7 - CONCLUSIONS

[7]
CONCLUSIONS

7.1 Semantic appraisal of film and music

From a first bulk of 22 pieces of music and 12 film excerpts, initially appraised by two groups of 107 and 102 members respectively, we selected the 6 soundtracks and 6 film scenes with lowest variance on Arousal and Pleasure scores. Other stimuli, whose ratings were significantly different between female and male subgroups, or between the musically trained and the musically untrained, have been excluded from further investigation. Therefore the musical tracks and film excerpts we used for our final experiment were the ones receiving the most consistent Pleasure and Arousal ratings in the first part of the research.

Each of the six movie scenes has then been paired with two different pieces of music, for a total of 12 film/music combinations.

In subsequent steps in the research, these combinations have always been presented to two different groups. Each participant, regardless of the group it was assigned to, viewed all six film excerpts and heard all six soundtracks but the audiovisual pairing was different between groups.

This subdivision allowed each piece of music, as well as each film scene, to be presented only once to the same participant, in order to prevent undesirable effects ascribable to multiple exposure to the same material.

Two groups, consisting of 18 members each (all different from the ones used in the arousal/pleasure appraisal) were asked to rate the congruency between the film scene and its underscore.

Fit ratings were significantly correlated with the distance of the musical and the visual part of the combination on the Pleasure axis (r=.645, p<.05). Correlation between fit and distance on the Arousal axis was instead not significant (r=-.009)This is probably ascribable the very nature of the arousal itself.

For instance a soundtrack perceived as happy/joyful is very likely to be rated as incongruent, when used to underscore a solemn and dignified scene. On the other hand, one soundtrack can be perceived both sad (low of pleasure) and interesting (high arousal level), and it could probably serve as an effective and congruent accompaniment for a calm (low arousal level) and tragic (low pleasure) scene.

Another good example of this can be the results for combination 5(A).In this particular combination film excerpt #5 (low Pleasure, high Arousal) was accompanied by music #17 (low Pleasure, low Arousal).The appropriateness of this pairing was rated as very high (fit=6.24).The opposite happened for combination 2(A), where music film excerpt #3 (high Pleasure, low Arousal) was accompanied by music #16 (low Pleasure, low Arousal).This pairing in fact was rated as rather incongruent (fit=2.53).We can see that scores on the Pleasure dimension are the most effective in differentiating congruent vs. contrasting audiovisual combinations.

In our final experiment two groups of 51 and 41 members respectively, again all different from the ones used in the preceding experiments, appraised the audiovisual combinations using the same semantic differential scales as in the first part of the research.

We compared Pleasure and Arousal scores from film-only condition with film + music A condition, and from film-only with film + music B.Scores from film + music A and film + music B conditions were also compared.

Music was found to play an important role by means of providing additional semantic and emotional connotation to the visual stimuli. Results highlighted two main effects:

1) Adding music did affect significantly the semantic ratings of the film/music combination.

2) A change of underscore also produced significant differences in the semantic ratings.

The first effect occurred in 92% of the cases (all but one) for the Pleasure dimension, and in 67% of cases for the Arousal dimension.

The second effect occurred in 83% of the cases for the Pleasure dimension, and in 50% of cases for the Arousal dimension. This somewhat different impact of the two semantic dimensions can possibly be ascribed to the different weight of the Arousal factor, as explained in the previous paragraph (7.2).

The final experiment also featured an additional item: participants were asked to choose from two given interpretations the one they considered the most suitable to explain what was happening in the scene.The soundtrack accompanying the scene proved itself to have a significant effect in our viewers/listeners understanding of what was going on in the film: in four cases out of six, a different piece of music polarized viewer's preferences on one specific interpretation. In one of the remaining two cases the same tendency was also there, even it did not reach significance. This polarization effect was also found to be significantly correlated (r=.658, p<.01) to congruency of the relationship between film and underscore (fit)As a matter of fact, participants' preferences polarized more when music/film combinations had a high fit level (i.e. music and film were perceived as congruent), rather than when they were perceived as incongruent. This result suggests that a coherent combination of auditory and visual stimuli is more effective in being interpreted in one specific way. Similarly, viewer/listener's interpretations are more spread between alternatives when they are appraising a scene accompanied by a contrasting soundtrack.

This difference can possibly explained if we consider that, when participants are presented a contrasting combination, the range of conceivable explanations is wider than when film and soundtrack are congruent. Therefore, since the scenes we used were all open-ended and did not lead to a clear interpretation by themselves, a soundtrack that is perceived as congruent to the visual elements will bring the majority of viewers/listeners to the one or very few interpretations. On the contrary, an underscoring perceived as contrasting will make the audience think of more different interpretations. As a consequence of this wider range of explanations we can expect a smaller agreement between subjects.

More aspects of the interaction between film and soundtrack could possibly be explored by further research.For instance, does a specific soundtrack highlight certain characteristics or elements in the film? Does the use of underscoring deepen the analysis of the film scene? How does this vary from when the music is congruent to when it is incongruent with the visuals? Participants could be asked to write a description of the scene they were presented, or fill in a questionnaire some time later, to appraise which details they remember better. These descriptions and memory test results could then be compared to verify whether different pieces of music, or different degrees of congruency between film and music, produced different results.

Moreover, do some specific soundtracks produce an effect on semantic differential rating regardless of the visuals they are paired with? Can we identify singular characteristic of music (tonality, tempo, timbre, etc.) possibly responsible for specific effects on the semantic appraisal of a film? A set of soundtracks, whose musical qualities are known, could be used to underscore a large number of different film excerpts to verify if there is a tendency of such kind.

Music is almost always used to accompany film and video, and filmmakers and composers certainly know how to obtain certain psychological effects, like increasing attention or providing additional closure to an event.Unfortunately there is no specific psychological theory to explain these phenomena. In summary, we believe that the interaction between visuals and music can provide researchers with many aspects of psychological interest.