Re: [LAU] tool to seperate elements of a song

On Sun, 2008-02-24 at 17:20 +0100, Jean-Baptiste Mestelan wrote:
> On 23/02/2008, schoappied <schoappied@email-addr-hidden> wrote:
> To say if a brain can do it, a machine can do it, is quit a
> degradation
> of the very complex organ of human that can let us experience
> this
> amazing world... Maybe you're sitting to much time behind your
> pc... ;)
>
> Sorry for the late reply : I was having some time off the PC ....
>
> Acknowledging the marvelous complexity of the human brain should not
> prevent us from putting this wonderful resource to use, and think up
> new ways of doing new things ... , should it ?
> Incidentally, this is how mankind has thrived during the last few
> millennia : by taking hints from nature, and designing tools and
> machinery. In turn, this machinery (instrumentation, etc ) has helped
> increase the amount of knowledge available for our brains to consume.
>
> Anyway, I am probably just preaching to the converted, as it is you
> who - in the first place - asked whether such a computer program
> already existed ?
>

IMO the way to do this would be to write a program that leveraged
several of the LAD resources. You would need a fairly hefty setup to
achieve it in realtime.

It would need to learn how to extract the parts from the composition and
recreate them with the standard tools. I imagine some kind of genetic
based learning algorithm would help similar to the work done on the self
driving cars, the robot face or the Lord of the rings animations.

It would be a very interesting project and a great addition to a CV but
you might need a couple of years and some serious cash to achieve any
significant results...

Cheers.

-- 
Patrick Shirkey
Boost Hardware Ltd.
_______________________________________________
Linux-audio-user mailing list
Linux-audio-user@email-addr-hidden
http://lists.linuxaudio.org/mailman/listinfo/linux-audio-user