Re: New Controller from Pianoteq / Ivory / Kawai
I agree with Ecaroh's comments - higher resolution will give a more human feel to the playing,
Well, as Ecaroh suggested, we should try to prove that by testing it, and it will have to be a carefully designed test - the player will not be allowed to know the MIDI resolution when he is playing, and nor will the listeners. Only when all the listening tests have been performed and the opinions assessed should the resolution for the recordings be revealed to us, to eliminate bias. Now, even if we can hear a difference, that still doesn't prove that the player is playing with more resolution than standard. To prove that, we'd need to go further - we'd need to record the player's performance using standard MIDI resolution, but then superimpose +/- 0.5 a step of random noise, to create a quasi high resolution MIDI signal. If the result of doing this produces an audible difference to the straight high-res signal, I will then accept that standard MIDI is not sufficient. Let me give an extreme example, to illustrate the point I'm trying to make. Consider a hypothetical instrument, that is indeed designed for high-res MIDI - it has 127 x 10 steps = 1270 steps. It is a strange instrument - each step produces a very unique sound, easily distinguishable from every other sound produced by every other step. When the player plays this instrument, he will not be able to hit any of these sounds consistently, however, he will learn the approximate strength required to cause any of the desired sounds to be triggered "sometimes" - not every time, but sometimes. Ok, now we'll process the MIDI, to reduce it's resolution down to 127 steps. The 10 intermediate steps between each of the coarse 127 steps have now been removed - the instrument can NEVER make those sounds. The player will OF COURSE notice that these intermediate steps are missing, and we have suffered a degradation in behaviour of the instrument. I.e - this doesn't mean that the player is playing with a physical resolution of 1270 steps - it simply means that the sounds that he could sometimes be triggering at random have been removed. I'm asserting that if the instrument is now altered, such that it automatically selects, at random, one of the 10 intermediate steps, the subjective behaviour of the instrument will be restored, and the player will be unaware that the keyboard he is playing is only recording his velocity at standard (127 step) resolution.
though granted, the current midi standard gives a reasonably good representation as skip as demonstrated.
Reasonably good? You're very hard to please. Remember, I increased the dynamic range of Pianoteq to maximum, and the difference in steps is still imperceptible! It's better than "reasonably good" - it's EXCELLENT.
Greg.
