Piano World Home Page Forums Our Most Popular Forums Piano Forum will you still buy acoustic piano?...

"Whoever heard of an electric violin, electric cello or, for that matter, an electric singer?"

- Andres Segovia


Also, when asked what he thought about electric guitars:

"That is not a guitar, that is a completely different instrument."

I love that Segovia quote. Thanks for sharing it!

I agree with Segovia. Just as electric guitars are different, Digital keyboards are much different than Pianos. The same as a Pipe Organ is much different.

I would guess that most pianist (~80%) have a digital keyboard somewhere in the house.

As others have mentioned, the 3 dimension aspect seems to hold a clue, or several.

Pianos are not perfectly spherical radiators. The initial attack and first part of a note radiate primarily in the vertical plane, while the sustain and decay radiate primarily in the horizontal plane.

Think about what that means as the piano sits in a 3 dimensional space such as a concert hall or living room. You can see how complex the instrument's interaction with its environment is. The human ear picks up on this in ways we don't even understand.

That may be a clue as to why the first "not right" thing I notice about a digital is the decay of each note. This is true even with digitals made by companies such as Yamaha or Kawai, that also make well-regarded acoustics. One would think their sampling techniques would be a little more sophisticated than simply sticking a couple of microphones above their best acoustic grand. But the decay is always a giveaway, at least to me.

Nevertheless, I suspect it may be perfectly possible, even with today's technology, to build a digital piano that any of us could not tell from the acoustic original. But it would be so complex, cumbersome and expensive as to be pointless.

Can anyone tell me whether, and if so how well, digitals can reproduce the resonances in any undamped strings of the notes which being played? This is surely a most important feature of the sound of a piano. A note with the sustain pedal depressed sounds quite different from the same note without the pedal.

I think some digital with resonance board can reproduce this pretty well besides I don't think this is the biggest challenge
here anyway. Big enough speaker in right place, close to something that can vibrate and you go...
Check out the kawai CA93 for instance.

Simple: I will decide when this moment has come. But for the time being...

The Oberheim Minigrand piano module from the 90's I have models sympathetic string resonance. It also simulated pedal noises. There is also a slight shift in timbre with una corda too, but that might be an illusion due to volume or dynamic range compression.

So a lot of this sort of thing has been done in the past.

I am also an EE and I have thought about how to reproduce a grand piano using purely electronic means. This is what I think would be the best approach to doing it:

1. Start with a shell of a grand piano, the rim and shape.
2. In place of the soundboard, use an array of many 6 inch or so wide range dynamic drivers in a pattern to simulate the soundboard. These would be dipoles and radiate both "up" and "down".
3. The movement of the soundboard is dependent on the excitation point provided by the bridge bearing point and we can via DSP tailor the output of each speaker to simulate the soundboard. Obviously each speaker receiving the same amplitude and phase signal would not be right. Each speaker would need to have different output depending on the note being played. So we're talking about a fairly sophisticated DSP and on the analog side separate amplifiers for each speaker.
4. We can outright synthesize the piano sound by physically modeling it, but sampling at different volumes and interpolating might work as well. But we would need to de-embed the influence of the piano rim from the samples and have many samples per note. You can fairly easily simulate sympathetic resonance.

By using the original piano body, we can eliminate a lot of the uncertainties with how the sound radiates from such an odd shape. And treating the soundboard as the transducer that it really is and simulating that would be the easiest approach imho.

of course I will continue to buy acoustic pianos. Whenever I play a digital instrument, I feel like I'm being ripped off, tricked, fooled, i dunno, I just don't like it!

In answer to the original question "Will you still buy acoustic piano when can't tell the difference anymore between digital ones?", the answer would be YES! (If I was going to be buying a piano, which I will not be, since I have one already.) They are more beautiful, and "alive" to me in a way which digitals just are not.

Wow... I am impressed . This would indeed sound like a piano... and you can use different sound samples in addition to piano sound and see how it works through that soundsystem (I am thinking on electric pianos, clavicords, pipe organs or even synthesized sounds). I would really like to test such a device!

This is my thoughts about it: this hipotetical digital device that sounds as a piano (the sound emitted by this device fills the threedimensional space in the same way as a piano does) seems very expensive to make... and it's bulky (size is one of the main advantages of DPs).

If it costs like a piano , has the size of a piano and sounds like a piano... it is not better just to get a piano? I mean... let the digital sound as a digital and the acoustic sound as an acoustic. Both are great instruments. But not equal.

It would be so nice to have a small and cheap device that would be as good as a full-fledged acoustic piano. But I think it is a fact that we are not near that point yet. In the future? Who knows... for now, give me an acoustic piano, and a digital for practising at night or playing with different sounds.

But when that device will be small and cheap (or at least, cheap)... mmm... let me try it, and probably I will get rid of the acoustic.

Regards,
Kurt.-

Daifanshi, I am curious what you mean by physical modelling. Do you mean some sort of finite-element analysis of the forces in the soundboard, done in real time, with the vibrating strings as inputs, with the objective of modelling the soundboard motion? This might work, but would have to be extremely sophisticated. Since the strings are part of the vibrating system, they would probably have to be included in the model. This approach would predict the separate oscillations of different parts of the soundboard which you would require to drive the individual speakers.

If you were going to adopt a sampling approach, you would have to sample separately the soundboard oscillations at each speaker position. I think that this approach would get into difficulties when combinations of notes are played at once, since it seems to me that each combination would have to be sampled separately. And the number of combinations, possibly at different volume levels, would be enormous.

The Kawai CN-43 does this. The effect is adjustable with menu settings. There are separate adjustments for string resonance (sympathetic vibrations in keys held down while another is struck, no damper pedal) and damper pedal resonance (entire resonance with the damper pedal down).

I would assume the entire CN series has this feature, and probably the CA series, as well.

"Physical modeling" in this context would be to mathematically model the interaction of the hammer to the string to bridge to the soundboard as a function of time. It would be a complex equation, but it can be done if one was motivated. Modeling an instrument has been done in the past. The equations describing something like the pipe resonance as found in an organ aren't really that complicated when you look at it. A piano I expect would be much more complex.

We can empirically determine how a typical soundboard behaves by attaching accelerometers to many points around a typical soundboard and along with some laser interferometry we can get a very good idea of how they work. Doing this using FEM is a possibility since the soundboard is basically a diaphragm. But creating an accurate model for the soundboard to even start the analysis could be difficult.



If we understand what the displacement vs. time function looks like for an arbitrary point on the soundboard for different frequencies which we should be able to figure out with good approximation experimentally, then we can de-embed the effect of the soundboard radiating transducer from the samples. I don't believe we would need to sample combinations of notes since each note would have it's own set of unique characteristics associated with it and any interaction will happen on the physical level in the speaker array and in the DSP on the mathematical level like in the way a Fourier Series can decribe it. But the vast majority of interaction will already have happened before the signal reaches a particular driver. Realistically I expect a modern hi-fi speaker to have a vanishingly low amount of distortion compared to a typical piano soundboard.

Any added harmonic content (both linear and non-linear) due to many simultaneous notes will be present in the signal. And any of the original harmonic content would already be present in the samples. If somebody manages to play 88 notes simultaneously with the sustain pedal I can see how we will probably need some serious DSP horsepower. But it's not in the realm of impossibility.

The title of the thread is a loaded question. It would have been better to ask "would you still buy an acoustic piano, if at some time you can't tell the difference anymore between digital ones?"

With respect to the topic of digitally modelling pianos -- as opposed to sampling them with microphones -- one ought to mention Pianoteq software. One advantage this software offers over acoustic pianos is that many variables can be changed. One example is the ability to model very long bass strings for a unique sound that strikes the ear as being a recording of an acoustic instrument -- one far longer than even the longest concert grands. (I looked for a video demonstrating this, which I saw a while ago, but couldn't find it quickly, unfortunately).

I don't think that will happen some day. In my opinion, the digital ones will continue improving ad infinitum but I don't think they will reach the perfect emulation. That's, in my very humble opinion, because of that word: emulation. The original thing can only been obtained with the original thing. No speakers will emulate ever the 100+ strings of a piano. Maybe in 20 years from now they will invent digital strings? I don't know. But with the technology of today, it's just impossible, telling only one of the hundreds of aspects a complex instrument like the piano has.

But this doesn't contradict the fact that digital pianos today are an impressive working tool. They are, as someone quoted Segovia before, in fact a new instrument. In my opinion, every pianist must have an acoustic instrument and also a digital, because of the advantages of each. And we must bless the fact we are living an era when very good DPs exist, because in some cases people who cannot have an acoustic one (whatever the reason, budget, neighbors...), can make music otherwise impossible to do.

This could be a new topic but we also can discuss about what's better: a low end acoustic upright or a top line digital piano. In my humble opinion, my Roland piano has a nicer tone and has a better action than my Yamaha B1 upright. It feels in some way more than a grand. It is also true that the B1 is more "authentic" (whatever it means) but if I could have only one of them in a deserted island, I'd probably choose the Roland (if it were somewhere to plug it). Of course, in my opinion still today digitals cannot beat in any way acoustic grands.

By the way, I'd love to try those Avantgrand "hybrid pianos". I'm really curious about them. If a great artist as Katsaris and others are speaking very well of them there must be a good reason, apart of the money Yamaha paid them... ;-)

I suppose its possible that in the future as digitals become closer in sound and touch to acoustic pianos there will come a time when people who have never owned an acoustic before will think the two sound and feel so similar that it doesn't make a difference. In the showroom they may not find they prefer one over the other, but I don't think people who have always played acoustics will ever not be able to tell the difference when they sit down and play. Maybe not in recordings, or live but at the other end of a room or in a concert hall, but not sitting at the piano playing it. There will always be some difference. The question will just be whether or not the difference matters. I would not buy a digital unless I had no other choice. To me the instrument itself is a thing of beauty that a circuit board just can't compete with. I suppose they could start making holographic "insides" for digitals so that it could look real while you're playing it...

If I have the funding, the room, and the chops to properly enjoy such a thing, yes. But until then, my digital suits my needs quite well. And let's say I do get a "real" piano -- would that mean I get rid of my digital? Not very likely at all, unless I replace it with a newer model.

The "real" piano isn't going away anytime soon. Some people still make and use fountain pens. Some people still make and use mechanical wind-up watches. Others insist on having their personal stationary be run off on letterpress. Some find exquisite pleasure in watching an image come out of nothingness, in the orange light of the darkroom.

Still others, with very deep pockets, insist on having a proper, Baroque-style, mechanical pipe organ. And because of that, there are still folks who make new ones, and service the old ones.

The piano's not dead. The number made may decrease, as digitals become more and more ubiquitous than they already are -- but the piano won't go away. Ever.

I mean.. people still make harpsichords and fortepianos. I thought the modern piano had made both redundant 120 years ago..

Another point that could be made is that ultra high-end audio speakers are very expensive, and have been very expensive for a long time. I'm not familiar with them, but I haven't heard that there is any reason to think that the price for speakers like that is going to down drastically in the future.

From this, one could conclude that even if the technology for a digital instrument arrives at a level where one can't tell the difference in a blindfold test (hard to imagine, but let's say it happens), the digital piano will be much more expensive than a comparable acoustic piano.

Another point that hasn't been mentioned (at least in this thread) is that some repertoire absolutely requires an acoustic piano. There was a recent thread full of videos of this kind of repertoire -- which calls for plucking the strings, hitting the strings, hitting the cast iron plate, etc.

In recent music, this isn't all that uncommon. It's not just a novelty, but is an established part of the artists' palette; listen to what Esbjoern Svensson did by putting a piece of paper between the hammers and the strings of a vertical, for example. Or listen to what many artists are able to do by using their hands to dampen the bass strings in innovative ways.