Sound designers: how to handle aftertouch and velocity conflicts

[Echoes in the Attic]

I've been having a lot of difficulty using patches that are modulated by both aftertouch and velocity and my cme xkey. For example if you want to be able to hit a key hard and have the filter open immediately or hit it softly and press in to swell the filter open. It's a pretty common type of patch you can see with a lot of roli equator sounds in seaboard videos for example.

The problem im having when I try to do this is that hitting a key hard seems to send an initial lower aftertouch so it glitches the sound and the two seem to interfere with each other. I get this problem both with Omnisphere and bitwig instruments, so I'm wondering if it's an issue with my cme xkey. On the other hand if I try an U-he synth with aftertouch and velocity modulation they don't seem to conflict and it seems to be smooth. But I'm not sure how conflicting modulation like this should work. Should they sum, or should one takeover.

I was recently playing some trilian synth bass patches for example which have aftertouch modulation and they are just totally screwed up if I play them with my cme. Unfortunately since I'm traveling right now the only other keyboard input I have is the bitwig touch screen so I'm not sure how other keyboards send aftertouch.

So I guess I have two questions:
Does anyone have this same glitchy issue with aftertouch keyboards playing patches like this? And
Is there a smart way to modulate a parameter like cutoff with both velocity and aftertouch so that it works smoothly?

Thanks

[aMUSEd]

I have this problem with my QuNexus and pretty much gave up on using it as I'd originally intended - it seems a limitation of the soft pad based keyboard approach that you must apply pressure to simulate velocity which is very hard to separate gesturally from applying additional pressure for aftertouch (unlike traditional keyboards where velocity is speed of depression and aftertouch is additional pressure applied after depression)

I did ask the QuNexus developers to make it possible to use tilt for aftertouch and pressure for velocity but they never updated their editor. Maybe Roli could do something similar and make them more distinctly different gestures? This is one reason, although I love the sheer tactile sensuousness of playing with the Roli keyboards, I was wary of getting one.

[Echoes in the Attic]

This is what I can't figure out. Do physical controllers have different ways of avoiding sending these conflicting messages? I mean I find it hard to believe that trilian has so many patches with aftertouch modulation that would be totally mangled with any keyboard that actually sends aftertouch so they must have work as intended wen designed. And maybe my cme works differently. It seems like it takes a second before the aftertouch is registered. So I hit with high velocity and there is a split second if fast attack but it seems like a default aftertouch value is then sentnwhich is lower, thus immediately reducing the modulation and then picks up again. I don't know.

But I bought (and returned) a qunexus as well and although I had other problems with it, the aftertouch seemed to work well I thought. Perhaps I just used it with synths that don't have this issue though. I remember trying imposcar 2 and diva, but I thought I tried Omnisphere too. But somehow diva doesn't get this issue like Omnisphere and bitwig synths do so I don't know if it uses some magical lag or something.

But I can see many videos or roli demo sounds having the type of response I'm going for: high velocity = fast attack. Lower velocity allows pressure to swell the sounds manually. Unless they always use separate oscs/layers for those different articulation.

[aMUSEd]

The CME uses a more traditional keyboard action though, it has firm keys not squishy pads, so the velocity of strike is easy to distinguish from pressure vs more pressure. They are distinct forms of midi data, the problem is down to how does the keyboard know what you are trying to do with your fingers.

[Gamma-UT]

This sounds as though it might be an issue of how different instruments handle control data - the u-he instruments may do more low-pass filtering of the controller data to make the filters respond more smoothly to sudden changes. One option is to increase the delay between a key-on event and when the keyboard starts sending aftertouch data - if that's an option for the keyboard.

[himalaya]

This is a complex topic.

Firstly, we have two different 'pressure' mechanisms being mentioned in this thread:

traditional aftertouch and continuous pressure.

Secondly, pressure/aftertouch just does not fit certain types of sounds.

Is there a smart way to modulate a parameter like cutoff with both velocity and aftertouch so that it works smoothly?

There is. You need expression curves, or 'envelopes' as a type of 'via' modifier between the mod source and the mod destination. So, if you take one of the new soft-touch surfaces, for example the Seaboard, which has continuous pressure, the best way to shape how this pressure will respond to your touch is by creating a suitable envelope shape, most likely an exponential shape, and even better, a shape with a 'null' modulation area, a flat zone, where there is no modulation, and this then joins with an exponential rise of the curve. This shape basically creates a pseudo-afterouch effect. The 'null' zone is faking the 'delay we experience when we use aftertouch mechanisms in normal midi keyboards. So, the end result is that you will be able to separate the velocity and pressure controllers more easily since the 'null' zone provides this buffer area where should you have a higher velocity value it will travel through the 'null' no-modulation zone, and then you will be able to press-in and get modulation on pressure.

This approach depends of your playing technique, which needs to be soft to begin with. This is a requirement for any keyboard design which uses 'continuous pressure'.

This is also dependant on the actual software as it needs to have these 'expression curves' in order to create this balance.

Another feature which is very helpful here is the 'slew' parameter, a time-based delay. Personally, I feel that this one needs to be used sparingly since large amounts affect the responsiveness of the continuous pressure modulation and playing a sound may feel like playing through some glue or something. However, when set carefully, most preferably with a suitable expression curve, it can really help to balance the velocity vs pressure response. I think the best example of the above two features are seen in Strobe2.

Another area which affects how we perceive the way aftertouch works is 'midi smoothing'. The stronger the midi smoothing algorithm, the more delay we perceive. Having less midi smoothing, and the sound feels instant but it will reveal parameter quantization artefacts.



Then, the actual use of aftertouch with certain sound categories.
Take a plucked bass as an example. Two things here:

1. This sound decays to zero amplitude.
2. Bass is normally played with great velocity force.

So if the sound decays to zero, the only way to bring it back to any audible level is to modulate the appropriate envelope -sustain- levels. It may be the amp envelope and the filter envelope sustain. This is normally possible in all synths. However, if you do so, and make the sound sustain vie pressure, then you risk modulating this pressure with high velocity values which reach into the pressure dimension, as it were. This happens on the soft-continuous pressure controllers. So at this juncture we are back to those expression-envelopes...but equally a question could be asked, why do we need pressure on a sound like a plucked bass which decays to 'zero' amplitude? It seems unnatural to sustain such a sound. And doing so, can create these issues, which may not be as troublesome as on a pad sound which is typically played softly, so the velocity does not 'fight' with pressure as much.


Lastly, some older midi keyboards/synths allow to tweak the aftertouch scaling. This can alter very slightly how easily aftertouch is triggered. I have this on my Roland XP80 keyboard which I use as my master keyboard. Not sure if CME allows to do this, maybe in some utility? Worth a look.

[himalaya]

By the way, on soft-continous pressure controllers, pressure can have the effect of 'velocity', in that if the sound does not have any velocity assigned, continuous-pressure will create a similar effect. It's uncanny, and really useful for certain sounds like reeds, woodwinds and such. Here, only a dash of velocity would ever be needed.

[Gamma-UT]

Then, the actual use of aftertouch with certain sound categories.
Take a plucked bass as an example. Two things here:

1. This sound decays to zero amplitude.
2. Bass is normally played with great velocity force.

So if the sound decays to zero, the only way to bring it back to any audible level is to modulate the appropriate envelope -sustain- levels. It may be the amp envelope and the filter envelope sustain. This is normally possible in all synths. However, if you do so, and make the sound sustain vie pressure, then you risk modulating this pressure with high velocity values which reach into the pressure dimension, as it were. This happens on the soft-continuous pressure controllers. So at this juncture we are back to those expression-envelopes...but equally a question could be asked, why do we need pressure on a sound like a plucked bass which decays to 'zero' amplitude? It seems unnatural to sustain such a sound. And doing so, can create these issues, which may not be as troublesome as on a pad sound which is typically played softly, so the velocity does not 'fight' with pressure as much.

I've programmed sounds with the characteristic of using pressure/aftertouch to a create low-level amp and/or pitch wobble after the initial attack, so the idea of adding this control to a plucked bass isn't very outlandish. The other thing I'd say is that a plucked bass is often not played with high force. Outside EDM styles, one of the beauties of a plucked bass sound is to have a lot of dynamics between a gently stroked note and one with a much stronger pluck, up to the full funk slap. I do occasionally program a pitchbend into the aftertouch for these, preferably with some kind of thresholding function using a mod-multiplier so that it only happens on high velocity and pressure.

[himalaya]

You should try programming those sounds on a continuous-pressure design. It's a totally different approach. You wouldn't be able to get away with using pressure without negative playability artefacts.

But, I'm not suggesting that modulating a sharply decaying sound with presure is wrong or "outlandish", I'm just providing another area to look at should aftertouch pose real problems with such sounds.

The other thing I'd say is that a plucked bass is often not played with high force

For 'plucked' read 'decay'. I'm not referring to an acoustic plucked bass, but a type of envelope response. A decay shape, which normally is associated with the word 'plucked'. But even so, my experience differs, in that even acoustic plucked bass sounds are played with higher velocities which would pose a lot of issues with continuous pressure mechanisms. It's a serious issue.

[Gamma-UT]

You should try programming those sounds on a continuous-pressure design. It's a totally different approach. You wouldn't be able to get away with using pressure without negative playability artefacts.

I'm using a Linnstrument as well as a regular controller keyboard. Some synths are better behaved than others. However, I'd say that those with a good mod matrix are better behaved on the whole. Filter modulation on IK's Syntronik with MIDI from the Linnstrument needs rather different treatment to Zebra, I agree (or a bit of preprocessing in something like Max).

For 'plucked' read 'decay'. I'm not referring to an acoustic plucked bass, but a type of envelope response. A decay shape, which normally is associated with the word 'plucked'. But even so, my experience differs, in that even acoustic plucked bass sounds are played with higher velocities which would pose a lot of issues with continuous pressure mechanisms. It's a serious issue.

I was thinking of the envelope response. However, I tend to approach those sounds with the mindset of a bassist, who are often careful not to overcook the attack because of the twang and string buzz that results. I think people have to take into account the reality that the more performance-oriented controllers like the Seaboard or Linnstrument are going to have less predictable playing characteristics because there is so much going on. The Eigenharp in particular is really tricky to control unless you dial back some of the sensitivity as well as adjust your playing style.

Then again, I'm not selling sounds, just setting things up for what I'm doing personally. So, the issue of trying to adjust for different playing styles is not an issue I'm forced to deal with.

[Echoes in the Attic]

The CME uses a more traditional keyboard action though, it has firm keys not squishy pads, so the velocity of strike is easy to distinguish from pressure vs more pressure. They are distinct forms of midi data, the problem is down to how does the keyboard know what you are trying to do with your fingers.

I don't think it really matters if its traditional keys or pads or squishy things. There is a measurement of how fast something is hit and then a continuous measurement of how hard it's presssed. In all cases a fast strike will give an initial hard pressure so I guess there have to be options about when a controller should start to send aftertouch and what the initial value. And indeed I did discover a cme xkey setup application that I should have already used, which has a setting to send the initial aftertouch value as the same as the velocity, which I think makes sense. Or you can send zero to start but then you probably get jumps. Also there is a setting for a delay in the aftertouch being sent. I actually found that this was causing more trouble and don't see a huge benefit so reduced this to the minimum setting which is 10 ms.

So with the hardware settings better, I've also discovered a few things I hadn't thought of before. For example if you modulate cutoff with aftertouch and velocity positively (so cutoff is initially lowish) then you can hit it softly and aftertouch will bring it up, or hit it hard Causing it to be initially bright. However modulating the cutoff directly causes the filter to stay open even if you adjust your pressure. But, if you instead use velocity on the filter envelope, along with direct aftertouch cutoff modulation, it can give the best of both worlds with the right moderate decay and a low sustain. So an initial soft velocity still lets you open the filter with aftertouch but then with a hard strike, you can keep it pressed hard to keep the filter open, or let off the pressure (after the decay time has passed) which will close the filter as you press less hard. This is because the sustain is low/nothing so in effect the sustain is however much pressure you apply because the aftertouch brings it up. So basically you're telling the velocity driven modulation to close quickly so that the pressure becomes responsible for how much the parameter is modulated.

Not sure if I explained that well but it seems like a decent way to be able to control something with both velocity and pressure.

Now I'll read the rest of the responses!

[Echoes in the Attic]

This sounds as though it might be an issue of how different instruments handle control data - the u-he instruments may do more low-pass filtering of the controller data to make the filters respond more smoothly to sudden changes. One option is to increase the delay between a key-on event and when the keyboard starts sending aftertouch data - if that's an option for the keyboard.

Indeed but I think I found a shorter delay to be better but thy might have been to do with the envelope decay I was working with. I think the key to the type of expression options I was going for was to make sure the aftertouch kicked in before the envelope decay was noticeable.

[Echoes in the Attic]

Thanks for the rest of the replies to. I think I'm following and I think I came to somewhat of a similar outlook using a velocity controlled envelope alongside pressure.

One thing I'm very surprised by though, is the aftertouch modulation all over so many bass synth patches in trilian that result in a complete mess of a patch if you actually send aftertouch. They don't seem like they were well designed to use it but they might have been using a different controller that worked differently, I don't know.

[himalaya]

One thing I'm very surprised by though, is the aftertouch modulation all over so many bass synth patches in trilian that result in a complete mess of a patch if you actually send aftertouch. They don't seem like they were well designed to use it but they might have been using a different controller that worked differently, I don't know.

It's likely that those patches were designed on a midi controller with deep exponential aftertouch scaling, which would explain why aftertouch is so responsive on a controller with a more sensitive pressure mechanism (maybe with linear scaling).

[aMUSEd]

The CME uses a more traditional keyboard action though, it has firm keys not squishy pads, so the velocity of strike is easy to distinguish from pressure vs more pressure. They are distinct forms of midi data, the problem is down to how does the keyboard know what you are trying to do with your fingers.

I don't think it really matters if its traditional keys or pads or squishy things. There is a measurement of how fast something is hit and then a continuous measurement of how hard it's presssed.

I think it does. A traditional key has to travel mechanically a certain distance at a certain rate to create velocity, it doesn't actually know how fast your finger is moving until it hits the key, and in fact many players, myself included, tend to keep their fingers on or very close to the keys and don't actually hit them from very far off but just press them. Aftertouch is produced when, once you have pushed the key all the way you push just a little more, there is some give in the key design for keyboards that support aftertouch that allows for that to then be measured as an aftertouch signal. With a softer pad again it can only know how fast your finger is moving by how far the pad is squished down, but the difference here is that it's the same signal that also tells it to produce aftertouch, there is no difference between squishing it to create a velocity message and squishing it to create aftertouch except in the timing so in practice one has to rely on velocity curves to tell it where to draw the line, which means it's harder to play in a way that clearly differentiates the two.