It sounds like gain no longer affects MIDI velocity in Tidal 1.7.
The amp param affects velocity, but I'm trying to track down what the default value for amp is. By my estimate, without specifying amp in a pattern, it results in a velocity of 50 out of 127, which is only about 39% of full velocity. Does this sound accurate, and is this desirable?
I think the pre-1.7 behavior was ~80% velocity.
this might be the reason i am having to crank my VSTs into the red to get any volume out of my synth patches. no, not desirable to me. let me do some tests on where my velocities are coming in at before specifying an amp value and i'll report back.
Isn't amp linear and expecting values between 0 and 1 ? That's how I've been treating it, ah !
Ps: solely using it for MIDI.
That's right, but the MIDI velocity is really low unless you specify # amp 1
-- quiet, velocity ~40%
d1 $ s "midi" # note "c5"
-- use amp 1 to get velocity 100%
d1 $ s "midi" # note "c5" # amp 1
I guess I am just arguing that the default velocity is too low.
With previous Tidal versions (I haven't used it much yet with 1.7, I treated it like # amp 0.5 equals 64, and 1 = 127 etc.. IIRC some simple test confirmed that.
I might be totally wrong though hehe
That sounds correct. That is consistent with what I've observed. I would just expect the default amp value (when not specified) to equal 1, or 0.8, or something higher. Instead, I think it defaults to something like 0.4 right now.
here are my findings. even though 0 127 is a weird scale to force into linear, it actually gets pretty close
default amp value when not explicitly specified = 0.4 (ccv 50)
amp 1 = ccv 127 (100% of total velocity)
amp 0.9 = ccv 114 (89.8% of total velocity)
amp 0.8 = ccv 101 (79.5% of total velocity)
amp 0.7 = ccv 88 (69.3% of total velocity)
amp 0.6 = ccv 76 (59.8% of total velocity)
amp 0.5 = ccv 63 (49.6% of total velocity)
amp 0.4 = ccv 50 (39.4% of total velocity)
amp 0.3 = ccv 38 (29.9% of total velocity)
amp 0.2 = ccv 25 (19.7% of total velocity)
amp 0.1 = ccv 12 (9.4% of total velocity)
amp 0.0 = ccv 0 (0% of total velocity)
not sure if we should petition to get the default set at amp 1 but maybe amp 0.5 would be better than amp 0.4.
Awesome, thanks for sorting all that out @kit-christopher!
Personally I would propose a default amp value of 0.8 or 0.9. As I recall, a lot of DAW software defaults MIDI velocity to about 80%, and amp = 0.8 is roughly 80%.
personally i would vote for a higher value but, for example, i know that when i load a sample into Ableton's Sampler instrument, it defaults to 45% velocity and -12db gain. to me, it's totally irritating and i have templates set up to override that but there's gotta be a logical reason for why they set the defaults like that...i just can't figure it out.
Ideally, a dedicated velocity function with the usual 0-127 range would make most sense IMO. No matter some weird implemantations (re: Live's Simpler), it's still the most common way to measure / quantize velocity.
I agree on both a higher default and a proper velocity function with the expected 0-127 range... This seemed like an odd quirk when I was leaning Tidal.
For what it's worth, the MIDI spec suggests a default velocity of 64 (50%) for non-velocity sensitive controllers. Not sure how common that is in practice, though.
Conforming to the spec seems sensible to me. Seems like the current default is amp 0.4 and a desired default might be amp 0.5. I might submit a github issue on this unless others have different input.
There's usually a convention in Tidal of scaling things between 0 and 1 when there are absolute minimum and maximum limits. I think this is why amp ranges from 0 to 1 instead of 0 to 127. However, this convention is not always honored in Tidal, for example the values for ccv range from 0 127 even though they have a hard limit. So I don't know what the right answer is for proper MIDI velocity control. Perhaps both an amp control that ranges from 0 to 1 and an ampAbs that ranges from 0 to 127? I don't know what's right.
So no more gain affecting midi anymore by design? Darn, I really liked that unification of commands.
perhaps restoring the gain functionality should be the github issue, or an additional issue.
Hi sorry I've been super distracted for some days. I'd say gain not working for midi is a regression so a github issue would be great. @julian and I don't really use midi so depend on others to suggest sensible defaults etc, so please do make an issue / send PRs for that too, when there's consensus for something.
Thanks @yaxu. I'll make a github issue(s) when I find a moment, unless someone beats me to it.
The gain issue has been resolved in this SuperDirt PR: midi: fix midi gain by telephon · Pull Request #236 · musikinformatik/SuperDirt · GitHub