on the theoretical scent intensity of n-Butanol vs the judged intensity is 0.48 for male and 0.47 for female judges, also shows an only small difference between genders. The t-test results (Table 3, p-values, 95% confidence level) on the perceived intensity of the male and female subgroup show that the p-values (0.47 - 1.0) are greater than 0.05. This means that there is no significant difference in the judged intensity based on gender. According to these results, the judged intensity of n-Butanol is not influenced by gender, which implies that it can be used as an odor intensity recognition training solution for preliminary judges. The perceived n-Butnaol odor intensities are further analyzed by using “Two Factor ANOVA with Replication” method (Table 4, gender and n-Butanol concentration, replication = 9). The p-value for n-Butanol concentration is 2.07E-06, far less than 0.05. This small p-value reveals that n-Butanol concentration has a critical effect on the judged intensity. However, the pvalue of 0.70 for gender is greater than 0.05, so gender has less effect on judged results.

### 3.2 n-Butanol correction effect on acetic acid odor intensity assessment

As mentioned above, judge training or perceiving correction with known odor intensity likely has a positive effect on the subjective assessment of unknown odor intensity. Acetic acid is selected as a sourness for appraising this effect. Five concentrations in water are prepared for acetic acid by ten-fold serial dilution of the 10% solution (concentration ratio = 10). The results of the male subgroup, female subgroup, and total group's evaluations are shown in Figure 2 (see detailed data in Table 5). The curves for all groups without n-Butanol correction exhibit reverse “S” shape, however, they transform into approximate line after correction. These changes obviously show the correction effects on the perceived sourness intensity. Moreover, most of the perceived odor intensities locate below the corresponding theoretical values except for that of the lowest concentration level of 10 ppm.

Statistical analysis reveals more information about the sourness intensity assessment. In case of no training with n-Butanol, the male is more sensitive to odor intensity change (slope of intensity-concentration equation, 0.72 for male vs 0.60 for female), and has lower detective limit (intercept 0.28 for male vs 0.56 for female) than female. ANOVA results (Table 6) show that there is no significant difference between the genders (F = 0.236 < Fcriteria = 3.96, p = 0.628 > 0.05). Namely, both gender groups can give similar perceived intensities as a whole without n-Butanol correction. This is further confirmed by two-tailed t-tests for the both on each concentrations (p = 0.321, 0.576, 0.229, 0.825, 0.362). They equivalently score the intensities at each odor concentrations. After correction, the male is still more sensitive (slope, 0.94 for male vs 0.88 for female), and has lower detective limit (intercept -0.61 for male vs 0.23 for female) than female. ANOVA result shows that there is a significant difference between the genders (F = 13.675 > Fcriteria = 3.96, p = 3.97E-4 < 0.05). Namely, n-Butanol correction seriously affects both gender groups’ perceived intensities as a whole. Two-tailed t-tests for both groups on each concentration show that they score significantly differently for intensities on the concentration of level 1 and 3 (p = 0.036, 0.214, 0.05, 0.457, 0.063). One-tailed t-tests demonstrate that the female gives greater scores than male for intensities on the concentration of level 1, 3 and 5 (p = 0.018, 0.107, 0.025, 0.229, 0.031). Thus, the correction has a severe influence for the female group on odor intensity assessment.

The male subgroup indiscriminate determines the sourness intensity with or without n-Butanol correction (ANOVA results, F = 1.52 < Fcriteria = 3.96, p = 0.221 > 0.05). This means that the correction has little effect on acetic acid odor intensity perceiving for the male group. Two-tailed t-tests on each concentration also confirm this conformity (p = 0.078, 0.536, 0.275, 0.641, 0.779). However, the male group with correction is more sensitive to odor intensity change (slope of intensityconcentration equation, 0.94 for correction vs 0.72 for no correction), and has lower detective limit (intercept -0.61 for correction vs 0.28 for no correction) than that without correction. On the contrary, before and after correction, female panelists estimate vast different odor intensities (ANOVA results, F = 8.33 < Fcriteria = 3.9, p = 0.005 < 0.05).

The ANOVA results indicate that the n-Butanol correction severely effects the female group’s evaluation of sourness intensity. Two-tailed t-tests on each concentration show that they score significantly differently for intensities on the concentration of level 3 and 5 (p = 0.426, 0.240, 0.031, 0.353, 0.0016). This is also the source of difference between both genders after correction. Similarly, correction increases female group’s sensitivity to sourness (slope of intensity-concentration equation, 0.88 for correction vs 0.60 for no correction) and lowers their detective limit (intercept 0.23 for correction vs 0.56 for no correction). Correction also increases full panel’s sensitivity to sourness (slope of intensity-concentration equation, 0.91 for correction vs 0.66 for no correction) and lowers their detective limit (intercept -0.19 for correction vs 0.42 for no correction). However, ANOVA shows that there is no significant difference between full panel’s evaluation results before and after correction (p = 0.265).

All groups’ evaluation results are compared with theoretical or standard odor intensity scales. The full panel’s assessments before and after correction are very different from the theoretical values (ANOVA, p = 4.429E-10 and p = 1.788E-06, before and after). This shows that the full panel can’t judge the sourness intensity accurately, whether it is corrected or not. The male group's assessment results, show that the perceived values on concentration of level 2, 4 and 5 before correction (p = 0.279, 0.025, 0.085, 0.044, 0.003; onetailed t-tests), and all scores for each concentration level are remarkably less than the theoretical values (p = 0.001, 0.004, 0.005, 0.025, 0.001; one-tailed t-tests). The female group’s scores before correction are significantly less than theoretical values (p = 0.011, 0.004, 0.001, 0.025, 1.79E-05; one-tailed t-tests).

However, after correcting with n-Butanol, there is no significant difference between their judged scores and the theoretical ones (p = 0.347, 0.397, 0.681, 0.559, 0.195; two-tailed t-tests). This shows that correction with n- Butanol has diametrically opposite effect on male and female group’s evaluation of acetic acid odor intensity. After correction, the male group scores lower, while the female scores higher. Furthermore, the female can accurately score the sourness intensity according to the ORIS. The full panel's perceived intensity for the lowest concentration before correction is notably higher than the theoretical one (1.44 vs 1.0， p = 0.028, one-tailed ttest). After correction, there is no significant difference between the two values (0.89 vs 1.0, p = 0.303, one-tailed t-test). As for other concentrations, the judged scores are remarkably less than the corresponding ones, not mater with or without correction (before correction: p = 3.88E- 4, 9.13E-4, 0.004, 7.29561E-07; after correction: p = 0.006, 0.012, 0.044, 9.13E-4, one-tailed t-test). Therefore, for the evaluation of sour odor intensity, female panelists with n-Butanol correction are more reliable and accurate than male or mixed gender panels. It would be very simple and useful to accurately evaluate sourness intensity on-site by female panelists using 2-point intensity calibration with acetic acid after n-Butanol correction.

### 3.3 n-Butanol correction effect on trimethylamine odor intensity assessment

TMA is selected as a fishy smell for appraising n- Butanol correction effect on odor intensity assessment. Five concentrations in water are prepared for TMA by ten-fold serial dilution of the 10% solution (concentration ratio = 10). The results of the male subgroup, female subgroup, and full group’s evaluations are shown in Figure 3 (see detailed data in Table 7). The curves for all groups without n-Butanol correction exhibit reverse “S” shape, are bent lines with two points at concentration level 2 and 4.

However, after correction, they transform into approximate straight with a slight bend at concentration level 3. These changes obviously show the correction effects on the perceived fishy odor intensity. Moreover, all perceived odor intensities locate below the corresponding theoretical values. The perceived intensities of the lowest concentration level of 10 ppm are close to zero no matter what gender and correction. The concentration of 10 ppm is almost thousands of times of the olfactory threshold of 0.032 - 5.89 ppb (Nagata and Yoshio, 2003;Schiffman et al., 2001), however, the Korean panelist cannot seem to detect it. Populations inhabiting different environmental conditions have different olfactory sensitivity (Sorokowska et al., 2015, 2013). Maybe Korean's living near the sea and habit of eating seafood result in their insensitivity to fishy smell.

As for the male panelists, n-Butanol correction has a strong effect on the TMA odor intensity assessment (Table 8, ANOVA, p = 1.46E-4 < 0.05). The male group with correction is more sensitive to fishy odor intensity change (slope of intensity-concentration equation, 1.12 for correction vs 0.93 for no correction), and has a similar detective limit (intercept -0.81 for correction vs -0.82 for no correction) with that without correction. One-tailed ttests on each concentration show that their perceiving results on the concentration level of 3, 4 and 5 clearly increase after correction (p = 0.332, 0.088, 0.046, 0.020, 0.017). Before correction, the male's results on each concentration are lower than theoretical values (p = 0, 0.004, 0.011, 0.005, 0.002, one-tailed t-tests), but after correction, their scores on concentration levels of 3, 4 and 5 are very close to the theoretical ones (p = 4.37E-05, 0.051, 0.594, 0.169, 0.035, two-tailed t-tests).

As for the female panelists, n-Butanol correction also has a strong effect on the TMA odor intensity assessment (ANOVA, p = 0.0025 < 0.05). The female group with correction is slightly sensitive to fishy odor intensity change (slope of intensity-concentration equation, 1.07 for correction vs 0.99 for no correction), and has a higher detective limit (intercept -0.67 for correction vs - 0.83 for no correction). One-tailed t-tests on each concentrations show that, their perceiving results on the concentration level of 2 and 5 clearly increase after correction (p = 0.277, 0.040, 0.082, 0.358, 0.015). Before correction, the female’s results on each concentration are lower than theoretical values (p = 2.18E-05, 0.001, 0.007, 0.007, 0.001, one-tailed t-tests), however, after correction, similar to the male group, the female’s scores on concentration levels of 3, 4 and 5 are also very close to the theoretical ones (p = 0.0007, 0.104, 0.169, 0.104, 0.035, two-tailed t-tests).

Gender factor almost has no effect on panelist evaluating the fishy odor intensity. Before correction, ANOVA result on gender factor shows the p value is 0.310, greater than 0.05; after correction p-value is 0.853, greater than 0.05. Results of two-tailed t-tests on each concentration also lead to the same conclusion (p = 0.347, 1, 0.710, 0.304, 0.444 for correction and p = 0.555, 0.750, 0.661, 0.641, 1 for no correction).

As for the full panel, correction with n-Butanol slightly increases their sensitivity to fishy smell intensity changes (slope of intensity-concentration equation, 1.09 for correction vs 0.96 for no correction). The correction also has a strong effect on the full panel’s TMA odor intensity assessment (ANOVA, p = 6.688E-7 < 0.05). One-tailed t-tests on each concentrations show that, their perceiving results on the concentration level of 2, 3 and 5 clearly increase after correction (p = 0.302, 0.023, 0.024, 0.074, 0.002). Comparing with the theoretical values, both perceived values before and after correction are significantly different from the theoretical values (ANOVA, p = 5.23E-13 for no correction and p = 2.75E- 26 after correction). Before correction, the full panel's results on each concentration are lower than theoretical values (p = 2.09E-12, 1.4E-5, 2.61E-4, 1.33E-4, 1E-5, one-tailed t-tests), however, after correction, only the score on concentration levels of 3 is close to the theoretical one (p = 2.5E-8, 0.004, 0.094, 0.015, 9.1E-4, two-tailed t-tests).

Thus, n-Butanol correction remarkably increases the male, female and full panel’s scores of fishy smell intensity on the concentration levels of 2 to 5. The TMA concentration of 10 ppm is too low for the panel in the study to score according to the ORIS and is not applicable in odor intensity training. It would be very simple and useful to accurately evaluate fishy smell intensity on-site by any gender panelists using 2-point (0.01% - 10%) intensity calibration with TMA after n-Butanol correction.

### 3.4 n-Butanol correction effect on methylcyclopentenolone odor intensity assessment

MCP is selected as a caramel smell for evaluating n- Butanol correction effect on odor intensity assessment. Five concentrations in liquid paraffin are prepared for MCP by ten-fold serial dilution of the 3.2% solution (concentration ratio = 10). The results of the male subgroup, female subgroup, and full group’s evaluations are shown in Figure 4 (see detailed data in Table 9). The curves for all groups without n-Butanol correction exhibit reverse “S” shape are bent lines with two points at concentration level 2 and 3. However, after correction, plots of the intensity scores versus concentration slightly leveled off at the bottom concentrations. These changes obviously show the correction effects on the perceived caramel odor intensity. Moreover, all perceived odor intensities locate below the corresponding theoretical values. Curves before correction intersect that after correction at a certain concentration between level 3 and 4, indicating that n-Butanol correction has opposite effects on intensity assessment for low and high MCP concentrations. ANOVA (Table 10) shows that, there are no notable difference between the perceived intensities of the concentration level 1 and 2 no matter what gender and correction (for male, concentration effect: p = 0.741, correction effect: p = 0.325; for female, concentration effect: p = 0.338, correction effect: p = 0.748; for full group, concentration effect: p = 0.349, correction effect: p = 0.639). In the following analysis, the perceived result of the concentration level 1 is excluded, and only the latter four points are considered.

As for the male panelists, the correction has a strong effect on their evaluation of caramel odor intensity (p = 0.029). The interaction of concentration and correction also significantly affects their measuring (p = 0.001). They give a similar score for the concentration level 2, a lower for level 3 and higher scores for level 4 and 5 after correction (p = 0.326, 0.040, 0.031, 0.002, one-way tailed t-tests). For the female, on the contrary, the correction has a weak effect on them (p = 0.232), even the interaction of concentration and correction is remarkable (p = 0.030). The female group gives similar scores for concentration level 2 and 3, and higher scores for level 4 and 5 after correction (p = 0.5, 0.088, 0.021, 0.017, oneway tailed t-tests). The correction also strongly affects the full panel's evaluation (p = 0.016), and the interaction of concentration and correction is significant (p = 0.030). Similar to the male group, the full panel gives a similar score for the concentration level 2, a lower for level 3 and higher scores for level 4 and 5 after correction (p = 0.382, 0.017, 0.0025, 1.14E-4, one-way tailed t-tests).

Gender has no significant effect on scoring the caramel smell intensity (no correction: p = 0.820, correction: p = 0.555). Two-way tailed t-tests further show that there are no notable differences between both genders on each concentration before (p = 0.653, 0.594, 1, 0.265) and after correction (p = 1, 0.461, 0.663, 1). All groups’ judged values are notably lower than the theoretical ones (all pvalues less than 0.01), indicating larger errors between measurements and theoretical values whether correction or not. In other words, n-Butanol correction doesn’t improve the accuracy of the caramel odor intensity evaluation.

Before correction, the female is more sensitive to caramel intensity change than the male (slope of intensityconcentration equation, 1.08 for female vs 0.93 for male), while they have almost similar sensitivity after correction (slope, 1.37 vs 1.36). All groups’ sensitivity to caramel intensity increase (slope, 1.36 vs 0.93 for male, 1.37 vs 1.08 for female, 1.36 vs 1.01 for full group), and the detective limits decrease after correction (intercept, -0.69 vs -1.68 for male, -1.10 vs -1.80 for female, -0.90 vs -1.74 for full group). The slopes are around 1.36, greater than the theoretical slope of 1, implying that the judged intensities for high concentrations are more closer to the theoretical values, but those values for low conventions may greatly deviate from the true values. The full panel’s evaluations before correction give a slope of 1.01, almost paralleling to the theoretical curve, showing that similar bias exists for low and high concentrations, and could be calibrated with an adjusted intercept. Therefore, the full panel could accurately assess the caramel intensity without n-Butanol correction comparing with a single gender. It would be very simple and useful to accurately evaluate caramel smell intensity on site by full panel using 2-point (0.0032% - 3.2%) intensity calibration with MCP without n-Butanol correction.

## 4. Conclusion

The present work shows n-Butanol correction affects the odor intensity evaluation for three odorants in different ways. In most cases, n-Butanol correction increases the panelists’ sensitivity to the odor intensity change, and resulting in a high score closing to the theoretical value. The female panelists can evaluate accurately the sourness intensity of acetic acid after n- Butanol correction. Any gender panelists can accurately evaluate the fishy smell intensity of TMA after n- Butanol correction. For evaluating caramel smell intensity of MCP, a full panel without n-Butanol correction is suggested. In this study, only three types of test solutions were used for selecting judges and confirming n- Butanol’s intensity correction effect. In the future, more evaluations will be conducted on various concentrations of a greater variety of substances.