Velocity to Parameter Modulation. I'm doing something wrong

[Jake Jackson]

I've tried the three variations illustrated below for Vel to Unison detune, and all have the same flaw. They play the note AND change the parameter, but the effect of changing the parameter is only heard if the note is played a second time. In other words, with any of these settings, if I play the note twice, the second strike reacts to the parameters changed by the first, but the first strike plays without any change in the sound.

(I would not use all three in one Mux in reality, nor have all three attached to the Event Output. They are all attached only to show the complete signal path.)




Uploaded with ImageShack.us


In the variations with an ADSR envelope in the chain, the attack time is set to 0 and the Decay Level and Sustain Level are set above 0. Clearly, I'm doing something wrong, but trying different arrangements and connections has led nowhere. (The variations in sound are still interesting, since the 2nd strikes do change, but...)

[AndreasD]

If you are talking about Mulab/MUX osc and the fatness parameter, indeed all those methods will fail. To my experience the MUX osc do not react to changes of fatness, detune and spread after the osc has started playing. Which means these params are only checked/used at note on.

Solution: delay the event going into the osc. I can be wrong, but you might need at least 5msec delay.

[Jake Jackson]

If you are talking about Mulab/MUX osc and the fatness parameter, indeed all those methods will fail. To my experience the MUX osc do not react to changes of fatness, detune and spread after the osc has started playing. Which means these params are only checked/used at note on.

Solution: delay the event going into the osc. I can be wrong, but you might need at least 5msec delay.

Thanks. But...I was using another VSTI and trying to modulate the parameters within it by velocity. (Not an outside effect.) I was using the unison detune, a parameter inside the VSTI, as an example. All of the parameters react the same way.

I tried to set up a delay going into the VSTI, using the middle arrangement in the screenshot above, by putting an Event Delay in between the Note Key\Vel module and the Note to Modulation, but it does not delay the signal. I also disconnected the ADSR from the Modulation Map and connected it to the parameter directly, but without the result I want.

If I put a delay anywhere else, I meet the simple problem: If I delay the event, I am also delaying the velocity modulation of the parameters, so everything reacts a little slower, but the result is still the same.

[Jake Jackson]

Sorry for two posts in a row. I did get the signal to delay, but the problem remains: the signal is delayed, but the delay only rarely lets the parameters shift before the sound is triggered, and only when the delay is set to several seconds. (The 5 milliseconds and up to several hundred milliseconds does nothing but delay the signal, which still is not changed by the modulation on a dual core 2.3 gHz system.) The fault is probably mine.

Somehow, although the delay is only on the Event to Mixer to VSTI side of the chain, adding the delay sometimes seems to prevent the parameters from changing at all until the second strike. Then the one-note behind situation sets in. Not noticeable if you play several hard strikes in a row, but very noticeable if you play a soft strike followed by a hard strike or the opposite--the second strike has the parameter change set by the first strike.

Sorry to keep posting about this. So close but so far. Does anyone have a setup in which the event modulates the parameters on the first strike?

[AndreasD]

I tried to modulate unison by velocity with instant success for
- Korg MonoPoly
- Waldorf Largo
- U-he Diva



No delay needed, no glitches, nothing. All worked instantly.
I guess your problem is an issue of your VST.

For debugging purposes you might check the output of your Parameter Event Generator with an Event Monitor (found under 'Other' modules)

[Jake Jackson]

Thanks. Working with another synth, I found that you are right--the modulation does occur as the note is struck. I was able to modulate a lp filter with a steep cutoff that was reflected on the first strike.

I've also tried:

Setting the various amplitudes and offsets in Pianoteq, and setting its parameters to their maximum and minimum in various combinations with vel.

Putting a CC Modulation Generator in the chain instead of the Parameter Modulation Generator. The result was the same that I'm getting with the PMG: Velocity changes the parameter setting instead of modulating the parameter as I play.

The problem may be that it is not a solvable problem: Pianoteq is of course physical modelling, so parameter settings may be calculated before the note is produced. Unlike a filter, that occurs later in the signal chain. On the other hand, part of me doesn't understand. If the parameter change reaches the vsti and intercepts the signal before the partials are generated, it seems that it should still affect the note, just as it affects the note if the user manually changes them on the interface before playing the note.

I'll write to the good people at Modartt to see if there is a way to create the mux. But any suggestions here would be appreciated. Thanks again.