Piano World Home Page Forums Our Most Popular Forums Digital Pianos - Electronic Pianos - Synths & Keyboards What if Chopin had MIDI?

Clearly humor impaired folks like me benefit from the appropriate emoticon.

I see.One word:

If I had a nickel for every time I heard this argument

If a DP only had 8 total velocities, I bet nobody here would be able to hit every note at the exact velocity they intended in their standard repertoire. I bet at 4 velocities it would still be the case. But I think a DP supporting only midi velocities 0-3 for ppp-fff would be extremely limiting and clearly noticable.

Imo, it's NEVER about whether you can hit the precise velocity you want, it's whether there's an audible difference between velocities actually played.

If you can't hear or feel a difference, it probably doesn't matter to have more. If you can tell the difference, then maybe more velocities can be heard as well.

Yes. I don't understand how the inability to reproduce velocities accurately means fewer velocities are necessary.

It’s not about being able to hit the exact velocity on instant. But rather in a sequence of notes, whether a trill, a phrase or a repeated note. While listening to your current note (and in the context of previously played notes that will create pretty accurate muscle memory in your forearms and fingers) you should be able to produce pretty consistent note velocities that follow. With that in mind, if you feel you target a particular note velocity (or range) but the actually produced values audibly jump around that and you miss some in between, then the velocity resolution is low.

With that in mind I think 127 is more than enough.

Yes, basically it is about the creative process. As an artist, you don't want to have any limitations to what you can or can't do. Listening to another pianist play on a digital piano might sound fine to you, but if you ask him or her if the outcome was what he or she expected they might say no. They might say they had other intentions throughout the piece he or she was playing but was somewhat limited by the technology.

That said, it's difficult to argue that more velocity layers are needed over the 127 currently available when most sampled libraries represent only a fraction of what is already available.

So I would never debate the velocity layer issue from a listener's perspective, but it is from the pianist's perspective and the limitations it may enforce on sensitive, experienced pianists. I think it is up to the individual to decide whether MIDI or digital technology will work for them. For anyone else to decide for them is hubristic.

All digital pianos and VST-s for at least the last 20-30 years support 127 velocity layers. I’m not sure what you’re implying, e.g. that only Pianoteq supports 127 velocities? In case you mean that, you’re wrong. All pianos can respond with smooth dynamic gradation in 127 volume steps. And many sample-based digital pianos and VST-s also produce 127 different timbres which is apparent in the excellent spectral analysis our forum member dewster used to perform on digital pianos in the past. No difference here between Pianoteq and digital pianos.

Not implying anything . Yes any MIDI capable digital piano is capable of 127 velocity layers, whether there is this smooth dynamic gradation in 127 different timbres exists in most sampled pianos today- I'd say that is open to debate. I certainly don't feel that smooth gradation- others may though. So to each his own. I happen to agree with the argument for and against modeled and sampled sounds by the way. They both have their respective strengths and weaknesses.

My suggestion is always to aim towards an acoustic piano at some point in your playing and all is solved.

I think that if a pianist has the feel that he/she isn't able to give the dynamics he wants in its playing on a digital piano, compared to an acoustic, the culprit is not the standard MIDI limit of 127 values for the velocity, but the imperfect mechanics/electronics of the keyboard action and/or the quality of the sound engine.

If you really need more than 127 velocity values, you should explain why. Let's suppose we have a VST with 127 velocity layers fully sampled one-by-one. Do you think a human ear could feel a difference (in timbre and volume) between velocity X and velocity X+1 (where X is between 1 and 127)? Would a person feel the need of a velocity X+0.5 ?

I guess we would never know for sure unless they got beyond that 127 level limit. No one's perfect when they play no matter what the level- and maybe 127 levels is all that can audibly be discriminated by even the most experienced pianists, but think about how many possible velocities a piano key can actually be struck. It must be close to infinity. I can't tell you how the brain is able to go from a thought to a tactile action when playing music, but I can guarantee it allows for far more than just 127 different velocities to strike a key. That key strike may be attenuated by your emotional state, the touch and feedback of the keyboard, the technical demands of the piece you are attempting and your own physiological/biomechanical constraints. All artists would prefer the maximum amount of sensitivity that an instrument would allow for. This 127 level limitation was a established at a time when I was using 5.25 floppy discs and Atari computers. Maybe this limitation was posed by the limitations of the technology at that time, but why should we assume this is the limitation of the human brain to discriminate? The point being, most artists/high level pianists would prefer the least amount of limitations to express their art to maximize the amount of creativity our brains are truly capable of.

I’ve seen similar arguments about 4K and 8K televisions not being needed because they are beyond the point of our brains to resolve the pixels. This may be absolutely true, however 4K televisions are nonetheless already here and 8K televisions are coming soon whether we need it or not. In the same way, the next version of the MIDI standard is coming and some number of years from now, it won’t matter if we can resolve 127 audio levels or 256 times as many because the MIDI then will simply have a higher limit.

And are we sure our digital pianos are not internally using much higher resolution velocity values? Maybe when exported through MIDI it’s only 127 but down-converted from a higher precision. In my DIY controller I have approximately 20000 steps between lowest velocity and highest velocity. And I recalculate that to produce a value between 1-127. I could easily imagine that my N1X might be even better than that taking in mind how advanced it is. Why would they bother to internally handicap it? Replaying the samples in the 16-bit dynamic range (even if not using the entire range) is exactly the same as replaying only 127 volume levels, there’s no difference.

Let's stop calling them "127 velocity layers" when they're not layers. They're just 127 digital stops on the volume control.

Of course an acoustic piano is capable of infinite timbrical and volume variations. That our fingers are able to generate much more velocities than 127, it's true. We are analogic beings, not digital. But this doesn't mean that the human brain is capable to "distinctly" perceive more than 127 timbrical variations in the typical dynamic range of a piano. Honestly, I don't think I'm able to clearly distinguish the timbre of all the 20 layers of my piano VST. For example, supposing they are all volume-normalized, if you listen to the 4th and then the 8th velocity layer, then you can hear a difference. But between the 7th and the 8th it's very very difficult. And they are just 20! So, why we would need more than 127?

Anyway, I don't want to limit our brain abilities, but without scientific proof of the contrary, personally I think 127 timbrical variations is enough for a piano. Maybe it's not for a violin.

Yes, it's possible. Most of the current DPs interpolate between layers, so the limit is just in the bit-depth they use to store audio values, and it would be stupid to limit a value like 32768 in a range 1-127 if you don't have to record a MIDI file.

Yes I don't know if this had ever been scientifically researched but from my understanding the 127 was just an arbitrary number that was established given the constraints of the technology available back in the late 70's/80's when MIDI was created. You bring up another point as well with the violin. This MIDI velocity limitation was not applied specifically towards MIDI keyboards, it applied to ALL MIDI instruments including drum pads and even electronic violins and more importantly beyond that these velocity layer implementation vary from one digital instrument to the next even within the same category of instruments. I remember because I played around with those. So clearly the MIDI engineers did not have any specific instrument in mind when they created the 127 velocity standard so it supports the notion that for the digital piano there more be more (or possibly less) velocity layers that could be discriminated by advanced pianists. I'm sure we can all agree that the mechanics involved in drum playing is different from playing the piano so why restrict both to only 127 velocity layers? Imagine if Da Vinci or Van Gogh were limited to only 127 levels of brush strokes. It must of had something to do with the practicalities of the implementation of MIDI at the time and the constraints of the technology available at that time as well.

I'm pretty firmly in the camp that these ultra high resolutions are a case of Emperor's new clothes really.

I would wager that one could even perfectly well use a digital with less than 127 velocities.

Yeah, but what of it? Do you think that aside from the retro-crowd, anyone would willingly buy a 720p TV, for example, when 4K TVs are about the same price?

and I'm pretty sure Kissin can make music on an 8 key toy piano but he won't be smiling.

Yeah that's a fair point. If it's around the same price then might as well buy it. I'd agree with that.