Piano World Home Page Forums Our Most Popular Forums Digital Pianos - Electronic Pianos - Synths & Keyboards What DPs really lack (compared to the acoustic pianos)

Of course the car analogy (like most analogies) is imprecise. Nevertheless, even someone who can produce only a dozen or so distinct values on demand (i.e. each within a repeatable margin without overlapping another) will benefit from having more values than that available (for example, in a crescendo). Again, it's not about your ability to repeat, it's about the ability of the instrument to respond to what you do, regardless of how much or how little precise control you have over what you do.

Yes, a piano with more resolution will respond differently than one with less.
But my ear's resolution matters.
The piano only has to be as good as my ears. Or yours.
More resolution or more precision matters only if you can hear the difference.

Okay, so what's the ultimate output of this? How many levels is enough for you? 127? Thousands? Milions? 10^10? I don't see the practical answer, apart from hints on "no finite amount is enough"...

The answer is that other things are SO much more important than whether a DP has the standard 127 levels or more that that should be about the last thing you should be considering at this point, even if there WAS universal consensus about it, which there isn't. Don't worry about it. Most people are playing on boards with 127, and most people are fine with it.

Yep. As far as volume is concerned, I agree. However, for timbre, I don't think a handful (1, 2, 3...) of different tone qualities is enough. It sounds very artificial.

As a side note, I'd be very interested how "interpolation between sounds" works...like, you recorded ppp timbre, mf and fff. Now let's say you play p...what algorithm interpolates between those pre-set timbres? As for velocity I kinda get it, just make ppp louder or mf quieter, but how do you combine two different sounds? (i.e. the question for me is..."how is linear combination of two tone qualities defined?" - or any combination, for that matter)

It's really frustrating for me because I mean no ill towards anyone here. I have written in the past my frustrations with digital pianos I have owned 3 or 4 digital grands over the past 20 years. All of them had sampled sounds and it was not the quality of the samples that bothered me it was the fact that I felt so disconnected from the music because the lack of sensitivity the digital piano offered. This is the same disconnect classical pianist Hugh Sung in those videos I posted comments about.

YES, the lack continuity, the lack of precision, the lack sensitivity produced by sampled pianos when trying to shape a phrase, produce coloration or maintain an even line IS VERY VERY VERY OBVIOUS even on the best sampled VSTs today because the technology that is used has not fundamentally changed.

Hugh Sung already explains the reason for this and don't know if I can make it any clearer.

When you learn the piano you spend decades trying to make connection between your brain and the amount pressure and velocity you apply to each individual keys with each individual finger you possess. You could call this sensitivity. (Good piano is not just about playing fast BTW). It's this sensitivity that allows you to express yourself through your music. Your brain is not limited to only 127 ways to press that key. The possibilities are limitless and your ability to manipulate those possibilities plays a part in your ability to be a good pianist. That's why music is considered a "fine art".

When you play an acoustic piano there is practically no limitations to how fast or hard you can strike a key the only limitations being your strength and speed and the physical limitations of the piano. But in the digital world due to limitations with technology and standards were established decades we are limited. We are limited by 127 levels/gradations/velocities whatever you want to call it and that IN ITSELF is limiting. But the forefathers of this technology felt at the time given the limitations of technology itself or the intention of digital instruments at the time when MIDI was conceived thought that 127 levels would suffice. I don't think they ever had trained classical musicians in mind when they designed the MIDI.

Now take those 127 gradations/possibilities/velocities (whatever) and now handicap it even further by saying I am going to take only 10 instances out of those 127 instances to represent the entire group of 127 possibilities and spread them out hoping that no one would notice. Well, that's exactly what sampled sounds do. They take a microphone and they have a machine or a pianist strike a single 10 different velocities and record them. They record from the softest for example to loudest. So instead of recording let's say at level1, level2 level 3, level 4- They record at level 1 and then immediately jump to say level 8, and then jump to level 16 all the way up to level 127. When you are playing a digital piano and you strike the key, all you are hearing are RECORDINGS, but those recordings are limited to only 10 levels per key.

Pianoteq doesn't do that. Because it's generating sound in real time using the same digital medium (ie 1's and 0's) as sampled sound using fancy algorithms you are given the full palette of instances that MIDI has to offer (127 to be exact). So theoretically with Pianoteq you can strike a key using 127 different velocities compared to the 8-10 velocities found not today's digital pianos or software VSTS that use sampled sounds. For the novice pianist 8-10 possibilities may be enough and appears more accessible, for the rock band keyboardist it may be just enough, a trained classical pianist or a serious learner this would be a disaster.

So to use an example if I asked you to play a series of 50 notes legato from ppp to FFF and back to ppp, if you only had 10 possibilities with a digital piano and 127 possibilities with Pianoteq, you would clearly be more limited with the digital sample then with Pianoteq and this IS exactly what us experienced pianists hear when we play with a digital instrument with digital samples. There are too many levels missing. There are too many gaps that it's hard to phrase a musical line cleanly or produce colorations or find the exact tone we looking for. If you are listening to just one note from attack to release, yeah I'm sure the sampled sounds will sound nice, but music is about connecting tones and this is where sampled technology falls flat and why many pianists consider digital pianos nothing more than "toys".

Cybergene likes to believe that sampled sounds solves this by "morphing" when what he actually means is blending sounds. Our brain has the ability to fill in gaps to a degree much like when you watch TV you fill in all the pixels. But to use a visual analogy. Can you tell the difference between compact disc and blu ray? I can. The brain has the ability to blend with compact disc but still the picture is not as clear as with blu-ray. Why? Because there is more data available visually in blu-ray versus compact disc. It's the same thing with Pianoteq versus sampled sounds. There is more data available to the auditory representation is more accurate. When it comes to the fine arts we want blu ray data not compact disc and yes you absolutely can hear these differences as clear as night and day if you know what you are supposed to be listening for.

anotherscott thinks there are other things more important than the levels. And I agree.

I think the most important thing is the fingers on the keys making music.

I prefer to judge the quality of the results, and leave alone any worries of how it got that way.

Modeled pianos, at least theoretically, can generate a different timbre for every velocity. A sampled piano that uses 3 samples per note will play the quietest samples (at different levels of loudness) for the lowest velocities (say, 1 through 40 on a system of 127), the mid-level sample for middle velocities (say, 41 though 80), and the loudest velocity sample for the hardest struck notes (say, 81-127). They may also use other techniques like filters or crossfades to smooth transitions or provide additional timbal changes. There are plenty of keyboards and software systems that use more than 3 samples per key, but certainly none that use anything like 127. Some people prefer modeling because of the lack of the limited number of available velocity timbres, other people feel that while it may have that technical advantage, modeling (so far) falls short in other ways compared to their sampled pianos of choice.

But ultimately, you'll get your P515, and either you'll love the way it plays, or not, and all the rest is just academic. Though sure, since there's a free Pianoteq demo, you should try it and see what you think about that as well.

I totally agree, but you need to have an instrument that responds the way it's supposed to respond. The best advice I would give is just stick to an acoustic. Can't go wrong there. I just had my Shigeru SK2 delivered today before the rain came and I am am in auditory heaven ....

Jethro, although I enjoy Pianoteq (I use it) - your theory kind of goes out of the window if it's true that most pianists, even concert ones, can only achieve about 10 (give or take) different levels of velocity. Hence, why the sample packs use 10 different samples.

However, there is weight to the fact that one person could have a slightly different, indeed offset, set of velocities. Pianoteq surely has them all covered in that case.

I just look at it as different pianos. Pianoteq, digital pianos, sample packs, acoustic pianos. I never thought there was an "ideal" sound for a piano. In fact I think it depends what you're playing.

Where did you read that concert pianists can only achieve 10 or so velocities per key?

There's a bit of truth in this:But it ends at the limit of human perception. More is only better until it makes no further improvement.

I can't see this at all:That's very poetic, but everything has limits.

This is mostly true:
But should we care about what the pianist does to the piano? I think only the results matter. It's the sound! The music!

As for 127 velocity levels being enough ... I'm sure it's more than enough. Way more.
If every level were spaced 0.5 dB apart you could span 64 dB of dynamic range. And no one can distinguish sounds that are a mere 0.5 dB apart.
Is that what led to the choice of 127 velocity levels 40 years ago? Or was it that 127 levels neatly fit into 7 bits of data? Or both?
I don't know. But the choice makes perfect sense.

Also ... to say that a digital piano has only 10 sampled levels was once true.
But even my ten-year-old copy of Ivory 1 has 18 levels.
Some newer virtual instruments have many many more. (Does anyone know?)
Also please re-read the many threads in which there is discussion of layer blending.

I think you meant DVD vs. Blu Ray?
If so ... yes, we can all see the difference. It is easy to conclude that DVDs do not reach the limits of visual perception. So more is better. Blu Ray is better.
But, again, more is only better when you can tell the difference.

Here's a nice article that explains a little about how sensitive trained musicians are. The differences between Pianoteq and sampled sounds CAN be to this degree but the major differences between Pianoteq levels of sensitivity compared to sampled sounds level of sensitivity should be clear to an intermediate to advanced pianist.

Can you hear the differences in tone in the recording. I can. And this is what I'm referring to when I say it's hard to appreciate something when you have a difficult time knowing what to listen for. These things won't appear to matter now but it will hamper your development as you reach the intermediate to advanced stages of piano study. This is what I mean when I refer to sensitivity to what you are listening to. Can you tell that the second tone was struck slightly harder and with more attack? You wouldn't be able to create this tonal difference with a sampled sound but I bet you could with Pianoteq. Listen to it carefully.

If you don't have the proper tools to reach this level of sensitivity you are only handicapping yourself.

http://physicsbuzz.physicscentral.com/2014/12/hearing-pianists-fingers-importance-of.html

Just a quick question: would I be able to hook up Yamaha P-515 to my computer and use Pianoteq for generating sound and send it back to its speakers?

Yes. I do this from my MacBook Pro with a single USB cable connected to the P-515.

Noice! Thanks.

But just remember what Joe T told you about the P515 and Pianoteq. The notes were all over the place and it sounded like it sounded unplayable. That's where the VPC-1's professionally calibrated touch velocities come into play. I don't know if you will be able to adjust the velocity setting for Pianoteq on a P515. For the onboard sounds yes, I don't see an editor otherwise but I may be wrong.

That's true... but I don't think we should assume that 127 velocity levels calibrated to be .5 dB away from each other would necessarily produce the most naturally playable piano,


If they change in timbre as well as volume, such small increments would be easier to detect.

VSL's pianos have between 60 and 100 velocity layers.

Isn't Pianoteq's Bechstein DG based on a sampled instrument, the Bechstein Digital Grand?