TransPose algorithm writes the soundtrack to novels
The automatic analysis of sentiment in text is fast changing the way we interpret and interact with words. On Twitter, for example, researchers have begun to gauge the mood of entire nations by analysing the emotional content of the tweets people generate.
In the same way, other researchers have started to measure the emotional temperature of novels by counting the density of words associated with the eight basic emotions of anticipation, anger, joy, fear, disgust, sadness, surprise and trust.
The next step, obviously was to write an algorithm that measures that emotional temperature throughout full length novels, and generate a musical soundtrack to accompany the text.
Interesting research, or pointless? Time will tell, but energy and money is increasingly being used to judge moods and allow software or equipment to react accordingly. The true value of this research might not become apparent until sometime in the future, even if it is only used so your phone can sing you a soothing song after you receive a nasty email from your boss.
In the same way, other researchers have started to measure the emotional temperature of novels by counting the density of words associated with the eight basic emotions of anticipation, anger, joy, fear, disgust, sadness, surprise and trust.
The next step, obviously was to write an algorithm that measures that emotional temperature throughout full length novels, and generate a musical soundtrack to accompany the text.
Interesting research, or pointless? Time will tell, but energy and money is increasingly being used to judge moods and allow software or equipment to react accordingly. The true value of this research might not become apparent until sometime in the future, even if it is only used so your phone can sing you a soothing song after you receive a nasty email from your boss.
Part of it is that while there is a standard vocabulary for associating particular musical patterns with emotions, we are also always playing with these patterns to shift our associations. So I'm wondering if the algorithmic approach will be able to have this kind of exploration and growth element added and, if so, how?