I've got a SoundFont with up/down/mute strums. It's mapped along the lines of C3=I downstrum, C4 = I upstrum, C5 = I mute, D3 = ii downstrum, etc. So it's relatively easy to transpose a song by changing the instrument from C Major Guitar to F Major Guitar. Of course, things break down when you need strums that aren't diatonic to the key, but it does a pretty good job.

Samples work pretty well for strummed guitar, bass and drum. That is, instruments which have an initial attack, but don't "shape" their sound. Some instruments work pretty well (flute, orchestral woodwinds) with sample-based versions, and I could see PG Music potentially building an "all in one" solution along those lines. But for something like a fiddle, sax or trumpet the articulation becomes a lot more important.

Some sound libraries are now going with phrases instead of individual samples. So if you're playing a run of eighth notes, it'll try to find a phrase that contains that particular type of run, and transpose it if necessary, splicing together various phrases into a final output. That way, not only do you get the individual note played with the correct articulation, but you also get a plausible connection between the notes. Think of it as similar to RealTracks, but at a much more microscopic phrase level - note to note.

It's a trade-off. RealTracks don't give the same level of control as a sample-based library, but a sample-based library quickly becomes huge, especially with the myriad of articulations that are available for some instruments (like the violin).


-- David Cuny
My virtual singer development blog

Vocal control, you say. Never heard of it. Is that some kind of ProTools thing?