Blossomer is a probabilistic melody maker by David Su. It takes as input a single voice of melody (represented by notes with pitch and rhythm attributes), performs single-order Markov chain analysis to calculate probabilities, then outputs an infinitely recurring melody generated using a simple sine wave.
Despite the program's primitivity you can still hear it churn out something "new" while retaining the original "flavor" of a melody. Additionally, the different methods of playback generation yield noticeably different results. Check out the to-do list below for plans on making Blossomer more powerful!
Future features (to-do list):
- More advanced methods of analysis, especially the ability to take into account higher levels of structure (e.g. groups of notes as phrases, 8-bar structures, overall musical form) as well as melodic variation and embellishment
- On that note, a better sense of rhythmic cognizance
- Capability to "mix" different melodies by combining their probability tables
- Addition of "noise"/chaos to control how far away to stray from the original melody
- Alternate methods of melodic generation such as those based purely on intervallic relationships or talea + color extraction
- A more robust audio interface (beyond a simple sine wave)
- Ability to easily re-blossom the probabilistically generated melodies
- Control of output size, rather than an infinitely recurring melody
1/21/13: += auto input generator, 2 new playback generation methods, playback visualizer
1/10/13: first build
about | index | faq |