Sending velocity

Use this forum to report bugs or request new features for WARBL, including fingering chart suggestions.
User avatar
admin
Site Admin
Posts: 680
Joined: Tue Mar 26, 2019 5:17 pm
Contact:

Re: Sending velocity

Post by admin »

I’ll take a look at this tonight or tomorrow. Is it v 1.8 that you installed? 1.9 is still in development and has a lot of untested features including changes to velocity, so I wouldn’t recommend installing that one.

I’d be curious to know if you put a lower limit of, say, 64 on the velocity output if you still get notes that don’t sound. In other words, I wonder if some notes have 0 velocity, or if they’re not getting sent at all.

Thanks!
Andrew Mowry
info@warbl.xyz
User avatar
gonzoB
Posts: 84
Joined: Thu Jan 28, 2021 9:45 pm
Location: Sydney, Australia
Contact:

Re: Sending velocity

Post by gonzoB »

Ah, yes, it's 1.9. I can't downgrade and still keep my fingering changes. Hell, I'll just be cutting edge!

And I just checked the incoming noteons. It's occasionally sending noteons with zero velocity, so that's the problem.

Gonzo
---------------------------
Gonzo Projects Page: https://gonzos.net/projects
User avatar
admin
Site Admin
Posts: 680
Joined: Tue Mar 26, 2019 5:17 pm
Contact:

Re: Sending velocity

Post by admin »

Okay, thanks. I’ll look at the changes we made in 1.9 and see if we broke something with the velocity. In general I do find that I have to set the lower limit of the velocity mapping higher so notes don’t get lost at lower pressures.
Andrew Mowry
info@warbl.xyz
User avatar
gonzoB
Posts: 84
Joined: Thu Jan 28, 2021 9:45 pm
Location: Sydney, Australia
Contact:

Re: Sending velocity

Post by gonzoB »

Hi Andrew.

What I've done with my (limited, admittedly) velocity programming on my wind controllers is
a) make sure there is a "hysteresis" gap between the noteon and noteoff pressures (stops losing the noteon)
b) wait about 20ms after the first CC2 to let the pressure stabilise
c) then send the noteon with either the current CC2 as velocity, or the "projected" initial peak value of CC2 (if using a delay time less than 20ms)

By the way, I reverted to 1.8 firmware and detected the same issue, so it may have been "broken" already.

Have I said how much I'm enjoying this device? It's really fun to play.

Gonzo
---------------------------
Gonzo Projects Page: https://gonzos.net/projects
User avatar
admin
Site Admin
Posts: 680
Joined: Tue Mar 26, 2019 5:17 pm
Contact:

Re: Sending velocity

Post by admin »

Thanks, I’m glad you’re enjoying it!

I tested out the velocity this morning, and what’s happening appears to be that when the 10-bit pressure value is mapped to the 8-bit velocity value there is a loss of precision, so that very low pressure values are mapped to a velocity value of zero. So, there’s not an error in the calculation per se but I agree that it doesn’t make sense that a note should ever be sent with a velocity of zero. You can set the output slider to a low number greater than zero to eliminate this, but maybe I should put a lower limit on the velocity output, like 1?

We’re calculating velocity in basically the way that you described. We were using a 20 mS delay when tonguing notes, but we’ve reduced that to 16 in v. 1.9. If you’re playing legato, there’s no delay as it’s assumed the pressure is changing more gradually.

I don’t normally use velocity myself because I find that the (necessary) staccato delay adds a noticeable amount of latency with fast tunes, but I use mostly tin whistle, which like you mentioned i your email is mostly just “off” or “on” :)
Andrew Mowry
info@warbl.xyz
User avatar
admin
Site Admin
Posts: 680
Joined: Tue Mar 26, 2019 5:17 pm
Contact:

Re: Sending velocity

Post by admin »

Ah! There actually was a big bug in calculating the velocity. I had a variable overflow where high tongued velocity values were overflowing and becoming very low values. It's fixed in the 1.8 version in the "archive" directory on GitHub now. I think you'll find the velocity works much better. Thanks for noticing this!
Andrew Mowry
info@warbl.xyz
User avatar
gonzoB
Posts: 84
Joined: Thu Jan 28, 2021 9:45 pm
Location: Sydney, Australia
Contact:

Re: Sending velocity

Post by gonzoB »

Morning Andrew.

I loaded your updated 1.8, and with "send pressure as velocity", the velocity for tongued noted is now locked to the minimum as set on the output slider. That's the only value that appears. With legato notes, the velocity varies.

Gonzo
---------------------------
Gonzo Projects Page: https://gonzos.net/projects
User avatar
admin
Site Admin
Posts: 680
Joined: Tue Mar 26, 2019 5:17 pm
Contact:

Re: Sending velocity

Post by admin »

I’m not getting that behavior, but I wonder if you may have downloaded the files before I made a final change about an hour ago. Would you mind downloading the “functions” file one more time (sorry)? When I’m testing it the tongued velocities mirror the legato velocities reasonably well, and I can get tongued velocities in the entire output range.

Thanks!
Andrew Mowry
info@warbl.xyz
User avatar
admin
Site Admin
Posts: 680
Joined: Tue Mar 26, 2019 5:17 pm
Contact:

Re: Sending velocity

Post by admin »

I just uploaded yet another versions of the functions. In fixing the previous issue I had caused another one where it was sending pressure as velocity even when that switch was turned off. Hopefully that's it for today :)
Andrew Mowry
info@warbl.xyz
User avatar
gonzoB
Posts: 84
Joined: Thu Jan 28, 2021 9:45 pm
Location: Sydney, Australia
Contact:

Re: Sending velocity

Post by gonzoB »

Hi Andrew.

Ships in the night! I downloaded the (latest) firmware and functions and the result is the same:
with "send pressure as velocity" on, only the slider minimum velocity is sent on tongued notes, legato works fine
with "send" off, the velocity is fixed for both tongued and legato at 64. Is there another setting that could be affecting it?

No rush.

Gonzo
---------------------------
Gonzo Projects Page: https://gonzos.net/projects
Post Reply