Evaluation of the goodness of musical tempo

Abstract

Numerous studies on how the tempo of music influences the assessment of physiology and emotion have concluded that “the higher the tempo derived the higher assessment.” However, these were based on experiments that also contained factors other than the tempo. Because the evaluation indices in those papers were not unified and the experimental conditions were different, it was not possible to obtain a uniform knowledge on the relation between tempo and evaluation. In this article, we chose “goodness of tempo” as the evaluation index and examined the influence of tempo on evaluation by changing its speed. Three songs, with different tempi in their original version, were used as test pieces. The findings demonstrated that each music piece had different assessment functions. One showed a constant assessment function at any speed, the other had an assessment function with optimum speed, and the third generated a step function with a low rating at the low speed and a high rating at the high speed. As a result, the proposition “the higher the tempo, the higher the evaluation value” was proved incorrect. Furthermore, it was considered that factors other than tempo affect the goodness of tempo.

Keywords

musical tempo evaluation functional representation music dependency preference

Problem presentation

The relationship between tempo and psychological states has been studied for a long time, and many have concluded that the faster the tempo, the better the evaluation, referring to the improvement of the emotional state and perception. For this reason, the proposition “the higher the tempo, the higher the evaluation” may be accepted as truth. However, sometimes results were different. These studies differed in music mode, tempo, and song genre, as well as within/between-participant differences. The differences in these conditions might cause discrepancies between conclusions. Consequently, it was difficult to verify under what conditions the above statement held only based on the results of each study. Furthermore, because the evaluation indices differed, it was not possible to compare whether the conclusions were identical or not. By elaborately designing experimental conditions, our aim is to verify that the proposition “the faster the tempo, the higher the evaluation” is incorrect.

Background of the problem

Rigg (1940) presented stimuli that combined six tempi and five different music samples and asked the participants to rate the serious-sad and pleasant-happy items. The results were summarized as fast tempo tends to make music happy whereas slow tempo has the opposite effect. Dobrota and Ercegovac (2015) showed that a fast tempo and major key (eight music samples) was more preferred than a slow tempo and minor key (eight music samples). However, tempo was only divided into slow/fast. Moreover, because the experiment was conducted under the condition of combining tempo and mode (major/minor), the effect of tempo alone was not shown. Gagnon and Peretz (2003) unified the melody, presenting stimuli consisting of 10 notes with varying tonality (major, minor) and tempo (110, 165, 220 bpm), and asked participants to judge whether each musical sequence was expressing happiness or sadness on a 10-point scale. The combination of a major tonality and a fast tempo was evaluated as happiness, whereas the combination of a minor tonality and slow tempo was evaluated as sadness.

In Webster and Weir (2005), short music samples of five-measure sequences were examined for 12 combination conditions: 2 texture (harmonized/nonharmonized) × 2 modes (major/minor) × 3 tempi (72, 108, and 144 bpm). It was a within-participant experiment that used evaluations on a continuous scale of “happy-sad.” As the tempo increased, happy ratings increased reliably and mostly linearly for nonharmonized samples in major, but they were neither linear nor quadratic for harmonized samples in minor. However, these main effects were qualified by higher-order interactions. Sims (1987) used 10 classic piano pieces containing slow (32–64 bpm) and fast (124–180 bpm) music, and asked the children (six categories from preschool, kindergarten, first through fourth grade) to evaluate the music by selecting three facial expressions (smiling, straight, frowning). Consequently, there was a strong correlation between tempo and preference, and fast music was favored over slow music except for kindergarteners. Hunter et al. (2008) cited earlier studies, including examples other than those above, and summarized that the faster the tempo, the more positive the emotion. Their experiment was aimed to verify the existence of complex emotions. They concluded that fast tempi and major modes associate with higher happy ratings.

Based on the results so far, the conclusion that “the faster the tempo, the higher the evaluation” seems to be valid. However, different results have shown less of an effect of tempo: fast or slow tempo did not affect preference, or it was completely different from the results mentioned above. Swaminathan and Schellenberg (2015) summed up previous studies and concluded that conflicting combinations (fast tempo and minor key, slow tempo and major key) could create more complex emotions than coherent combinations. When happy and sad were given as evaluation items at the same time, the participant’s evaluation did not fluctuate between the items, but instead they felt happy and sad at the same time. In Bramley et al. (2016), tempo alone did not influence participants’ physiological or subjective arousal levels, or their opinions of the musical stimuli in terms of liking or familiarity. They concluded that there were some circumstances under which the effect of music tempo did not work. Holbrook and Anand (1990) presented 14 stimuli created by varing only the tempo from one jazz song and found that ratings were a convex function with respect to the change in tempo (logarithm). In the study of Iwanaga and Tsukamoto (1998), the tempo directed by the composer was preferred in the well-known music and the medium tempo in the unknown music. Furthermore, in Yarbrough’s (1987) experiments using fast and slow music by Mozart and Chopin, it was preferable to have a slow tempo for fast music and a fast tempo for slow music.

Geringer et al. (2016) showed that preference differed depending on the combination of song genre and tempo. The well-known rock excerpts were preferred at their original recorded tempo, the originally fast samples of other styles (jazz, western, classical, and Brazilian popular) were preferred at slower tempi, and originally slow tempo samples were preferred at faster tempi. In Confredo et al. (2021), the relationship with music proficiency was investigated. The music majors preferred the faster tempi, but the band director preferred the original one.

The existence of results against the proposition “the faster the tempo, the higher the evaluation” was revealed. It is necessary to clarify whether the existence of contradictory results is due to errors in the proposition itself or to differences in the condition of each experiment.

Purpose of this article

Many papers conclude that “the faster the tempo, the higher the rating.” However, several other studies led to different results. When an experiment is conducted with multiple factors as independent variables, confounding occurs, affecting the result; therefore, it is difficult to ascertain whether the evaluation increases along with the tempo. To verify the proposal, it is desirable that the independent variable is represented only by the tempo factor. If only one type of music is used, there is a possibility that only the characteristics particular to that music are observed. When the tempo is extremely fast or slow, the meaning of the music may change. We decided to use three types of music with a different tempo, and each was within the range of 15% of increase and decrease of the original tempo.

In the above-mentioned studies, emotion-related assessments included happiness, liking, and preference. Happiness can be influenced by melody and may not include tempo ratings. It is quite possible that the “tempo preference” refers to examining only the tempo, which is separate from the content of the music. Therefore, the goodness of tempo—which is a straightforward evaluation item—was used as an index.

The purpose of this study can be rephrased as proving that “increasing the tempo does not necessarily increase the evaluation” is correct. If even one of the multiple music pieces does not apply, it is proved that the proposition “the higher the tempo, the higher the evaluation” is false by the rule of reductio ad absurdum. Only three similar types of music were used in this experiment. If the proposition were proved by this music, there is no need to investigate further music.

Method

Experimental design

The tempo evaluation in Tamir-Ostrover and Eitan (2015) was affected by the increase/decrease from the initial tempo. Dobashi et al. (2021) showed that mean and peak power outputs during the second set after listening to fast tempo music were significantly higher compared with those after the slow and no-music conditions. To avoid the influence of this presentation order of stimuli, three randomized patterns were prepared.

Gabrielsson and Juslin (1996) clarified that the evaluation of the participants differs depending on the intention of the performer and the instrument used. In our experiment, these problems were avoided using a stimulus in which only the tempo of the same song was changed.

Because it is difficult to ensure quietness and maintain the concentration of the participants in the experiment, the number of participants in one experiment was limited to less than 15. We planned to equalize the number of participants in each group, but the number of applicants varied: the number of experiments was five, and the number of people in each pattern was not equal.

To limit the evaluation to the tempo, only the speed of the same music was changed, thus keeping the other characteristics (e.g., mode, meter, melody) constant. The experiment participants were asked to evaluate the “goodness of tempo” for each music piece. Three music pieces were selected for this experiment. Because our purpose was to find music that ultimately denies the proposition, in principle, it was not necessary to unify or change elements such as mode or genre when selecting the music. The optimal tempo might be determined by the content of the music, therefore, for this experiment, it would be advantageous to choose completely different tempi for validation. However, because the goal was to verify the proposition within a realistic tempo, we decided to select music with different original tempi. By selecting from slow to fast tempo music, the degree of tempo deviation was limited for each piece; however, in the end, we could investigate the relationship between tempo and “goodness of tempo” within a wide range of tempi.

Experimental environment

We used a classroom with good audio equipment that can reproduce natural sound from the speakers. The air conditioner was turned off during the experiment to avoid noise. Three classrooms were used for experiments. The speakers were Bose Panaray 402 IV and Bose Model 101, and the amplifiers were Bose PS604 or equivalent quality products. The stimuli were played on a Toshiba laptop computer using foobar2000 at a 1,411 kbps. The level of intensity of the musical stimuli was set to about 50 dB at the participant’s position.

Participants

The participants in the experiment were 37 university students. After deleting the data containing inappropriate responses, the number of participants was 26 men and 9 women (35 people in total), with an average age of 22.0 years (SD = 1.2 years). The sample size for multiple comparisons among the seven groups by one-factor analysis of variance (ANOVA) was 32 at a significance level of .05, a population effect size of f = 0.25, and a power of 0.8. The actual size was appropriate.

Stimuli

It is known that musical genre affects tempo ratings. However, the purpose of this article is to demonstrate that tempo preference ratings can differ even for similar songs.

Three types of music were selected: slow, normal, and fast. Those were Chopin’s “Etude Op. 10, No. 3, in E major (Tristesse),” Liszt’s arrangement of “La Campanella,” and the third movement (“Presto”) from Vivaldi’s Four Seasons (Summer). As the tempo fluctuated during the pieces, the two authors measured the tempo of the stimulus portion of each piece of music. On average, those values were around 55 bpm for “Tristesse,” 90 bpm for “La Campanella,” and 150 bpm for Four Seasons. We created musical stimuli at seven tempi, including the speed of the original music. About 20 to 25 s from the beginning of the music pieces were cut out and used. The music editing software Audacity was used to cut out the music and edit the tempo. The tempo of the original piece was set at 1.0, and stimuli with tempo ratios ranging from 0.85 to 1.15 were created at 0.05 intervals without changing the pitch. The last 2 s of the stimuli were faded out. This operation was performed for the three music pieces, and a total of 21 types of stimuli were created.

Procedure

The participants consisted of five groups, and the experiment was conducted in groups in a classroom with 3 to 13 participants per session. Three stimuli with 1.0 times the tempo (original tempo) were given prior to the beginning of the experiment, as practice. In both practice and experimental phases, the order of presenting the music was fixed: “Tristesse,” “La Campanella,” and the Four Seasons. For the first task of the experimental phase, one stimulus randomly selected from each music piece was presented in the set order, followed by three randomly selected music pieces, in order, from the non-selected stimuli. The stimuli used in the practice phase differed from the first stimulus of this first task. Three patterns were created through the same procedure, and one of them was assigned to each group. Immediately after listening to the stimulus, the “goodness of tempo” was evaluated from a maximum of 20 points to a minimum of 1 point. If a participant gave an extremely high or low score first, it would not be possible to score a later stimulus. Therefore, original tempo music was assigned as practice, and the participants were asked to rate it around 10 points, without using extreme values. After completing the practice task, we confirmed with the experimental participants that there were no problems with the volume, response method, and so on. The participants were reminded that the stimuli were to be played back continuously and that they were to write their rating of the stimuli in the appropriate section of the answer sheet, and began the main task.

Although a questionnaire was administered after the experiment, the result was not used in this analysis and therefore was omitted. The number of participants in each experimental group used for the aggregation was 11, 5, 13, 3, and 3 persons, in the order of the experimental period. The first and second groups, as well as the fourth and fifth groups had the same stimulation order for each, being assigned one of the three patterns.

Results

The results of the experiments are presented in Table 1, which shows the mean ratings and SDs for the seven tempi of the three songs.

Table 1.
Result of Tempo Ratings for Three Pieces.

85% 90% 95% 100% 105% 110% 115%

M SD M SD M SD M SD M SD M SD M SD

A 10.31 3.21 9.11 3.72 9.69 3.70 9.89 3.99 10.03 3.64 9.86 3.76 10.46 3.28

B 7.46 3.24 8.66 3.27 10.46 3.02 11.34 3.22 12.57 3.22 12.09 3.52 11.60 4.45

C 7.71 3.53 9.46 3.11 8.26 3.31 11.91 3.01 11.09 2.99 12.03 2.86 11.97 3.28

Note. A = Tristesse, B = La Campanella, C = Four Seasons, percentage means the tempo ratio to the original tempo.

Tristesse

An ANOVA was performed using the statistical package R with the evaluation value of “Tristesse” as the dependent variable and the tempo ratio as the independent variable. There was no significant difference between the evaluation value and the tempo, F(6, 204) = 1.53, n.s., $η_{p}^{2}$ = .043. Because it was not easy to recognize the evaluation tendency from the results of the ANOVA, a linear first-order regression equation was obtained and shown in Figure 1.

Figure 1.
The Relationship Between Tempo and the Evaluated Values of “Goodness of Tempo” for “Tristesse.”

The regression equation is y = 1.61x + 8.29 (residual sum of squares [RSS] = 0.98), where y is the evaluation value and x is the tempo ratio to the original music. Although it seems that there is a slight slope, it can be regarded as an almost horizontal straight line because there are no significant differences between the tempo ratios from the results of analysis of variance. That is, in the range of the measured tempo ratio, the evaluation value does not vary depending on the tempo ratio.

La Campanella

The analysis of variance using the evaluation value for “La Campanella” as the dependent variable showed the main effect of the tempo ratio, F(6, 204) = 15.43, p < .01, $η_{p}^{2}$ = .31. In multiple comparisons done by Bonferroni with a level of 5%, the ratios 0.85 and 0.95 were considered significant difference relative to the other ratios, except for those between ratios 0.85 and 0.90 or 0.90 and 0.95. The ratio 0.95 had significance between 1.05, but no significant differences between the other ratios. To facilitate this trend, a regression equation was added to Figure 2.

Figure 2.
The Relationship Between Tempo and the Evaluated Values of “Goodness of Tempo” for “La Campanella.”

The residual sum of squares for the quadratic regression equation was the smallest when compared with those of the linear and cubic equations. The regression equation is y = −91.29x² + 197.87x − 95.07 (RSS = 0.59), and the maximum evaluation value was 1.06 times of the original tempo. The tempo ratios from 1.0 to 1.15 had no significant difference for each combination, whereas significance was observed for the ratio of 0.95 or lower; therefore, it can be said that the higher the tempo ratio, the higher the evaluation.

Four Seasons

For the Four Seasons, the main effect of the tempo ratio was admitted, F(6, 204) = 17.86, p < .01, $η_{p}^{2}$ = .34. The results (Bonferroni corrected for multiple comparison at the 5% level) showed no difference for any combination of 0.85, 0.90, and 0.95. There was a significant difference between the tempo ratio of 0.90 and 1.00 or higher, except for 1.05, and there was a difference between the ratio of 0.95 and higher, but there was none for 1.00 and higher. When the residual sum of squares for the linear, quadratic, and step regression equations was calculated, the step function had the smallest value (RSS = 2.19). In the step function, the average tempo above the original music was 11.75, and the average slow tempo was 8.48. From these results, the result was a step-like evaluation with a high evaluation at a tempo equal to or higher than the original tempo and a low evaluation at a slower speed than the tempo of the original music. These results were shown in Figure 3.

Figure 3.
The Relationship Between Tempo and the Evaluated Values of “Goodness of Tempo” for Four Seasons.

Verification of participants group differences

We performed a two-factor mixed-design analysis of variance for each song: tempo and presentation pattern. No interaction was seen in any of the songs, indicating that the order of stimuli had no effect. The values of F(12, 192) in the order of “Tristesse,” “La Campanella,” and Four Seasons were 1.19, 0.57, and 0.00, and all p values were greater than .05. There was no effect on the evaluation of the stimulus presentation order, and it was shown that there was no bias in individual characteristics.

Discussion

In the past, there was no research that used “goodness of tempo” as an evaluation index. The evaluation function that estimated the relationship between the tempo ratios and the evaluation values differed for the three types of music used. It was proved that the proposition “the faster the tempo, the higher the goodness of tempo” is false. By comparing the differences from the conventional research results, while considering differences such as experimental conditions, evaluation indexes, we reveal the problems of the existing research, and the validity of this research will be verified.

Comparison with experiments with similar evaluation items

Rashotte and Wedell (2012) is one of the few studies to treat tempo itself as a rating item. Their evaluation item “tempo pleasantness” is very close to the “goodness of tempo.” The first experiment used a 95-bpm song by the Beatles, and 12 clips (five clips of the same song with only the tempo changed and seven clips for context) were created. In the context clips, the tempi were −30% to −9% for the slow group and 9% to 30% for the fast group, and the target tempi were −12% to 12%. After rating the tempo speed of each stimulus on a 9-point scale, the “comfort” of the tempo was rated on a 9-point scale. The tempi of the targets were evaluated faster in the slow context than in the fast one. The peak pleasantness value shifted to the context tempi. The second experiment used a fast (156 bpm) Beatles song. In the first experiment, the evaluation value of tempo had a contrast effect given by the context; whereas, in the second experiment, the contrast effect was less than in the first experiment. In the third experiment, two songs and two tempo conditions (slow/fast) were used, and four target songs (the same song, a song with similar tempo, and two songs with dissimilar tempi) were set as the context. The tempo pleasantness had a contrast effect only for the songs with the same tempo, and none for the songs with similar or dissimilar tempi. In other words, it was suggested that the type of the song affected the evaluations associated with changes in tempo other than the same song.

Rashotte mentioned that they did not dare to use “preferred tempo,” but future research would be needed to assess the difference from “tempo pleasantness.” Regarding the evaluation of “tempo pleasantness,” it seems that the evaluation of only the pleasantness of the tempo was required, ignoring the content of the music. However, a matter of concern that the tempo evaluation of the stimulus might be influenced by the context was dispelled by the fact that the tempo speed rating increased with the tempo. In addition, the evaluation of “tempo pleasantness” also differed depending on the context, and the maximum point of each evaluation value was different. It shows that this is not simply an assessment of the tempo itself. It must have been interpreted in the same manner as our “goodness of tempo.”

From the experiment in which the tempo of the same music was changed, “tempo pleasantness” became an inverted U-shaped evaluation function. The maximum value shifted to the faster tempo in the context of the fast tempo, rather than in the slow tempo. The inverted U-shape of the evaluation value is consistent with the result of one of our music pieces. Holbrook and Anand (1990) also suggested that the evaluation value would be expected to show an inverted U-shaped characteristic of Berlyne (1971). However, there was no direct explanation for the relationship with tempo. However, Iwanaga and Tsukamoto (1998) analyzed the relationship between the perceived tempo of a music piece and the tempo felt as emotion. Although it depended on the familiarity of the music, they found that the relationship between tempo and emotion was an inverted U shape. From these, it can be inferred that the music with a faster tempo is not always preferred.

Comparison with experiments with similar experimental conditions

A. Kawakami et al. (2013) evaluated a listener’s perceived emotion by asking, “How would normal people feel when listening to this musical stimulus?,” and they evaluated felt emotion using a traditional question, “How did you feel when listening to this musical stimulus?” The result showed that the emotions perceived through music did not always match the felt emotions.

Huron (2011) also argued that sad music can evoke positive emotions. These results meant that even a music that evokes a sense of happiness was not necessarily rated high. Because tempo is an important component of a piece of music, it may not be directly linked with the perceptual evaluation of the “goodness of tempo,” as in the above discussion. Although the possibility cannot be denied, there have been many results that lead to a positive evaluation if the tempo is faster than the original tempo. Therefore, this study considers that a positive evaluation is likely to lead to a good tempo, and compare our result with Rigg’s (1940) result.

In Rigg’s experiment, three happy and two sad music pieces were used as stimuli. The tempo was changed in seven steps for each music. The participants rated their emotional state and judged whether music were happy or sad. For each piece, the selection rate for each tempo of the experiment participants was tabulated. From the tendency of selectivity, they concluded that the faster the tempo, the more happiness was chosen. However, no statistical tests were conducted. We performed a χ² test for each music piece based on the selection tables (Rigg, 1940: Tables I–V). The selection rate of happiness tended to be significantly higher at tempos of 120 bpm or higher, and the selection rate of happiness tended to be significantly lower at low tempi of 80 bpm or lower. Overall, the faster the tempo, the more monotonously the number of people selected; however, in some cases, the happiness selection rate was saturated before the fastest tempo, and the happiness selection rate was the highest at the fastest tempo. The fastest tempo of one music piece was set to the limit that the pianist could perform. In this case, it can be said that the faster the tempo, the higher the selection rate of happiness. However, considering the results of other studies, it cannot be true that happiness is always selected as tempo increases.

In our experiment, we used a rating value of 1 to 20. The step evaluation function meant that the value of “goodness of tempo” was low in the slow tempo area, and high in the fast tempo area. The quadratic curve evaluation function with a convex upward was maximum at a tempo slightly faster than the original. Therefore, assuming that the choice of happiness is based on the “goodness of tempo,” the one piece is consistent with Rigg’s results, but the others do not match.

In Webster and Weir (2005), 12 variations of four different musical phrases in major and minor, created in Finale, were used. There were three tempos: 72, 108, and 144 bpm. The happiness rating increased as the tempo increased. Furthermore, in Husain et al. (2002), the subjective enjoyment rating of the combination of a major key and fast tempo (165 bpm) was significantly higher than that of the combination of a minor key and slow tempo (60 bpm). Predicting the result of “goodness of tempo” from these results, the slower tempo matched the content for sad music, and the faster tempo, the subjective feeling was inferred positive. However, in Yarbrough (1987), the fast tempo was Presto (quarter note = 176–192) for Mozart and Molto vivace (faster than quarter note = 156–176) for Chopin, whereas the slow tempo was Adagio for Mozart (quarter note = 56–63) and Andante (quarter note = 63–76) in Chopin. The slow tempo for the fast music and the fast tempo for the slow music were preferred. This result was inconsistent with Rigg’s result. The fact that different results were derived depending on the music suggests that the evaluation may change depending on the tempo of the original music and the content of the music.

Comparison with experimental results of physiological/emotional levels

Tempo has also been shown to affect physiological levels. Carpentier and Potter (2007) reported that fast classical music increased skin conductance response (SCR) activation and rock music decreased it. Van der Zwaag et al. (2011) investigated the relationship between tempo, SCR, and heart rate variability (HRV). The SCR had a positive correlation with arousal and tension, whereas the HRV had a negative correlation with them. The result generated by Rock music did not coincide with Carpentier’s result. Among the rock pieces, there was a tempo of 82 bpm, which was classified as a slow tempo. It is possible that the result was affected by the fact that it was slightly faster than the slow tempo of other experiments. The SCR was higher in high-percussiveness than in the low-percussiveness condition. Interaction with tempo was significant. As for mode, the arousal level was lower in major than in minor, and this result was different from previous studies. The slightly faster tempo compared with the study by Carpentier and Potter (2007) is probably the main cause of the discrepancy. Differences in music content indicated different emotional effects, as the evaluations were strongly influenced by mode and percussiveness. Bauer et al. (2015) used rock songs to investigate the relationship between tempo and preference in the 90 to 180 bpm range. Electroencephalogram (EEG) results revealed a significant positive correlation between motor beta and preferred tempo. In Brownlow (2017), Galvanic skin response was significantly higher in the fast tempo music. These results showed that the faster the tempo, the more physiologically and emotionally activated it was. However, using various types of music, Liu et al. (2018) investigated the difference in the effect of low, medium, and high tempo by functional magnetic resonance imaging (fMRI). The evaluation of medium speed was the highest.

Husain et al. (2002) asked participants to listen to music that combined tempo (fast 165 bpm/slow 65 bpm) and mode (major/minor), and to report their mood on a questionnaire. The data were analyzed by principal components analysis. The result showed that the arousal level increased at the fast tempo. The relationship between arousal and emotion is not always clear. According to multidimensional scaling representation by Russell (1980), emotions such as happy and delighted are situated in the pleasure and arousal position. High arousal at a fast tempo is consistent with high happiness or preference at a fast tempo in this theory. In H. Kawakami et al. (2006), 20 people who received music training were participants in the experiment. The activity of the stimulus that increased the tempo was highly evaluated. In Karageorghis and Jones (2014), participants exercised at six intensities during which they were exposed to music tracks from four tempi and a no-music control. The main effects revealed that medium tempo preference scores were significantly higher than very fast tempo music, as were scores for fast tempo than very fast tempo music.

These results showed that the tempo had an effect even at the physiological level, but the effect did not necessarily indicate that the faster the tempo, the higher the activity physiologically. Physiological level responses were not always linked to cognitive level assessments. However, it cannot be denied that the fact that the reaction at the physiological level differs depending on the song is also reflected in the influence on the tempo at the cognitive level.

Regarding differences in music and performances

The music used in this experiment was classical; however, the relationship between tempo and evaluation was different for each music piece. The evaluation of the experiment was “goodness of tempo,” and it was presumed that the evaluation values would be different for each piece because whether or not the tempo matched the content of the music was a factor that affects the evaluation. Holbrook and Anand (1990) used a jazz song to show that there was a correlation between fast tempo (logarithm) and perceived activity, but emotional response was the highest at medium tempo (108 bpm).

In our experiment, “La Campanella” showed an inverted U-shaped characteristic. However, in Holbrook’s experiment, there were as many as 14 types of tempi, while only one music piece was used. It could have different characteristics being dependent on the content of the music used. Furthermore, because it was an emotional evaluation, it might have different characteristics in terms of “goodness of tempo.”

In Gagnon (2003), the melody of the stimuli was the same, and a 10-point emotional evaluation (from 1 = happiness to 10 = sadness) was conducted using stimuli consisting of 10 notes with varying mode (major/minor) and tempi (110, 165, 220 bpm). The combination of a major key and fast tempo was evaluated as happiness, whereas the combination of a minor key and slow tempo was evaluated as sadness. For the combination of major and slow tempo and minor and fast tempo, the evaluation results were based on the tempo information. They concluded that tempo had a greater effect than mode. Dobrota and Ercegovac (2015) also compared eight music pieces in a major key with a fast tempo and eight pieces in a minor key with a slow tempo. The faster tempo in the major key music showed a higher preference than the slower tempo in the minor key music. Moreover, in the analysis related to the personality traits, they were the highest factors in predicting preference. The former was consistent with the results of Gagnon and Peretz (2003), and the latter suggested that individuality contributed to tempo preferences.

Conclusion

We have experimentally shown that the proposition “the faster the tempo, the higher the rating” is wrong. Furthermore, the validity of the experimental results of this article was clarified by comparing them with studies that used similar rating indices and experimental conditions, studies that examined relationships with physiology and emotion, and studies that examined the effects of different types of music. Considering these results comprehensively, it was found that the tempo itself might directly affect the evaluation, which might change depending on physiology and emotions. In other words, the evaluation changed not just because of the attributes of the music but also due to the emotional characteristics of the listener. Because many factors were involved, it was shown that even slight differences in experimental conditions produced different consequences.

In each experiment, the music used was different; therefore, it is highly possible that the evaluation changed depending on the music. Characteristics such as original tempo, mode, and genre could be considered as the classification for the differences between music pieces; however, conventional research had concluded that a fast tempo was effective for evaluation, without fully considering the differences. Alternatively, the effects of multiple factors were not so sufficiently separated that the effects of pure tempo could not be obtained. The results of our experiment, in which the experimental conditions were controlled and these problems were excluded, showed that it was feasible to investigate the effect of tempo alone, and it became clear that a faster tempo does not always generate a higher evaluation. However, it does not ensure that the proposal will not always be valid when multiple factors are present. To obtain characteristics under compound factors, it is necessary to conduct experiments under conditions limited to the factors of interest.

	85%	90%	95%	100%	105%	110%	115%
A	10.31	3.21	9.11	3.72	9.69	3.70	9.89	3.99	10.03	3.64	9.86	3.76	10.46	3.28
B	7.46	3.24	8.66	3.27	10.46	3.02	11.34	3.22	12.57	3.22	12.09	3.52	11.60	4.45
C	7.71	3.53	9.46	3.11	8.26	3.31	11.91	3.01	11.09	2.99	12.03	2.86	11.97	3.28

Footnotes

Acknowledgements

The authors are grateful for the research support of Ritsumeikan University, where the experiment was conducted.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This study received the research support of Ritsumeikan University.

ORCID iD

Masaomi Oda

References

Bauer

A. K. R.

Kreutz

Herrmann

C. S.

(2015). Individual musical tempo preference correlates with EEG beta rhythm. Psychophysiology, 52(4), 600–604.

Berlyne

D. E.

(1971). Aesthetics and psychobiology. Meredith Corporation.

Bramley

Dibben

Rowe

(2016). Investigating the influence of music tempo on arousal and behaviour in laboratory virtual roulette. Psychology of Music, 44(6), 1389–1403.

Brownlow

B. N.

(2017). The effect of music tempo on the psychophysiological measures of stress. Continuum Undergraduate Research Journal, 2017, 8–16.

Carpentier

F. R. D.

Potter

R. F.

(2007). Effects of music on physiological arousal: Explorations into tempo and genre. Media Psychology, 10(3), 339–363.

Confredo

D. A.

Geringer

J. M.

Parisi

(2021). Effects of experience on tempo preference for selected wind band masterworks. Journal of Research in Music Education, 69(2), 188–206.

Dobashi

Matsuura

Ando

(2021). Listening to fast-tempo music during a post-exercise passive rest period improved subsequent sprint cycling. Perceptual and Motor Skills, 128(4), 1747–1764.

Dobrota

Ercegovac

I. R.

(2015). The relationship between music preferences of different mode and tempo and personality traits—Implications for music pedagogy. Music Education Research, 17(2), 234–247.

Gabrielsson

Juslin

P. N.

(1996). Emotional expression in music performance: Between the performer’s intention and the listener’s experience. Psychology of Music, 24(1), 68–91.

10.

Gagnon

Peretz

(2003). Mode and tempo relative contributions to “happy-sad” judgement in equitone melodies. Cognition and Emotion, 17(1), 25–40.

11.

Geringer

J. M.

Brittin

R. V.

Confredo

D. A.

(2016). Tempo preferences for fast and slow music excerpts of diverse styles: Jazz, western classical, pop/rock, and Brazilian pop. In Boal-Palheiros

(Ed.), International perspectives on research in music education: Proceedings of the 26th international seminar of the ISME Commission on Research (pp. 118–127). International Music Education Research Centre Press.

12.

Holbrook

M. B.

Anand

(1990). Effects of tempo and situational arousal on the listener’s perceptual and affective responses to music. Psychology of Music, 18(2), 150–162.

13.

Hunter

P. G.

Schellenberg

E. G.

Schimmack

(2008). Mixed affective responses to music with conflicting cues. Cognition and Emotion, 22(2), 327–352.

14.

Huron

(2011). Why is sad music pleasurable? A possible role for prolactin. Musicae Scientiae, 15(2), 146–158.

15.

Husain

Thompson

W. F.

Shellenberg

E. G.

(2002). Effects of musical tempo and mode on arousal, mood, and spatial abilities. Music Perception, 20(2), 151–171.

16.

Iwanaga

Tsukamoto

(1998). Preference for musical tempo involving systematic variations of presented tempi for known and unknown musical experts. Perceptual and Motor Skills, 86(1), 31–41.

17.

Karageorghis

Jones

(2014). On the stability and relevance of the exercise heart rate–music-tempo preference relationship. Psychology of Sport and Exercise, 15, 299–310.

18.

Kawakami

Furukawa

Katahira

Kamiyama

Okanoya

(2013). Relations between musical structures and perceived and felt emotions. Music Perception, 30(4), 407–417.

19.

Kawakami

Shimano

Matsumoto

(2006). Musical impression and contribution of structural factors. In 9th International Conference on Music Perception and Cognition (pp. 1424–1430). http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.605.101&rep=rep1&type=pdf

20.

Liu

Wei

Yuan

Wang

Zhao

(2018). Effects of musical tempo on musicians’ and non-musicians’ emotional experience when listening to music. Frontiers in Psychology, 9, 1–11.

21.

Rashotte

M. A.

Wedell

D. H.

(2012). Context effects on tempo and pleasantness judgements for Beatles songs. Attention, Perception, & Psychophysics, 74, 575–599.

22.

Rigg

M. G.

(1940). Speed as a determiner of musical mood. Journal of Experimental Psychology, 27(5), 566–571.

23.

Russell

J. A.

(1980). A circumplex model of affect. Journal of Personality and Social Psychology, 39, 1161–1178.

24.

Sims

W. L.

(1987). Effect of tempo on music preference of preschool through fourth-grade children. In Madsen

C. K.

Prickett

C. A.

(Eds.), Applications of research in music behavior (pp. 15–25). University of Alabama Press.

25.

Swaminathan

Schellenberg

E. G.

(2015). Current emotion research in music psychology. Emotion Review, 7(2), 189–197.

26.

Tamir-Ostrover

Eitan

(2015). Higher is faster: Pitch register and tempo preferences. Music Perception, 33(2), 179–198.

27.

Van der Zwaag

M. D.

Westerink

J. H. D. M.

van den Broek

E. L

. (2011). Emotional and psychophysiological responses to tempo, mode, and percussiveness. Musicae Scientiae, 15(2), 250–269.

28.

Webster

G. D.

Weir

C. G.

(2005). Emotional responses to music: Interactive effects of mode, texture, and tempo. Motivation and Emotion, 29, 19–39.

29.

Yarbrough

(1987). The effects of musical excerpts on tempo discriminations and preferences of musicians and non-musicians. In Madsen

C. K.

Prickett

C. A.

(Eds.), Applications of research in music behavior (pp. 175–189). University of Alabama Press.