EMI Composing Computer Challenge

I was reading a thread in the composer's forum about a computer built by David Cope that can compose readily in the style of any composer. At some point in the discussion, someone mentioned that they felt the computer would end up composing music that was somehow without soul.

Well.... what we have here are two chorale style midi files. One written by the EMI computer and the other written by no other than J.S. Bach. I challenge you all to pick the one composed by EMI and maybe get some interesting discussion to boot.

Good luck.

Is #1 even on a chorale tune that Bach wrote to???? It probably is, but it just sounds too Bacchian to be Bach.

Seriously, I hate to say this, but even if they end up creating the same song, there is a HUGE difference between the spitting out of music by a computer, and the work of an acknowledged master writing music for the glory of God...
... so take that, mr. Cope.

I'm going to risk shooting myself in the foot here by hazarding a guess: bach-2.mid was written by Bach, with the other having been written by a computer.

It's probably worth mentioning at this point that I'm absolutely no Bach expert, by the way.

OK, now time to hide.

I said number one was bach as i recognised a fair few of his favoured progressions in there, more so than in 2.

I'm probably wrong though

Rats, I voted on #2 being Bach, but on second listen changed my mind....

I believe #1 is Bach. #2 repeates variations of the same melody three times. Very programmatic, but much easier to duplicate with a computer. It must be using some sort of algorithm. #1 seems more free flowing, more human. #2 also pauses in between each variation, which to me now screams ... ro8ot!

All I know is, #2 would be much easier to create from a programming perspective.
Hmmm. I guess it must be form A (pause) B (pause) A' (pause) B' (pause) A'' (pause) B'' (pause) A''' (pause) B''' etc, etc ad infinitum...... or just repeat, repeat, repeat, repeat, a number of times determined by integer X, where in this case, X=3. Each pause gives the opportunity to compute the next sequence.

Then again, I may be wrong, and that would probably be interpreted as an serious insult, but oh well ...

This is pretty interesting. Have any more challenges?

#2 is Bach's - i guess

That's what I thought. It seems obvious that #2 is Bach.

I'll let a few more gutsy people post here before I give the answer away.....

I have no connection with # 2 so I am saying Bach wrote # 1 while computer wrote # 2. Funny, but I find # 2 to have no soul and bland.

But again I am prob wrong... waiting impatiently for the answer...

For those wanting to kill David Cope, this guy NEVER had ANY intention on producing music similar to the style of Bach. This is part of his Experiment (that's why the whole project is called Experiments in Musical Intelligence). He tried to recreate the human way of analysing music (the way a completely untalented musician would learn how to compose, for example) and the human processes that are required for that in a computer. That's what computers have ALWAYS been about: trying to create an emulation of the brain with machinery, in an attempt to a) make things faster and easier, b) understand the brain itself (yes, some cognitive psychologists had based most of their models on computers, and they also used computer terminology such as "input" and "output" and "storage devices" etc).

So, this guy used previous composers as part of his experiment because he knew how their style sounded like, and he could judge whether the computer was doing a good job or not. So, after he finished creating all the algorithms and stuff, he just gave it lots of Bach and asked it to give him Bach back to him. And the computer analysed all the Bach scores it had as input, found and identified the "musical DNA" of the pieces (i.e. what elements of the music make us say "oh, this is bach, not corelli" or "this is bach, not scarlatti"), and then, it recombined these elements to create the pieces.

Cope did this because he was seeking a way to "cure" his composer's block. Back in 1981, when the project started (and mostly finished), he had a composer's block, yet he had to compose a full opera which he was commissioned for. So, slangly, he was in dip sheet, and he asked a friend of his at a party whether there is a computer that can help him write music, and his friend (who was into the computer industry) told him that nothing like that existed. So, David Cope decided to make one for himself. And so he did. In the beginning he just gave the PC rules (counterpoint and harmony rules from the Baroque, say), but the results were ugly. So, he decided that, instead of feeding it rules, he'd just give it scores, and "teach" the computer how to analyse scores and find the "musical DNA" of the composer. He has since then done it with Mozart, Wagner, Beethoven, Chopin, Joplin, Bach and another 100 composers. This is all part of the experiment.

What REALLY interested David Cope is giving the machine his OWN music, so the result would be an unheard piece in HIS OWN STYLE. And this is how he "'cooperated" with EMI to finish the rest of the opera, in just 10 days He has been working like that with EMI since then. And I find it totally logical and good. The computer is not composing for him, it's recombining his musical elements to produce a new piece, in his style, which -however- could have NOT been produced if the composer didn't have compositions to start with. So, this machine doesn't compose music for any amateur composer, just composers who already have established a style of their own, and composers who have finished compositions.

So, get off this "omg a computer is composing music, FIRE 2 DA WITCHES!!!" kind of thinking, it starts to get really irritating...




Anyway, I won't vote because I've heard the computer-generated piece before Performed by a choir. And trust me, it sounded amazing... I will give you the link to it once the answer has been given You HAVE to hear it