Take Your Mixes To The Next Level

Tuesday, August 25, 2015

Computer Or Bach?

Etude by Kulitta image
Maybe the robots are finally taking over. Yale computer scientist Donya Quick has developed a program that can create pieces of music that are so close to Johann Sebastian Bach's work that people have trouble telling the difference.

She calls the program "Kulitta" and it can learn the music quickly and create new pieces on demand in a matter of seconds.

In tests over 100 students judged the pieces to be the work of a human, and even some of the experts were occasionally fooled.

Kulitta analyzes the works of a composer then applies the rules it's learned from them when creating a new piece of music.

The one thing it can't do is determine a good piece for a bad one, but that's such a personal thing that even us humans can't seem to agree on.

Quick looks forward to having Kulitta do mashups between genres, with her ultimate goal a combination of Metallica and Mozart.

Could this be a new income source for artists where your musical algorithm is copyrighted and new compositions are created for the listener on the spot? Time to bow to our musical overlords.



3 comments:

Jordan said...

This seems like an amazing piece of technology! Maybe Kulitta will help us understand more about music and how it's created. If not, we'll at least get some beautiful pieces of music from it. Thanks for sharing!

duuuhhh said...

So who owns the rights to the songs Kulitta writes?

Rand said...

After the amazement wears off, let's not overlook the elephant in the room; what ATM's did for banks, this technology may do the same for musicians who studied for years to perfect their craft.

More sophisticated than Band-in-a-Box, this will allow practically anyone to produce music 'sampled' from their own favorite artists.

So yes, "who owns the rights to the songs Kulitta writes?"

Reminds me of EZMix compared to Mixers. Like any tool this technology can be abused or used wisely.

LinkWithin

Related Posts Plugin for WordPress, Blogger...