Hugely excited by MIDI 2
I have always been dissappointed with the sound of classical instruments as VSTs, though I own a lot of them. Here are some of the reasons. Many acoustic instruments have a varied vibrato that goes on after a note is struck, one needs specialist devices to control this and I have found even with them, the sound is a bit lame. Secondly, the attack of a note on a acoustic instrument is a world in its own right. I play sax and every note is crafted intuitively in the reed, same with the embroucher of brass instruments and strings like violins where the bow creates the attack in different ways. In scoring for violin, there are many articulations called for and even those articulations that are written the same, can vary widely.
For example, staccato can be of different lengths and amplitude of attack, the wrong type of staccato can just sound silly.
Then there are the question of players. On trumpet a performance of Purcell's Trumpet voluntary calls for very different muscles that a Chet Backer or Miles Davis Solo. The Aural differences are massive.
I am wondering what people think the impact of MIDI 2 would have on these issues.
What are the dreams and abilities this new standard is capable of releasing?
We have profile and property exchange, but how do people see these actually being implemented say with Violins? I know we don't yet know, but what are the rumours and speculations? Could we see further devlopments in non sample based instruments?
Opinions? Guesses?
I'd not get too excited about this, if I were you. Unless you're happy to wait 10-20 years?
I'd suggest that most of what you're talking about could be done within Midi 1, it would just be VERY fiddly and maybe more trouble than it's worth.
All the midi sounds we have now are dependant on samples, usually modified using 'synth' type processes. There are cases now where multiple sample sounds, variants of the same instrument, have been made available. For example, a trumpet as you suggest. You could build a sound bank of as many different samples as you care to each representing a particular aspect of the sound, as you describe, even down to the same instrument but being played by different players using different styles. Then you could create a midi performance that uses a Program Change midi command tp select the required one, on a note by note basis. Fiddly? You'd need to do this pretty much manually, I'm not sure there's be any way that midi 'recording' could distinguish between the different sounds to do it automatically, as it can do with volume/velocity/frequency.
Maybe things will develop to allow more options with the 'synth' type processes, but I'd suggest that these are not THAT much to do with 'midi', either Spec 1 or the new Spec 2. Yes, the latter will allow many more conrollers, and might allow for more sensitive variation, but I don't see this doing much to help your complaints - the massive use of extra/alternate samples could help here.
Like I suggested, using alternate samples, layering, maybe lots of SYSEX to do 'synth' type extras, you could achieve a lot, but would it be worth the trouble?
I suppose the question I'm asking is 'how do you get this sort of detail INTO midi?' If you cannot get the detail IN, in a practical way, then is it worth thinking about how to get it OUT again?
Dream On.
Geoff.
For me Miroslav is the best orchestral VST out there. For ultimate realism though I think Kontakt is the only answer. Some of the string libraries are mind blowing!!