Impact of Crust Creation on Techno-Functional and Organoleptic Properties of Meat with Different Fat Contents

Abstract

In this study, the effects of crust formation were investigated on the quality properties of pork meat, including texture, colour, cooking loss, pH, water-holding capacity (WHC), and sensory attributes. Two cuts of pork, (1) tenderloin (lean) and (2) collar (fatty), were subjected to three treatments as follows: (1) crusting before cooking, (2) crusting after cooking, and (3) cooking without crusting. Texture was analyzed measuring hardness, cohesiveness, springiness, and chewiness. Colour was analyzed measuring lightness, redness–greenness, yellowness–blueness, and hue angle. Cooking loss and pH and WHC were measured using standard methods and a pressing test. Results showed that meat type and treatment significantly impacted texture, with tenderloin exhibiting greater hardness and chewiness than collar (p = 0.05). Crusting after cooking produced the highest values for hardness and chewiness, while cohesiveness and springiness were unaffected by treatment (p = 0.05). Colour measurements revealed significant treatment-induced differences, with samples crusted after cooking being darker and redder, on average 14% and 33%, respectively. Cooking loss varied considerably, with no significant treatment differences, although a trend suggested lower loss in samples crusted before cooking. pH values were slightly higher in crusted samples, while WHC was notably higher for samples crusted after cooking, likely due to oil retention.

1. Introduction

Meat is an important element of the human diet, providing high-quality proteins and essential micronutrients [1,2]. However, for consumers, sensory quality is one of the most critical attributes of meat [3]. Among sensory attributes, the most important are the tenderness, juiciness, flavour, and aroma of cooked meat [4,5]. These sensory qualities can be assessed through both instrumental and subjective sensory evaluations [6]. Improving the properties of ready-to-eat foods, such as roasted meats, requires a comprehensive understanding of how to develop crusts with optimal techno-functional properties that maximize consumer benefits.

Crusting or sealing is defined as the formation of a crust, a denatured layer, on the surface of meat at high temperatures [7]. Through the Maillard reaction, the surface of the meat is heated between 160 °C and 250 °C to achieve a more attractive and flavourful crust [8]. This crust, along with the addition of natural antioxidant components, can even protect the meat from lipid peroxidation [9]. Heterocyclic aromatic amines, known to have carcinogenic effects in animals and on cell lines, are also formed during crusting [10]. Crust formation results in darkening and reddening of the meat’s surface [11]. Various marinades are often used to influence the colour, flavour, and other quality characteristics of the meat [12]. A summary of previous studies can be seen in

Table 1. Summary of previous studies.

Study	Material	Methods	Findings
Kilic et al., 2023 [7]	Beef muscles (M. Longissimus thoracis et lumborum) were purchased from Erzurum Meat Combine. M. Longissimus thoracis et lumborum muscles obtained from three different male cattle of the same race, gender, and age at 24 h after slaughter.	Sealing on a griddle pan until 250 °C meat surface temperature, then until 20, 30, or 40 °C internal temperatures. (Control was not sealed.) After the sealing process, the samples were baked in an oven at 180 °C until 71 °C internal temperature.	The high cooking loss and greater rate of protein denaturation observed in the sealed samples suggested that sealing may not be a way to make the meat juicier. Moderate sealing reduced the hardness of the meat. The sensory evaluation showed that the general acceptability of the longer sealed samples was higher because of the appearance change caused by the Maillard reaction.
Barber and Broz, 2011 [8]	Four cuts of rib eye beef steaks from the rib primal taken from the same primal with a nearly identical ratio of fat to lean.	Heat treatment at 205 °C in electric oven for 10 min. The searing pan was heated for 5 min at the burner’s highest setting, then meat slices were heat treated on the pan for 10 s per side.	Searing improved the flavour and caused browning through the Maillard reaction. But flavorful juices were not preserved, the cooking loss increased.
Asido et al., 2024 [9]	Ground turkey and entrecote from commercial stores and kept at −80 °C until use.	Heat treatment on an electric stove to 180 °C. The cuts were placed on the pan without oil and fried for 3 min exactly per side or were microwave treated at 800 W for 1 min. Internal temperature reached 70 °C.	The crust, as a semi-natural antioxidant, reduced lipid peroxidation and possibly extended shelf life.
Kondjoyan et al., 2016 [10]	Longissimus thoracis muscles (ribeye or loin eye) taken from carcasses of 18-month-old heifers immediately after slaughter aged for 12 days under vacuum-packing and stored at −20 °C.	Heat treatment with 160, 190, 225, or 260 °C air temperature during 20, 30, 60, and 90 min cooking times.	Less heterocyclic aromatic amines form in coloured crusts developing on the surface of meat pieces than in meat slices.
Wang et al., 2021 [11]	Pork legs with moderate fat from 10 pig carcasses (100–120 kg live weight, about ten months of age and 24 h post-mortem).	Frying at 150, 175, 200, 225, and 250 °C during 0.5, 1.0, 1.5, 2.0, and 2.5 min.	L* decreased, a*, b*, and amount of heterocyclic aromatic amine content increased with increased frying temperature and time.
Okpala et al., 2023 [12]	Pork neck (collar).	Marination then oven-grilling at 180 °C during 5 min. Internal temperature reached 75 °C.	Marination after heat treatment had an effect on chemical components, pH, colour, and texture.

.

Cooking loss is another important characteristic of cooked meats. Previous research has investigated whether creating a crust before cooking reduces cooking loss and juiciness, with findings suggesting it does not [7,13,14]. However, other studies have reported different results [8,15]. For pork, cooking loss is typically around 30% [16].

It is noteworthy that despite the mathematical description of the dynamics of crust formation [17], there has been a limited focus on the instrumental and sensory evaluation of meat quality attributes using various crusting methods.

This study investigates, from a practical perspective, the impact of meat type and crusting methods on both the quality of the meat and the characteristics of the crust formed. The aim of the present study is to investigate the effect of fat content (collar—fatty; tenderloin—lean) and crust formation method (crusting before cooking, crusting after cooking, and cooking without crust formation) on various techno-functional properties (texture, colour, cooking loss, pH, and water-holding capacity), as well as the sensory attributes of the meat cuts used. The novelty of the study is that no study has yet compared the effect of crust formation in the fatty and lean meat parts of the same animals, and previous research has not compared what occurs when the order of technologies (crusting, cooking) is changed.

2. Materials and Methods

2.1. Material

In this study, commercially available pork cuts, sourced directly from a meat plant, were utilized. The DanBred pigs (Sus scrofa domestica) were born, raised, and slaughtered in Hungary, a member of the European Union. The pork was obtained from female animals aged 6–7 months, raised under industrial farm conditions in accordance with Central European standards. These conditions complied with the technological requirements for meat from domestic ungulates, as outlined in Regulation (EC) No. 853/2004 of the European Parliament and of the Council of 29 April 2004, which established specific hygiene rules for foodstuffs. The meat underwent a 1-day cooling period between slaughter and deboning and cutting. The slaughter of livestock, along with the deboning, cutting, and vacuum-packing of meat, was conducted in an industrial slaughterhouse. The samples were taken from different livestock (from different 30 animals). The meat samples were treated and measured the day after being vacuum-packed, similar to the handling of commercially purchased meat. Two types of meat cuts, representing extreme levels of fat content—the fattiest and the leanest cuts—were selected for this study. These were the tenderloin and the collar, both of which can be prepared using the same cooking technique. For the experiment, 2 cm thick slices were cut from independent samples of both tenderloin and collar. Between the steps of vacuum-packing and cooking, the meat samples were stored in their packaging under refrigerated conditions at an average temperature of 3.2 °C. Prior to cooking, the samples were allowed to warm to room temperature. The surface of the meat reached 20 °C, while the core temperature remained at 14 °C before cooking. No seasoning or marination was applied to the meat, as this could have led to the release of exudates, potentially influencing cooking loss and water-holding capacity.

Three different treatments were applied. In total, 30 samples (from 30 different animals) were taken from each meat cut, with 20 samples per treatment type for each cut. So, 30 tenderloin and 30 collar samples, meaning 60 samples in total were treated. Of these, they were divided into 20-20-20 samples, so 10-10-10 samples from both meat types were treated in different ways. The replicate samples were measured multiple times by different methods, and the average of the parallel measurements of each sample was taken as the measured value of a sample for the evaluation of the results. Thus, the degrees of freedom of the statistics presented reflect the number of parallel samples. In the case of colour measurements, the sample number is double the sample number presented for other measurements because two different sides of the samples were interpreted as different samples.

The meat cuts were cooked in a combined steaming oven (Lainox VE051P, Type LX, Pozzo d’Adda, Italy) using the oven’s metal tray, covered with aluminum foil, in convention mode at 200 °C for 40 min. Based on earlier empirical observations and after reviewing several standard recipes and conducting preliminary tests, a temperature–time combination of 200 °C for 40 min was determined to be optimal, considering the thickness of the slices and the specific types of meat used. Bark crusting was performed in a non-stick pan with sunflower oil. The initial pan temperature for crusting was 230 °C. Three types of treatments were applied. One-third of the samples from each meat cut underwent crust formation before cooking, one-third after cooking, and the remaining third were cooked without crust formation. For crusting before cooking, the collar was crusted for 5 min per side, while the tenderloin was crusted for 1.5 min on the bottom side and 1 min on the top side. For crusting after cooking, the collar was crusted for 1.5 min per side, while the tenderloin was crusted for 0.5 min per side. These time intervals were sufficient to achieve the desired crust. The crusting technique was based on non-scientific culinary sources and methods applied by renowned chefs, and it was validated through preliminary experiments. The treatment conditions are presented in

Table 2. Treatment conditions of the experimental materials.

Slaughtering	with full bleeding slaughtering technology
Cooling	1 day to below 8 °C core temperature
Cutting, deboning, vacuum-packaging, and cooling	1 day, technologies below 12 °C, cooling below 4 °C core temperature
Resting before baking	until 20 °C surface temperature and 14 °C core temperature
Crust forming before cooking	in non-stick pan on sunflower oil, which had a temperature of 230 °C at the start of crusting: collar: 5 min per side tenderloin: 1.5 min for the bottom side and 1 min for the top side
Cooking	in tray covered with aluminum foil, in the oven with air-mix mode at 200 °C for 40 min
Crust forming after cooking	in non-stick pan on sunflower oil, which had a temperature of 230 °C at the start of crusting: collar: 1.5 min per side tenderloin: 0.5 min per side

.

2.2. Method

2.2.1. Experimental Design

This study presents the results of a 2 × 3 full factorial experimental design aimed at comparing the effects of two different meat cuts with varying fat content (collar—fatty; tenderloin—lean) and three distinct treatments (crusting before cooking, crusting after cooking, and cooking without crust formation). This design yielded six combinations of factor levels. Different sample sizes were used for each measurement, as outlined in the statistical results. The primary objective of the research was to assess the impact of these factors (fat content, meat dryness, and treatment method) on various techno-functional and sensory properties, with a focus on evaluating the quality and cooking loss of the meat cuts.

According to the USDA Food Data Central, the tenderloin contains 3.9 g of fat per 100 g of raw meat, while the collar contains 5.7 g of fat per 100 g.

It is important to note that the connective tissue content also influences the measured characteristics. Although this study specifically focused on the leanest and fattiest cuts of pork, which are commonly consumed in Europe, it is worth mentioning that the fattier collar cut contains a higher proportion of connective tissue.

Sample sizes were determined based on preliminary experiments. After the experiment, the sample number was doubled and the statistical analysis was performed again. The paper now contains only the final sample number and statistical indicators.

2.2.2. Texture Measurement

The texture of meat samples was assessed using the Stable Micro Systems (SMS) TA XT Plus texture analyser (Stable Micro Systems, Godalming, UK). The movement of the upper probe was controlled by a computer. The instrument allows for the use of various probes to measure different types of shears, compressive, and torsional stresses applied in multiple directions and over varying durations, thereby enabling a comprehensive analysis of the samples’ textural properties.

Texture profile analysis (TPA) was performed, and the following attributes were measured: (1) Hardness: The resistance [N] at maximum compression during the first compression cycle. It represents the force required to achieve a specific deformation and reflects the hardness of the sample at the initial bite. (2) Cohesiveness: The ratio [-] of the positive force during the second compression cycle to that of the first cycle (downward strokes only). It indicates the strength of the internal bonds that maintain the structure of the sample. (3) Springiness: Expressed as a ratio [mm·mm⁻¹] or percentage of the product’s original height. Springiness is typically measured by dividing the height during the second compression cycle by the original compression distance. (4) Chewiness: The force [N] required to chew a solid sample until it reaches a steady state for swallowing. It is calculated as the product of hardness, cohesiveness, and springiness [18,19,20]. The meat samples were diced (40 mm × 40 mm) with a height of 20 mm. A P/75V steel cylinder plate was used as the probe. The pre-test speed of the probe was 10 mm/s, with both the test speed and post-test speed set to 1 mm/s. Maximum compression was set to 50% of the sample height (10 mm) and lasted for 1 s, after which the sample was allowed to relax. The sampling frequency was 50 Hz, and each sample was measured three times.

A limitation of this investigation may arise from the uneven distribution of adipose and connective tissue in mammalian muscles. Although the number of samples was considered adequate, there is a possibility that some pieces contained more or less connective tissue than average. However, since no outliers were observed, this effect was deemed to have a minimal impact on the results.

2.2.3. Colour Measurement

A Minolta CR-400 Chroma Meter (Konica Minolta, Inc., Tokyo, Japan) was utilized for reflection colour measurements. This instrument evaluates the CIE Lab colour characteristics by plotting the ratio of three different wavelength lights within the CIE Lab colour space. The colour data are quantified by the device [21]. Calibration of the instrument was performed using a standard white etalon. Each sample was measured in triplicate on a white plate under consistent illumination to minimize potential interference. The chroma metre’s measuring head was positioned directly on the surface of the treated meat samples. A Minolta CR-400 colorimeter was used with an 8 mm aperture, a 2° observer, an illuminant D65, and a pulsed xenon lamp as the default light source. During the measurement, a glass cover with a light shield tube (CR-A33a, Konica Minolta, Tokyo, Japan) was placed over the aperture. The convex protective glass plate prevents the sample from entering the instrument and facilitates cleaning between measurements. The instrument was calibrated with a white tile standard before each analysis. The illumination of the samples was uniform and continuous throughout the measurements, which were performed on a white tray between white walls. However, it is important to highlight that the environment could not influence the colour measurement, because the measuring head was directly in contact with the surface of the samples so that no light other than the light from the chroma metre was shining on the measured spots. The colour attributes assessed included redness/greenness (a*), yellowness/blueness (b*), and lightness (L*). Additionally, the hue angle (h(ab)) was calculated as a significant colour attribute using the following formula [22]:

h_ab = arctan (b*)/(a*)

(1)

2.2.4. Determination of Cooking Loss

The mass of the sliced raw meat samples was individually measured using a KERN EMB 1200-1 scale (Kern and Sohn GmbH, Balingen, Germany) both before and after treatment following cooling to room temperature. The difference between the two mass measurements was used to calculate the cooking loss. Cooking loss was expressed in grams per 100 g of the initial mass.

It is important to note that the measurement was influenced by the oil used during the crust formation process. For samples where the crust was formed after cooking, the cooking loss was reduced by the mass of oil absorbed and adhered to the samples.

2.2.5. Determination of pH

The pH of each sample was measured using a Testo 206 pH stick (Testo GmbH, Innsbruck, Austria) [23]. Prior to each measurement series, the device was calibrated with two standard liquid buffers (pH 4.0 and pH 7.0). Each sample was measured in triplicate to ensure the accuracy and reliability of the results.

2.2.6. Determination of Water-Holding Capacity and Fattiness

The test conducted in this research is originally a determination method of water-holding capacity (WHC). However, this method is also suitable for quantifying the moisture lost by the meat during pressing. A pressing test was performed based on a modified methodology of Grau and Hamm for measuring water-holding capacity [24,25]. The result of this test, expressed in mm²·mg⁻¹, indicates the surface area of filter paper that can be moistened by 1 mg of meat sample under mechanical pressure, accounting for the amount of water and fat that cannot be retained by the sample.

A 1.5–2 mg sample, measured using an analytical balance, was pressed onto a filter paper placed between glass layers using a 0.5 kg mass for 5 min. The moisture released by the meat formed a blot (patch) on the filter paper. The mass of the entire filter paper, including the non-moistened portion, was measured using a Kern ABJ-NM/ABS-N (Kern and Sohn GmbH, Germany) analytical balance. The moisture transferred from the meat into the filter paper was composed of water, fat, and oil released during the crust formation process after cooking. Each sample was measured in triplicate.

2.2.7. Sensory Analysis

A questionnaire-based sensory evaluation was conducted in a controlled laboratory environment, free from confounding factors, with panel members spaced in time and/or location to prevent bias. The samples used in the evaluation were the analytical samples, which were cut into small, uniform pieces. Each sample was assigned a random three-digit numerical code, and the panel members received the samples in a randomized sequence to minimize the potential influence of sequential tasting on their perceptions. To neutralize their palates between samples, panel members were provided with water and bread. The panel consisted of 30 participants, all of whom were non-expert regular consumers who had not undergone specific training for the test. The participants, a balanced mix of men and women aged 18 to 45, were not asked to provide demographic information. Panel members were instructed to evaluate and rank their preferences for colour, aroma, texture, taste, and overall characteristics using a 10-point category scale in a descriptive sensory test. They indicated their preferences on pre-printed forms, with ratings ranging from 1 to 10 for each attribute. The order of attributes was arranged according to the typical sequence in which human senses perceive them. Participants were also encouraged to provide written comments. The collected data were subsequently entered into a spreadsheet programme for further analysis.

2.2.8. Statistical Analysis

The measurement data were analyzed using IBM SPSS Statistics v28 (IBM Corp., Armonk, NY, USA) and presented in Microsoft Excel 365 (version 2010, build: 13328.20356). To evaluate the effects of meat type and treatment on texture, colour, cooking loss, pH, WHC, fattiness, and sensory attributes, a range of statistical methods were employed, tailored to the specific dependent variables. A multivariate analysis of variance (MANOVA) was conducted to compare the means of different sample groups for related variables [26]. The normality of residuals was assessed using the Shapiro–Wilk test [27]. Although the normality of residuals for sensory attributes did not fully meet the assumptions required for the MANOVA, the analysis was nonetheless performed across all three attribute groups. The homogeneity of variances was tested using Levene’s test [28]. Given that a MANOVA is robust to violations of the equality of covariance matrices, Box’s M test [29] was applied to sensory attributes that did not satisfy the assumption of homogeneous covariance matrices according to Levene’s test. The MANOVA was applied to each dependent variable, and the unexplained variance (Wilks’ lambda) was evaluated. Homogeneous groups were determined using Tukey’s HSD post hoc test.

3. Results

3.1. Texture

After evaluating the equality of error variances [Hardness: F(5,54) = 1.368, p = 0.251; Cohesiveness: F(5,54) = 1.010, p = 0.421; Springiness: F(5,54) = 0.667, p = 0.650; Chewiness: F(5,54) = 1.926, p = 0.105] and the normality of residuals [Hardness: W(60) = 0.954, p = 0.023; Cohesiveness: W(60) = 0.974, p = 0.231; Springiness: W(60) = 0.960, p = 0.044; Chewiness: W(60) = 0.922, p = 0.001], a MANOVA was performed. Given that meat samples are biological materials with unevenly distributed adipose and connective tissues, the texture measurements exhibited a considerable degree of variability. Despite this, the analysis of unexplained variance revealed that both meat type and, consequently, the fat (and connective tissue) content [Wilks’ lambda = 0.322, F(4,51) = 12.245, p < 0.001] and the treatment [Wilks’ lambda = 0.386, F(8, 102) = 3.127, p < 0.001] exerted significant effects on the texture properties of the samples. Additionally, the interaction between these two factors [Wilks’ lambda = 0.367, F(8,102) = 3.009, p < 0.001] was also statistically significant, albeit with a lesser degree of influence. The meat type and, consequently, the fat content had the most significant effect. Based on the results of the Tukey post hoc test, no significant differences were found in cohesiveness and springiness with respect to the treatment (p = 0.05). However, for hardness, samples crusted after cooking exhibited an average 33% higher value compared to the other samples. This difference in hardness resulted in a 37% increase in chewiness for the samples crusted after cooking, which was statistically significant (p = 0.05). No significant differences were observed in cohesiveness and springiness according to meat type; however, significant differences were found in the hardness and chewiness between tenderloin and collar samples (p = 0.05). Tenderloin samples were, on average, 53% harder and 40% more chewy than collar samples. This difference was expected, as tenderloin is a drier and denser cut of meat. In contrast, sliced collar contains several smaller pieces of meat encircled by fat, resulting in a more tender and softer texture. This distinction was clearly measurable in the experiment. The results can be explained by the fact that unlike in the tenderloin, the high amount of fatty tissue that permeated the collar resulted in a less chewy texture with a lower hardness. A tenderloin is basically a muscle without fatty tissue. The method chosen for the treatment ensured that it did not dry out, yet it was clearly harder. In the case of tenderloin, the crust creation after cooking created a hard layer on the surface of the meat, which accounts for a larger proportion of the total sample size than in the case of collar. This explains the very high hardness of the sample. A previous study also measured higher hardness and chewiness in lean meat compared to adipose tissue [30]. Grilled (crusted) meats also showed increased chewiness compared to meat prepared without crusting (cooking sous-vide) in a previous study [31]. Since chewiness is the multiplication of hardness, cohesiveness, and springiness, as well as hardness being four orders of magnitude higher than the other two variables, it is easy to explain why chewiness increases with hardness. The hardness is increased mainly by the heat denaturation of meat proteins and water loss in the areas below the crust. Considering the Wilks’ lambda values obtained from the statistical analyses, it is clear that both meat type and total fat content influenced the experimental results. The strongest effect was observed in the texture results, and for texture, the effect of meat type was greater than that of the treatment when the Wilks’ lambda was below 0.5. Texture parameters are shown in Figure 1, Figure 2, Figure 3 and Figure 4.

For improved clarity, the key results of the statistical analysis for texture, as well as all other variable groups, are presented in

Table 3. Results of MANOVA in case of all examined dependent variants.

Dependent Variants	Factor	MANOVA Result
		Wilks’ Lambda	p Value
Texture attributes	Meat type	0.322	<0.001
	Treatment	0.386	<0.001
	Interaction	0.367	<0.001
Colour parameters	Meat type	0.617	<0.001
	Treatment	0.277	<0.001
	Interaction	0.499	<0.001
Cooking loss, pH, WHC—fattiness	Meat type	0.549	0.004
	Treatment	0.196	<0.001
	Interaction	0.442	0.005
Sensory properties	Meat type	0.617	<0.001
	Treatment	0.671	<0.001
	Interaction	0.608	<0.001

.

3.2. Colour

Following an assessment of the equality of error variances [L*: F(5,114) = 2.962, p = 0.015; a*: F(5,114) = 2.792, p = 0.020; b*: F(5,114) = 6.389, p < 0.001; h(ab): F(5,114) = 5.165, p < 0.001] and the normality of residuals [L*: W(120) = 0.984, p = 0.184; a*: W(120) = 0.974, p = 0.021; b*: W(120) = 0.934, p < 0.001; h(ab): W(120) = 0.968, p = 0.005], a MANOVA was conducted. The unexplained variance analysis indicated that both meat type and treatment had significant effects. However, the type of meat—and by extension, the fat content—was not a determining factor [Wilks’ lambda = 0.617, F(4,111) = 28.221, p < 0.001]. The treatment effect on colour was found to be substantial [Wilks’ lambda = 0.277, F(8,224) = 24.969, p < 0.001]. The interaction between the two factors was also significant [Wilks’ lambda = 0.499, F(8,224) = 11.538, p < 0.001]. However, given that Wilks’ lambda was just below 0.5, it cannot be definitively concluded that the interaction between the factors provides a strong explanation for the variation in sample colour. Therefore, in terms of colour, the meat type emerged as the most significant explanatory variable for the dependent measures. This observation is understandable, as the colour of fatty tissue in meat differs from that of the muscle tissue itself. This principle is employed in methods for measuring the fat content of minced meat using a simple optical sensor, without the need for analytical measurements.

In terms of the treatment’s effect, significant differences were observed based on the Tukey post hoc test (p = 0.05). The meat samples crusted after cooking were, on average, 14% darker and 33% redder than the other samples. Conversely, the meat samples crusted before cooking were 16.5% more yellow than the others. The hue angle results for samples crusted after cooking were, on average, 17% lower than those for the other samples. The hue angle values for all samples were within the range of 0–90°, centred around 66°, which corresponds to an orange colour. However, with lower lightness scores, this indicates a brownish colour. A similar colour has also been described as golden brown [16].

This result can be easily explained by the fact that the surface of the meat undergoes browning during the process of crust formation. The increase in the yellow colour factor for samples crusted before cooking may be attributed to the browning of the meat surface prior to cooking, which faded during the cooking process. An interesting observation was that although meat type was a stronger determinant of meat colour than treatment, no significant differences were observed in the lightness factor or hue angle. Nevertheless, the colour of collar samples, which had a higher fat content, was 12% less red and 18% less yellow than that of the leaner tenderloin samples. In a previous experiment, there was also a decrease in the lightness and an increase in the redness of grilled (cured) meats compared to cooked and sous-vide treated meats [31]. The denatured globin hemochrome caused the red colour, and the denatured globin hemichrome caused the darker brown/tan/grey colour. This red and brown/tan/grey colour intensifies as the final temperature of the heat treatment increases [31,32]. In our study, the crust creation temperature was much higher than the cooking temperature. This explains the difference between the “without crust” and “crust after cooking” sample groups. However, samples that were crusted before cooking had lost the substances from the surface that were responsible for the discoloration during cooking, so the same trend was not observed. The color factor results are shown in Figure 5, Figure 6 and Figure 7 and the Hue angle results in Figure 8.

3.3. Cooking Loss, pH, WHC—Fattiness

Another statistical analysis was conducted to examine the results of cooking loss (Figure 9), pH (Figure 10), and water-holding capacity (WHC) (Figure 11) in relation to fat content. This approach was used based on the expectation that these three techno-functional properties would be interrelated. Typically, the distance of the current pH from the isoelectric point is inversely related to water-holding capacity, as proteins at their isoelectric point acquire a neutral charge and are unable to retain their solvation shell [33]. Furthermore, it can be anticipated that higher cooking loss would correspond to lower moisture content in the samples. As the results indicate, the relationship between the dependent variables was not straightforward. However, among the independent variables, the treatment was found to be a significant determinant of the dependent variables. After checking the equality of error variances [Cooking loss: F(5,54) = 9.328, p < 0.001; pH: F(5,54) = 3.302, p = 0.011; WHC—fattiness: F(5,54) = 16.739, p < 0.001] and the normality of residuals [Cooking loss: W(60) = 0.879, p < 0.001; pH: W(60) = 0.981, p = 0.493; WHC—fattiness: W(60) = 0.887, p < 0.001], a MANOVA was carried out. The unexplained variance associated with meat type and fat content did not demonstrate a strong relationship [Wilks’ lambda = 0.549, F(3,51) = 6.014, p = 0.004]. However, the treatment showed a significant and strong effect on cooking loss, pH, and water-holding capacity (WHC) related to fat content of the samples [Wilks’ lambda = 0.196, F(6,102) = 9.218, p < 0.001]. The interaction between the two factors [Wilks’ lambda = 0.442, F(6,102) = 3.703, p = 0.005] also had a significant, though less dominant, effect. Therefore, the treatment was the most significant determinant.

According to the results of Tukey’s HSD post hoc test, the pH of sample groups without crust formation was significantly lower than that of the other sample groups (p = 0.05). Although statistically significant, the pH of the samples without crust was, on average, 0.6% lower than that of the other samples. This may be attributed to the denaturation of proteins and the decomposition of organic acids during cooking. The small difference observed is likely due to a compound formed during the crust formation process. This change was nominally very small. The average pH of all samples was 6.08, the average pH of samples without a crust was 6.05, and the average pH of samples with a crust before or after cooking was 6.09. So, the pH of all samples was on the border of the neutral and acidic range with very little difference.

In terms of WHC related to fat content, the samples crusted after cooking exhibited significantly higher values than the other sample groups, with an average value of 0.58 mm² g⁻¹ in the case of tenderloin and a value above the measuring range in the case of collar (p = 0.05). This can be explained by the fact that the crust was formed on the samples using oil after cooking and the oil adhered to the surface of the meat. As a result, the filter papers used during the pressing test were likely moistened by the oil. The WHC related to fat content values for all collar samples crusted after cooking were above the measurement range. Since collar is basically fattier, it was probably its fat content that gave it its moisture content measured by the pressing test. The average result of all collar samples (if the value above the measuring range was calculated as 1 mm² g⁻¹) was 0.59 mm² g ^{− 1}, and the average result of all tenderloin samples was 0.25 mm² g⁻¹. On average, the results for samples that were crusted before cooking were 0.06 mm² g⁻¹ lower than the results for samples without crusting.

No significant difference was found in cooking loss between the sample groups (p = 0.05). While a trend in Figure 9 indicates that the samples crusted before cooking had the lowest cooking loss, the large variance prevents this difference from being statistically significant. The trend suggests that samples crusted after cooking had slightly higher average cooking loss compared to those without a crust, although this difference was not significant. The higher average cooking loss in the crusted samples can be attributed to the oil adhering to the surface of the samples after baking. Previous studies have also found that crusting before cooking did not reduce cooking losses [7]. On average, samples crusted after cooking released 2.4 times more moisture during the pressing test than the other samples, with measurements exceeding the maximum range being considered as 1 mm² g ^{− 1}. Previous research has suggested that cooking loss and WHC depend mostly on the temperature of the heat treatment (final sample temperature) for the same type of meat. Since crust creation was carried out at a higher temperature than cooking, we confirmed this relationship based on the difference between cooking loss and the WHC values of the “crust before cooking” and “without crust” sample groups [34]. In the case of the samples for which the crust was created after cooking, some of the oil used in the crust was absorbed by the samples and no longer released, so the previous correlation cannot be confirmed by the values of these samples.

In general, the higher the cooking loss of a meat, the higher the WHC of the heat-treated meat is. This can be explained by the fact that during cooking, the unbound and weakly bound water content is removed, and the strongly bound water content is better retained during pressing test which measure the WHC. This could also be seen in the case of samples where crusting was carried out before cooking or when samples were cooked without crusting. However, because the oil absorbed by the meat from the pan was also pressed out of the meat during pressing, the WHC also measured the fattiness—the “oil release” of the meat. Thus, the latter relationship was not observed for samples crusted after cooking.

3.4. Sensory Attributes

After verifying the equality of error variances [Smell preference: F(5,354) = 3.941, p = 0.002; Colour preference: F(5,354) = 1.215, p = 0.302; Texture preference: F(5,354) = 1.771, p = 0.118; Fattiness preference: F(5,354) = 4.606, p = 0.001; Taste preference: F(5,354) = 6.912, p = 0.001; Overall characteristics: F(5,354) = 2.212, p = 0.053] and the normality of residuals [Smell preference: W(360) = 0.980, p < 0.001; Colour preference: W(360) = 0.988, p = 0.004; Texture preference: W(360) = 0.983, p < 0.001; Fattiness preference: W(360) = 0.973, p < 0.001; Taste preference: W(360) = 0.977, p < 0.001; Overall characteristics: W(360) = 0.914, p < 0.001], a multivariate analysis of variance (MANOVA) was performed. It is important to note that for fattiness preference, taste preference, and overall characteristics, the assumptions for the MANOVA were not fully met. However, due to the robustness of the MANOVA under such conditions and in the interest of maintaining the completeness of the analysis, the test was conducted. In the case of untrained panellists and preference testing, this approach is considered acceptable.

The unexplained variance associated with meat type (and thus fat content) was found to be significant [Wilks’ lambda = 0.617, F(6,349) = 36.097, p < 0.001], as was the variance for treatment [Wilks’ lambda = 0.579, F(12,698) = 18.268, p < 0.001]. However, values above 0.5 were not particularly indicative of strong effects. The interaction between these two factors [Wilks’ lambda = 0.608, F(12,698) = 16.404, p < 0.001] yielded results similar to those of the individual factors. This suggests that neither the fat content of the meat nor the type of treatment significantly influenced the panel’s preferences for the samples. This can be explained by the fact that some consumers preferred drier meat, while others favoured fattier meat. Similarly, some consumers preferred darker-coloured meat, while others preferred lighter-coloured meat. The results of sensory analysis can be seen in Figure 12, Figure 13, Figure 14, Figure 15, Figure 16 and Figure 17 and

Table 4. Sensory results of sample groups made from different meat parts with different fat contents with different treatments.

Sensory Attribute	Meat Part	Treatment	Mean	Std. Deviation
Smell preference [-]	tenderloin	without crust	5.57	2.37
		crust before cooking	4.15	2.60
		crust after cooking	5.80	2.67
	collar	without crust	3.13	1.65
		crust before cooking	4.90	2.86
		crust after cooking	5.92	2.45
Colour preference [-]	tenderloin	without crust	4.08	2.31
		crust before cooking	5.07	2.22
		crust after cooking	3.07	2.16
	collar	without crust	6.52	2.27
		crust before cooking	5.30	2.79
		crust after cooking	6.88	2.53
Texture preference [-]	tenderloin	without crust	3.97	2.20
		crust before cooking	7.35	2.25
		crust after cooking	4.08	2.30
	collar	without crust	6.82	2.08
		crust before cooking	8.57	2.40
		crust after cooking	6.30	2.71
Fattiness preference [-]	tenderloin	without crust	5.27	2.93
		crust before cooking	3.87	2.17
		crust after cooking	5.98	2.83
	collar	without crust	5.77	2.84
		crust before cooking	7.62	3.00
		crust after cooking	7.22	2.42
Taste preference [-]	tenderloin	without crust	6.67	2.38
		crust before cooking	7.07	2.65
		crust after cooking	7.22	2.78
	collar	without crust	5.98	3.27
		crust before cooking	4.13	2.12
		crust after cooking	6.27	2.79
Overall characteristics [-]	tenderloin	without crust	8.48	1.81
		crust before cooking	8.60	2.21
		crust after cooking	6.98	1.89
	collar	without crust	5.40	2.40
		crust before cooking	9.08	1.88
		crust after cooking	8.12	2.33

.

Unfortunately, it was not possible to isolate distinct consumer profiles using k-means cluster analysis, as consumer preferences were not consistent across different meats or treatments. Interestingly, although the smell of the meat theoretically should not have been influenced by the experimental factors, given that no spices were added and all the meat was allowed to cool to the same temperature before tasting, preconceptions about the different colours led most consumers to expect differences in the smell of the samples. According to the Tukey HSD post hoc test, the sample groups crusted after cooking were more favourably rated in terms of smell compared to the other sample groups (p = 0.05). These samples were darker and redder than the others. This suggests that, with a few exceptions, there were no significant differences in the sensory attributes of the various samples based on the preferences of untrained consumers. Consumer preferences can be very diverse. The sensory characteristics learned as a child will shape for a lifetime what a consumer expects from a known food. In addition, fashion also influences consumers. In the case of meat, a good example is the preference for thick medium rare meat patties and crispy fully cooked smashed burger meat patties. Previous research has highlighted that appearance also has a strong influence on consumer preference [7]. Some consumers prefer meat with a crust, while others dislike it, considering it burnt. In other words, regardless of how the meat was treated or the type of meat used, the samples were similarly enjoyed by the general consumer panel, and overall consumer opinions were comparable.

In addition, the texture, taste, and overall characteristics of the sample groups crusted after cooking received higher ratings compared to the other sample groups (p = 0.05). This can likely be attributed to the fact that the crust formed after cooking rendered the surface of the meat crispier, and the flavours developed during the crusting process were more pronounced on the surface of the meat compared to meat crusted prior to cooking. Therefore, despite the observed high standard deviations, it can be inferred that the general consumer panel in this study exhibited a preference for meats crusted after cooking. A typical cooking method in Hungarian cuisine involves briefly frying both sides of the meat in fat, followed by the addition of water, which is then allowed to evaporate, thereby resulting in the formation of a crust. As such, it is plausible that most participants in this study were more familiar with meat prepared using this technique. However, due to the substantial standard deviations, only minor differences in preferences were observed between the various meat treatments. Sensory results are presented in Figure 12, Figure 13, Figure 14, Figure 15, Figure 16 and Figure 17 and Table 4.

It is well-documented that consumers generally prefer a crispy or crunchy crust in a variety of foods [35,36,37,38]. While samples crusted after cooking were overall more preferred, this difference did not reach statistical significance in the present study. Interestingly, these findings do not corroborate those of previous studies. This discrepancy may be attributed to the differing tastes and preconceptions regarding pork in Central European cuisine, which contrast with those found in other regions. Furthermore, online recipes, particularly those outside of the scientific literature, often recommend forming a crust prior to cooking in order to retain moisture within the meat. Based on the results of this experiment, trend value findings suggest that more moisture is pressed out of the meat when crusted before cooking as opposed to after cooking. However, the crust formed post-cooking, along with the oil retained by the meat during the process, may impart additional beneficial properties to the pork.

It is essential to acknowledge that consumer preferences can vary across time and geographical location. A review of the existing literature revealed challenges in establishing direct comparisons, as previous studies often involved different types of meat (e.g., beef), varying treatments (e.g., dry-ageing), and diverse consumer preferences across distinct regions. Therefore, it is recommended that future studies focus on specific consumer segments. Previous research has indicated that the preferences of a significant proportion of consumers are predominantly influenced by juiciness, while smaller segments of the population are more influenced by texture and fragrance [39].

4. Conclusions

The texture of the meat samples was significantly influenced by both the type of meat and the applied treatment, with fat content (meat type) exerting a dominant effect. While cohesiveness and springiness did not exhibit significant variations, notable differences were observed in hardness and chewiness. Tenderloin samples, being denser and drier, were significantly harder and chewier than the more tender collar samples. The crusting treatment applied after cooking resulted in increased hardness and chewiness, attributed to the formation of an additional crust that occurred without softening during cooking. This treatment also significantly impacted the colour attributes of the samples. Specifically, samples crusted after cooking displayed substantial changes in colour, including increased darkness and redness, while samples crusted before cooking appeared more yellow. Treatment emerged as the most influential factor affecting cooking loss, pH, and water-holding capacity (WHC), alongside fat content.

Consumer preferences were not significantly affected by either meat type or treatment, suggesting a wide range of preferences among untrained panellists. Nevertheless, samples crusted after cooking were generally favoured for their enhanced texture, taste, and overall sensory characteristics. This preference is likely due to the development of a crispier surface and intensified flavours resulting from the crust formation process.

In conclusion, both the meat type and cooking treatment play significant roles in shaping the techno-functional and sensory attributes of meat. However, the treatment—particularly crusting after cooking—tended to enhance desirable qualities, such as texture. The findings of this study provide valuable insights for food producers, offering a better understanding of the effects of different crusting techniques on various meat types. These results have practical implications for the food industry and restaurants, guiding the formulation of ready-to-eat meat dishes and offering recommendations for meat product utilization.

Since the samples were not minced or homogenized meat, the non-uniform distribution of fat and connective tissue in the mammalian muscles may have influenced the results. This effect was minimized by the large sample size. One limitation of the study was the limited opportunity to build the sensory panel. It is worth investigating using a panel of specific consumer segments to examine preference and a professional panel to examine sensory characteristics. In addition, it is worthwhile to investigate the effect of current experimental factors on the heterocyclic aromatic amine content of meat in a similar experimental design. It is important to mention that the results are applicable to industrial and kitchen technologies, but as described, they are only valid for the examined meat cuts of the examined pig variety.