You guys all know I am an algorithm fanatic.
Some of you might even know I try to put everything I stumble across into an algorithm… or a process.
Lately, Raphael showed me a very interesting product of Microsoft Research.
It is supposed to be easy and simple:
- You choose a rythm which makes a drum base.
- You sing along, with a song you made up (or not)
- The music is automatically generated from your voice.
I imagine artists and mathematicians have been working together on that project.
Lots of users actually used the produt in another interesting way.
They take a classic hit, then remove the background music. Then generate the new music with songsmith.
Here’s the result with Eye of the Tiger.
Ok, quite good isn’t it? Well, the others are epic fails.
Anyway, this takes me to other considerations…
If the music is no longer an issue for a wannabee artist, then the voice will no longer be an issue neither.
I imagine softwares of the futures will be able to generate any voice you like and add those effects you would have to take years to master.
And sentences are difficult to make… why not use tags that would generate the sentence ?
In the future, you will be able to give a few tags, and a targeted audience, and it will generate thousands of possible songs. You could even add a virtual singer.
That’s a bit like what Animoto does with the pictures into a video clip.
We’re definitely heading somewhere with all those algorithms.
Maybe T-800 the new terminator robot, will be a singing robot.