Thanks for your comments. 😊
Some of what you say is true, but only up to a point. I am a very experienced chamber music pianist, and I can tell you that your ex. 26 does not sound at all the way a violinist would play it. String players work for years to be able to play this kind of melody without a noticeable gap in the sound. There is a specific noise as the bow changes, but the sound does not stop.
Solo strings are (along with voices) the most difficult thing to simulate, for a reason: there are so many things they are controlling, at the same time, to make the music come alive:
1) intonation 2) vibrato speed and depth 3) bow direction 4) harshness or softness of bow attack 5) changing the pressure of the bow during a phrase or even one note 6) changing the speed of the bow during a phrase or even one note 7) where they play (nearer to the fingerboard or the bridge) 8) on which string they play the note 9) where to make the shifts 10) portamento or not. And there are probably other things I can't remember right now!
My point here is not to denigrate VSL, not at all. As I have said elsewhere, VSL libraries are the cornerstone of my sample banks. It's just that there is a contradiction between real live musicians, who never play two notes the same way (and if they are good, the differences will be musically meaningul), and recording samples, where, for obvious reasons, they try to get the players to play each note as similar as possible to the others around it.
In my case, the thing which often causes me the most difficulty in mockups is a lyrical melody with sustained notes. If I just play it as it, the sustained notes are too smooth and even. I often really do go in and sculpt them, one at a time, with the appropriate midi controllers, but that is a LOT of work, and sometimes it's not enough. For example, not all wind instruments can fade away to nothing, so a controller 11 fade is not realistic. And if the dynamic samples don't have exactly the right duration and amount of fade out, they can't solve the problem.
Of course these are subtleties, and I agree that most users don't have the interest or the real world musical experience to bother about these things. But this is where the state of the art is at this point. And, as in the past, I am sure VSL will be bringing new and better tools for us in the future. 😊
@Beat Kaufmann said:
Hi Alan
About "Bow Changes" Even if we don't have such specific up- and down-bow articulations we can simulate a natural "bowing" close to the reality. These points help...
1. First of all. Really "draw in" the bow directions into the score how in reallity.
2. Try to find those articulations which support the sound and the feeling of up- and down- stroke.
3. Let correct gaps between notes for certain bow directions. Example: Twice a down stroke needs time between the two notes for moving back the bow for being ready for the second down stroke.
Here is a little example (No26.). For fast notes see >>> here at youtube, This helps as well.
Further, I believe that chosing "good" articulations can make those new articulations nearly unneccessary. This example (No21.) shall demonstrate my statment. I also believe that 90% of the users would have more a new problem than a help with such new up-/down-stroke articulations.
Beat
PS I agree that power users of VSL-Libraries shoud have a lot more articulations and of course the story around "dynamic vibratos" should be a next project - at least with solo instruments.
