Log in

11 July 2013 @ 08:35 pm
research on musical emotion - results!  
Around two months ago I posted in this community a link to research on musical emotion, now I would like to thank everyone who participated and tell you about some results.
My primary aim was to research the musical features that are correlated with emotions, but this is too technical to discuss here, so I'll just discuss the "psychological" side.
It turns out, that men and women prefer different music based on its emotional content. The difference was quite small, but men prefer more solemn and calm music, while women prefered nostalgic and sad music.

Moreover, listeners seem to define an emotion differently depending on genre.
Listeners of classical music feel "amazement, wonder" mostly when they hear quick, joyful music, while listeners of electronic music feel amazed when they hear dark, awe and fear inspiring music.
For listeners of classical music nostalgia is a sorrowful and tragic feeling, and for listeners of rock term "nostalgic" applies more to quite, slow and calm pieces.

Here is a graph, which shows how music in different genres induces in its listeners different emotions.
Pop music induces mostly tenderness, nostalgia and calmness. Electronic music induces tension, joyful activation.
Classical music induces mostly calmness, solemnity or joyful activation.


Thanks to everyone who participated and if you still want to take part in research, its still going on and probably in the end I will be able to build a better computer algorithm for music classification.
A link to survey:
http:// www.emotify.org
James Camien: On a treefelephant on July 12th, 2013 01:44 pm (UTC)
Thanks for this! I'm still pretty concerned about the methodology, as I argued in the previous post, but this is interesting nonetheless. (And perhaps its being interesting isn't affected by the particular concerns I raised: maybe they can go through or not without affecting this.)
vieni: fire_flowervieni on July 12th, 2013 01:56 pm (UTC)
Yes, I remember our previous conversation about difficulty of verbalising emotions. When I started to look into data I discovered that there is also a problem of translation - probably people mostly verbalize emotions to themselves in their mother tongue first, and emotional terms are barely translatable, because there are all these nuances and cultural connotations.
But nevertheless this emotional model proved to be quite good! People reached sufficient (but not excellent) agreement on which emotion a piece induces.