Humanization In digital music and perfection in played music.

Has anyone ever thought about why we as composers of the digital age put so much work in making our virtual players sound as human as possible, with different dynamic range, changing tempo, even errors, performance noises etc, when on the other hand as performers we are encouraged to make the piece sound as perfect as possible, steady tempo, perfect dynamics, no unnecessary noises...??

Because when we are with human performers we are aiming for the 'perfect' recording, that we have heard through thousands of times, of myriads of different composers and recordings. Because humans are supposed to be un-perfect, and thus we aim 'higher'. On the other hand computers and digital means are perfect by default and this is strange, weird and plainly wrong for our listening. so we try our best to emitate the behaviour of humans, by destroying perfection.

Makes any sense?

The setup is wrong. If you're writing electronic music to EMULATE human-played music, you need to sound like what humans playing would sound. If you want electronic music on its own, it can be whatever else regardless of how humans play or not (since humans can't "play" what electronic music is capable of.)

If you on the other hand ask humans to play something, they're not trying to copy a machine and it's sort of the contrary. Any good performer has room for playing a piece differently in different performances, and still be "good." It's encouraged otherwise might as well make a recording and be done with it.

Because I don't get paid if it doesn't sound relatively real.

Because I don't get paid if it doesn't sound relatively real.

...and, I don't get paid if I sound like a robot.

;)

Because I don't get paid if it doesn't sound relatively real.

Yet, it sounds perfectly "real" if you just make electronic sounds without imitating humans at all. It won't sound human, but it will sound like real electronic sounds. The problem of "not sounding real" only arises when electronics are already trying to match something they are not, in which case, yes, you'll have to go all the way if you want to make the illusion perfect. But even in the case of electronic sounds carrying a strong resemblance to human-created sounds (which they will generally have if you are, for instance, using instrumental samples), "sounding perfectly real" is generally no musical criterion, but as the quote above indicates much more a financial one. (Since without money as the deciding factor, you could just use humans right away.)

The musical questions, therefore, are rather different ones than the one in this thread's title. And for musical reasons, I'd personally never even consider "humanizing" the things I produce digitally. It's made by a computer and I don't mind if it shows. Often I even want it to show.