Virtual Playing Orchestra - official release

[Michael Willis]

That's fantastic news! Which opcodes are you using to handle expression?

[bigcat1969]

When I get Bill Gates to sponsor bci, I'm gonna hire you to do the sfz versions!

[Paul Battersby]

(03-02-2017, 11:15 PM)Michael Willis Wrote: That's fantastic news! Which opcodes are you using to handle expression?

The secret to handle expression is to use these opcodes:

xfin_locc11
xfin_hicc11

xfout_locc11
xfout_hicc11

---

(03-02-2017, 11:30 PM)bigcat1969 Wrote: When I get Bill Gates to sponsor bci, I'm gonna hire you to do the sfz versions!

Sfz is fine. Just don't ask me to loop samples. I don't think I could handle a job where I was paid to loop samples all day long! I'd need some serious hazard pay to do that again.

[peastman]

Quote:The can of worms is open, I'm deep in the rabbit hole now.

But what if the rabbits don't want worms in their hole???

This sounds like a great feature!  Really looking forward to 2.0.

[Michael Willis]

(03-02-2017, 11:32 PM)Paul Battersby Wrote: The secret to handle expression is to use these opcodes:

xfin_locc11
xfin_hicc11

xfout_locc11
xfout_hicc11

According to https://linuxsampler.org/sfz/, LinuxSampler supports all of those opcodes. I'll happily try it out when it is ready.

[Paul Battersby]

(03-03-2017, 12:51 AM)peastman Wrote:

Quote:The can of worms is open, I'm deep in the rabbit hole now.

But what if the rabbits don't want worms in their hole???

Too late. The worms are everywhere. It's total anarchy.

---

(03-03-2017, 02:06 AM)Michael Willis Wrote: According to https://linuxsampler.org/sfz/, LinuxSampler supports all of those opcodes. I'll happily try it out when it is ready.

I believe those opcodes are standard sfz 1.0, so hopefully they are widely supported. I'll let you know when I think the next version is ready for testing.

[Samulis]

Yeah, those are from SFZ 1.0.

The cool part is you can assign it to any CC- just change "11" to anything else (e.g. "1" might be a popular choice too).

You may find you get better or worse results from overlapping the crossfades partially versus completely.

How did you stop normal velocity from affecting the crossfade? Is that automatic? I've been approaching this challenge from a purely theoretical direction (I honestly haven't had time to just sit down and test with different opcode implementations)- any findings would be greatly appreciated- and possibly implemented into the automapper in the future.

[Paul Battersby]

(03-03-2017, 03:15 AM)Samulis Wrote: The cool part is you can assign it to any CC- just change "11" to anything else (e.g. "1" might be a popular choice too).

I considered CC1, but that's for a mod wheel by default, which I'm already using for another patch. CC11 is labelled as expression, which goes along with CC7 (volume) and I figure those are typically controlled by adjacent sliders because you'll want to move both around at the same time if adjusting volume and expression live.

(03-03-2017, 03:15 AM)Samulis Wrote: You may find you get better or worse results from overlapping the crossfades partially versus completely.

I found that to be true of normal crossfading of velocity layers (for brass at least). For expression (even for brass), it seems best to have a complete overlap because you can set the volume and overlap to anything you want. I can then mix in any amount of each sample I want and adjust it's volume, but I may experiment further with that. I may have phasing issues as I try this on other brass instruments but so far (after minimal testing) it seems to work fine for french horns, solo and section.

(03-03-2017, 03:15 AM)Samulis Wrote: How did you stop normal velocity from affecting the crossfade? Is that automatic?

It does seem to be automatic. Under CC cross fading control, it's as if my MIDI controller is no longer touch sensitive. The volume is no longer controlled by key strike velocity but only by a velocity fader. I don't see how it could work any other way.

EDIT: On closer inspection, there is still velocity sensitivity affecting volume. By default, it appears the CC7 slider (in sforzando) sets the maximum volume. That combined with the key stroke velocity determines the actual volume. I removed the velocity sensitive cross fade opcodes so both layers play at the same time, but are faded together based on the CC11 setting. That's why key stroke velocity no longer affects crossfade. Hope that makes sense.

[peastman]

I assume this is an optional feature? If we want an instrument to continue using velocity to set volume and to cross fade, we can?

[Paul Battersby]

(03-03-2017, 05:57 AM)peastman Wrote: I assume this is an optional feature?  If we want an instrument to continue using velocity to set volume and to cross fade, we can?

Yes. I will simply be providing some additional .sfz files. I have no reason to remove the original ones and I don't think everyone will want to be forced to mess with velocity and expression sliders. I want version 2.0 to be compatible with 1.x.