Emotions of subject and object affect beauty differently for images and music

What role do the emotions of subject and object play in judging the beauty of images and music? Eighty-one participants rated perceived beauty, liking, perceived happiness, and perceived sadness of 24 songs, 12 art images, and 12 nature photographs. Stimulus presentation was brief (2 seconds) or prolonged (20 seconds). The stimuli were presented in two blocks, and participants took the Positive and Negative Affect Score (PANAS) mood questionnaire before and after each block. They viewed a mood induction video between blocks either to increase their happiness or sadness or to maintain their mood. Using linear mixed-effects models, we found that perceived object happiness predicts an increase in image and song beauty regardless of duration. The effect of perceived object sadness on beauty, however, is stronger for songs than images and stronger for prolonged than brief durations. Subject emotion affects brief song beauty minimally and prolonged song beauty substantially. Whereas past studies of beauty and emotion emphasized sad music, here we analyze both happiness and sadness, both subject and object emotion, and both images and music. We conclude that the interactions between emotion and beauty are different for images and music and are strongly moderated by duration.

Barplots showing counts of subject happiness and sadness ratings, split by mood induction group Figure 2. Barplots showing counts of participant happiness and sadness ratings on a 7-point Likert scale from "Not at all" to "Very," taken from the PANAS mood questionnaire (Watson, 1998).Plots include ratings for each stimulus (24 images and 24 songs) from all 69 included participants.
Barplots showing counts of image and song beauty, happiness, and sadness ratings Figure 3. Barplots showing counts of image and song beauty, happiness, and sadness ratings on a 7-point Likert scale from "Not at all" to "Very."Plots include ratings for each stimulus (24 images and 24 songs) from all 69 included participants.

Table 1 .
Linear mixed-effects model for prolonged (20 sec) songs (N = 35).The model explains 65% of the variance in the data.Bold values indicate statistical significance.

Table 2 .
Linear mixed-effects model for brief (2 sec) songs (N = 34).The model explains 61% of the variance in the data.Bold values indicate statistical significance.

Table 3 .
Linear mixed-effects model for prolonged (20 sec) images (N = 33).The model explains 53% of the variance in the data.Bold values indicate statistical significance.

Table 4 .
Linear mixed-effects model for brief (2 sec) images (N = 36).The model explains 67% of the variance in the data.Bold values indicate statistical significance.Linear mixed-

Table 5 .
Linear mixed-effects model for prolonged (20 sec) songs (N = 35).The model explains 61% of the variance in the data.Bold values indicate statistical significance.

Table 6 .
Linear mixed-effects model for brief (2 sec) songs (N = 34).The model explains 60% of the variance in the data.Bold values indicate statistical significance.

Table 10 .
Linear mixed-effects model for brief (2 sec) songs (N = 34).The model explains 65% of the variance in the data.Bold values indicate statistical significance.

Table 11 .
Linear mixed-effects model for prolonged (20 sec) images (N = 33).The model explains 54% of the variance in the data.Bold values indicate statistical significance.

Table 12 .
Linear mixed-effects model for brief (2 sec) images (N = 36).The model explains 67% of the variance in the data.Bold values indicate statistical significance.

Table 13 .
Linear mixed-effects model for prolonged (20 sec) songs (N = 35).The model explains 65% of the variance in the data.Bold values indicate statistical significance.

Table 14 .
Linear mixed-effects model for brief (2 sec) songs (N = 34).The model explains 61% of the variance in the data.Bold values indicate statistical significance.

Table 15 .
Linear mixed-effects model for prolonged (20 sec) images (N = 33).The model explains 54% of the variance in the data.Bold values indicate statistical significance.

Table 16 .
Linear mixed-effects model for brief (2 sec) images (N = 36).The model explains 67% of the variance in the data.Bold values indicate statistical significance.

Table 17 .
Linear mixed-effects model for prolonged (20 sec) songs (N = 35).The model explains 65% of the variance in the data.Bold values indicate statistical significance.

Table 18 .
Linear mixed-effects model for brief (2 sec) songs (N = 34).The model explains 62% of the variance in the data.Bold values indicate statistical significance.

Table 19 .
Linear mixed-effects model for prolonged (20 sec) images (N = 33).The model explains 54% of the variance in the data.Bold values indicate statistical significance.

Table 20 .
Linear mixed-effects model for brief (2 sec) images (N = 36).The model explains 68% of the variance in the data.Bold values indicate statistical significance.

Table 21 .
Linear mixed-effects model for prolonged (20 sec) songs (N = 35).The model explains 64% of the variance in the data.Bold values indicate statistical significance.

Table 22 .
Linear mixed-effects model for brief (2 sec) songs (N = 34).The model explains 64% of the variance in the data.Bold values indicate statistical significance.

Table 23 .
Linear mixed-effects model for prolonged (20 sec) images (N = 33).The model explains 61% of the variance in the data.Bold values indicate statistical significance.

Table 24 .
Linear mixed-effects model for brief (2 sec) images (N = 36).The model explains 61% of the variance in the data.Bold values indicate statistical significance.

Table 25 .
Linear mixed-effects model for prolonged (20 sec) songs (N = 20).The model explains 62% of the variance in the data.Bold values indicate statistical significance.

Table 26 .
Linear mixed-effects model for brief (2 sec) songs (N = 19).The model explains 56% of the variance in the data.Bold values indicate statistical significance.

Table 27 .
Linear mixed-effects model for prolonged (20 sec) images (N = 20).The model explains 59% of the variance in the data.Bold values indicate statistical significance.

Table 28 .
Linear mixed-effects model for brief (2 sec) images (N = 19).The model explains 59% of the variance in the data.Bold values indicate statistical significance.

Table 29 .
Linear mixed-effects model for prolonged (20 sec) songs (N = 15).The model explains 70% of the variance in the data.Bold values indicate statistical significance.

Table 30 .
Linear mixed-effects model for brief (2 sec) songs (N = 15).The model explains 66% of the variance in the data.Bold values indicate statistical significance.

Table 31 .
Linear mixed-effects model for prolonged (20 sec) images (N = 15).The model explains 57% of the variance in the data.Bold values indicate statistical significance.

Table 32 .
Linear mixed-effects model for brief (2 sec) images (N = 15).The model explains 73% of the variance in the data.Bold values indicate statistical significance.