sjm wrote:jancivil wrote:People who frequent KVR or GS or VI may listen to tracks to detect what's real. This is not normal behavior and I've not seen it outside these fora.
There's an element of truth in that. Well OK, more than an element. But there are also different degrees of expectation from casual listeners. We're all so used to hearing things like workstation string ensemble patches filling out tracks, that we kind of thinks that's what strings sound like. But even the casual listener will be able to tell you that a classical solo piano piece sounds like robots playing if you hard quantise everything and leave everything at default velocity.
So I think there's a a shifting continuum there. Sometimes it doesn't matter if it sounds fake (80s workstation strings in a hip-hop song = legit), sometimes it does (My First Casio guitar solo climax in a metal epos = fail). And while the casual listener might not know why something sounds "weird" or "wrong", they can tell - at least when it's obvious.
The more acquainted someone is with an instrument, the easier they'll find it to spot a fake. I can normally tell a guitar is fake that convinces others, simply because I play the guitar every day. I'd probably have a hard time telling a halfway decent programmed xylophone from the real thing.
I'd definitely draw the line at cupped brass; I don't think there's any way you can convincingly program that.
Yeah, no, there is no way unless it's [its character type is] sampled. There are many things like this one. So a samples library will have gestures, or the samples are of a cupped trumpet et al and a legato instrument may be made out of it. For nearly everything I've used which uses 'gestures' or loops as the terminology of the kids today has it, I hate working this way and it's not usually feasible. WarpIV Music (brass and winds, jazz-type) seems extremely well thought-through as I've done nicely with theirs.