need for larger velocity scales, 0-128 to small

[EvilDragon]

Best solution: global custom velocity curve(s) (why not having several user velocity curves?), then per program velocity curve override, then in Combi mode velocity curve override per timbre/zone.

[ozy]

Best solution: global custom velocity curve(s) (why not having several user velocity curves?), then per program velocity curve override, then in Combi mode velocity curve override per timbre/zone.

i.e. buying a pc3

ok, ok, sorry, I shouldn't have said that. It was unwarranted and inappropriate, sorry.

(and: the base curve is not exactly "custom". And: the ovverride is not a total override, it is an offset+scale parameter).

Told ya: it was inappropriate.

But the sensible part is this: it can be done. Now.

[EvilDragon]

Yep, it's almost there on the Kurz, except having a custom user velocity curve.

[jemkeys25]

hey evil dragon, what about the kurz?

[EvilDragon]

On the Kurzweil you can change the velocity curve per zone in Setup mode.

[synthjoe]

It could be that we are both correct. Because there is a Pre-MIDI and Post-MIDI setting (on Korgs) as well.

I believe it works like that since Trinity. Korg M1 and T-series had the curve only prior to MIDI - if I'm not mistaken. The selectable curves are unfortunately fixed, though.

So, speaking generally... what this shows is that it's not enough to just say that MIDI doesn't provide enough steps (values) for velocity. We have to distinguish between mapping MIDI velocity values to form a new response curve, or, leaving the MIDI velocity curve alone (typically linear) and mapping those values within the sound engine to achieve different response curves.

That's why I agree with the OP and would prefer to have more values with linear mapping rather than a bunch of proprietary curves, to each manufacturer their own... Hence HD MIDI looks to be a good solution. Also, I'm fairly sure that post MIDI curves were implemented only to cover for keybeds/controllers not having an adjustable curve, in my mind it should be tied to the keyboard's response rather than to the sound engine's (though I'm sure more than a few will debate that ).

[Kayemef]

There's a lot of talk around here about problems with MIDI range and how this translates to "limited dynamic range" on the keyboard. I don't think the problem is with the current MIDI resolution, but rather on how this range is used to quantise a player's playing...

If I understand correctly, "velocity" is calculated as an "average velocity" - basically, its the distance travelled by the key from top to bottom divided by the time it took to reach it. As we all know, the distance from "offset" to "onset" is a constant... so the only variable here is the time it takes for the key to travel it. All you need are 2 sensors, one that triggers when the note starts moving and another one at the bottom that triggers when the note hits the "note on". The "faster" you play, the louder it gets, its not about pressure, its about speed.

Now here's the problem... that not how a real piano works !

When you strike a key on a real piano, a hammer starts moving in on the note's strings, the "loudness" of the note depens on the energy left in the hammer at the MOMENT when the hammer strikes the strings... Anybody who has ever done some physics and 2nd degree math with speed and acceleration might see where I'm going here...

Try this example :

Lets say that you have an motion from key up to key depressed that lasts for 5 ms. On a "velocity sensitive" keyboard, that means that every motion that lasts for 5 ms will have the same MIDI velocity output.

On a real piano, this is not the case. You could have 2 very different articulations both with a motion time of 5 ms. If you want to play a fast leggerissimo, you have to move your fingers very quiclky and gently "snap" the keys, without pushing all the way through. This flings the hammers towards the keys as they slowly decelarate before finaly hitting the strings with whats left of its energy (preferably, very little). Its then possible to play very fast, yet very softly.

The opposite is also true. Remeber Bruce Lee's famous 1 inch punch. The idea was to acelerate into the target and have as much energy as possible at the MOMENT of the impact. You can do the same thing a real piano, even if the motion lasts for as long as 5 ms...

Everyone feels that MIDI keyboards aren't as accurate as a real concert grand, but maybe it's not about midi range, I think it's a lot more about the fact that they probably use only 2 sensors from start to finish. If they would use 3 or more, they could accuratly map the key's path in time and then with some pretty simple math calculate the key's speed at the MOMENT of impact.

We're barking at the wrong tree here, we don't need another MIDI standard, I really believe 127 steps is quite enough. What we need are better keybeds that emulate the physics in a real piano.

I don't really know how they mesure velocity, but from my playing experience, I'm pretty sure that what I'm saying here is right on the spot. If anyone can provide some detailed info on this, I would very much like to know...

The keyword is MOMENT...

Any thoughts ?

[peter_schwartz]

I think this whole discussion, interesting as it's been at times, is academic. To a few of your points... I don't think a reasonable person expects an electronic keyboard -- weighted or not -- to feel like a real grand. I don't think they expect the accuracy of nuance either (unless they're unreasonable). Otherwise, we're comparing apples and oranges. If you want the nuance of a grand piano (assuming that its action isn't beat to sh!t, is in tune, doesn't have a squeaky pedal and so on) then gowon and play a piano! But don't expect it to be multi-timbral or have any of the 10 billion other features that you'd find in an electronic keyboard (Kronos or otherwise).

Here's what I don't get... people have been making music on MIDI synths for nearly 30 years. And it's rare that any one keyboard gains a reputation for not having good velocity scaling in its programming. Hmmm... eh? And with that, the whole argument abut 127 steps not being enough starts to lose substance...

So yeah, would more steps be nice? Sure! But are the 127 steps we have now substantially preventing anyone from making good music? Doubtful. Highly doubtful. I can't imagine anyone here can honestly say that they abandoned playing live, or they've given up on completing a track because the 127 steps of velocity weren't enough.

[Kayemef]

Well, sorry for being a geek, for we are quintessentially academic, but this is exactly what progress is about and how new ideas are born...

I don't think a reasonable person expects an electronic keyboard -- weighted or not -- to feel like a real grand

A few years back, I don't think any reasonable person would have expected to see what happened with the smartphone revolution (espicially iPhones)... Guess who didn't buy Apple shares 10 years ago ? People expected their phones to do one thing : call. Today, we feel cheated if we can't film a bloody movie with it.

We expect our keyboards to behave differently than their accoustic counterparts simply because they have always behaved that way.

With today's processors and their computational power, it would be so simple to integrate a smarter way to determine note velocity. Would I pay and extra 500$ to have a keybed fitted with multiple sensors on each note ? If it means I can really get the same feeling I had when playing on 15k+$ concert grand, I thnk the answer is F&CK YEAH ! As for your suggestion that if I wanted to play a piano, I should "gowon" and play a piano, I'll kindly inform that I don't have the money for a concert grand, I wouldn't know where to put it and on top of that, I'd need some pretty devoted roadies to carry it around !

Fact is : people do feel that something just doesn't feel right with their velocity response. 5 pages of forum discussions say so ! I thought it might be interesting to consider an alternative to the velocity curve dilemna.

There is a lot stuff I used to be able to do on a concert grand that I can't do on my Oasys, although I do have to admit that the opposite is also true. Bottom line is, we don't have to choose between apples and oranges, or having the cake and eating it, or w/e... We can have both.

Yes we can

The technology is out there, why aren't they using it ?

I was hoping for some feed back from the nice folks at Korg... maybe I got it all wrong and maybe they do implement some sort of "Positional Motion Sensor" or PMS technology... maybe they don't call it that way. Who makes the RH3 keybeds anyway ? I heard somewhere that Yamaha and Fatar are the big players on the keybed making biz... Anyone ?

And dude, if your piano is beat to s**t, isn't in tune and has a squeaky pedal, you should definitely sample it - you'll find that the sound has a lot of character.

[cello]

Great points Kayemef!

I would hate my piano to play like my O & vice versa! They are instruments in their own right and should be enjoyed as such.

I like my Radias and T3 actions because they DON'T feel like a piano - anyone who plays a piano even just rudiments, will know that the when playing a synth (ie non-piano action) needs to hold keys down longer (or shorter) than the not value suggests simply to get the selected waveform to 'do its thing'. My O has a great action in that it's quite close to a piano but definitely not, and when playing 'synth' sounds it's easy to convert playing style to suit.

The piano sounds however are all related exactly to note value and expression and of course dynamics (and pedal action too is in the mix).

So, frankly it's a debate I've never really got my head round. I don't want any of my keyboards to be anything like a piano action. Yes, the protocol is common (ie black and white notes) but the execution and requirements are different.

[jemkeys25]

music is about feel, conveying an emotion to the listener, and yes electronic music has been around for decades with great success, as well as contiued success, to this day, no one will dispute this, your common non musician doesn't know a velocity scale from al LFO, they only know if they like a song or not, but as musicians who've been using these instruments for decades have learned the limits of these instruments and have played within those limits, really having little choice, but we have also learned where improvements can be made, and if someone said to me" this can improve the emotional content of your music,would you like it?" i'd say absolutely, i wouldn't say, no , 127 is enough, everywhere you look all electronic equiptment is constantly evolving around us, aren't our instruments also electronic, shouldn't they evolve as well?

weather its better hardware or software or HD MIDI, for what ever reason you'll notice it just sounds better , you can't really pinpoint why, it just does, a piano sound that gives you everything you got and more if you have it, not" oh to bad dude, you've already hit 128, don't have anymore to give.

[EvilDragon]

But the point is that 127 is perfectly enough, with proper velocity curve and velocity sensitivity of the patch.

[Zeroesque]

@Kayemef, develop a keybed that puts the two timing sensors on the hammers, on the very last fraction of the hammer throw. That should get close to what you're talking about. Of course, this is only valid for the piano -- a percussion instrument. This would be completely wrong, however, for a harpsichord or organ...which is why I wouldn't particularly want it, unless you created a good approximation of those mechanisms ("waterfall" anyone?) on the same keybed and switched between them for different sounds. Could be an interesting experiment in playability.

[danatkorg]

I was hoping for some feed back from the nice folks at Korg... maybe I got it all wrong and maybe they do implement some sort of "Positional Motion Sensor" or PMS technology... maybe they don't call it that way.

Our keybeds use the standard method with two sensors.

Another way to do it is with continual position sensing, such as Ensoniq used in several of their keyboards. That gives you both velocity and polyphonic aftertouch at the same time. I get the impression that there were reliability issues with that implementation, however.

Who makes the RH3 keybeds anyway ? I heard somewhere that Yamaha and Fatar are the big players on the keybed making biz... Anyone ?

For most products, at least, Korg now makes our own keybeds.

Hope this helps!

- Dan

[EvilDragon]

Another way to do it is with continual position sensing, such as Ensoniq used in several of their keyboards. That gives you both velocity and polyphonic aftertouch at the same time. I get the impression that there were reliability issues with that implementation, however.

That was quite some time ago, yes - might be that as technology advanced in the last few decades, this method could be reinstated with better reliability? That would be pretty awesome, as I presume it's much cheaper than "standard" PAT implementation, right?