Piano World Home Page Forums Our Most Popular Forums Digital Pianos - Electronic Pianos - Synths & Keyboards How close are we to digital beating acoustic pianos?

Pianos are actually being fully modeled with technology like Pianoteq. It's a bit disappointing they haven't put this technology in more mobile stage instruments.

I think that this is an excellent point. Being able to mess about on a keyboard with headphones on at 1 in the morning without worrying about family, housemates or neighbors, low purchase price and maintenance costs, portability, little space taken up in the house, fun of playing around with other voices. I know digitals initially got me playing again!

I wasn't satisfied after a while though and did buy an acoustic, and I intend to keep buying digitals as new, improved ones come out as digital has benefits for practicing over acoustic.

I'm a PhD candidate on electronics. I’d like to say some technical stuff.

What we hear are analog sound waves. In theory, sound recorded on analog tapes ALWAYS contains more details than digitally recorded sound, because sampling and representing each sample’s amplitude by finite bits, 16/24/32, introduce distortion. However, as sampling rate and bit-width increase, distortion can be undetectable by human ears. Therefore, in practice, DPs will replace acoustics bit by bit, just like digital camera replaces film camera. But high end acoustic and film camera will always exist.

Digital recording can achieve high fidelity, as long as the memory is sufficient. The bottleneck is more at the playback. It is like this: the speaker generates sound by vibrating a surface at certain amplitude and frequency. Even if a versatile speaker can generate each individual note accurately, when multiple notes sound simultaneously, the vibrations of different frequencies and amplitudes on the same surface will interfere with each other. The notes will not be ‘pure’ any more. Possible solution is using 88 speakers, one for each key. In fact, strings of acoustic do interfere with each other. But the interference is hard to emulate.

At the end of the day, when playing a digital you'll always be listening to the sound of a recorded piano and that sound will not have the richness of an acoustic.

You can increase the satisfaction of a digital by listening thru studio-quality headphones, such as Sennheiser's HD-580's or their newer 600's, and by buying third party samples or pianoteq's modelled piano (which to my ears sounds synthetic when chords are sustained), but you'll still be listening to recorded sounds.

Digitals do allow you to expand your piano universe. I currently play Art Vista's sampled Steinway B. Proaudiovault sells a sampled, company authorized Bluther Model One and Gary Garritan will soon release a company authorized Steinway D. The Garritan Steinway will included samples which can be triggered by an una corda (soft) pedal.

Playing any of these requires connecting your 'board to a computer. It's not complicated, but it's something to consider.

If you're considering a digital because of frustration re finding an acoustic that's within your budget, it might be worthwhile to make a piano shopping trip to a city which has a number of dealers.

Great explanation, ge_lw. That is exactly what I wanted to find out.

For the record, if I had the means, I probably always will prefer to have an acoustic concert grand than even the finest digital.

It's an emotional thing.

Reaper978,
I second your sentiments very much.
While I (wrongly or not) largely dismiss uprights considering current advances of DPs, there is a lot of sentimental value and charm in grands plus the sound!!
So far, I did not have a chance to test a DP imitating/approximating grands fully satisfactory to me.

this is the digital forum, so I've got to root for digital. That's why i'm in this forum

We all know the answers.
you want pure piano sound? You'll always need an acoustic.
you want more than that? You'll always need a synth.

Beauty is in the eye of the beholder.
I like synths better.
Pianos are beautiful.
But a piano isn't enough for me.

Someday if I ever get rich, I'll cram an acoustic into the livingroom (somehow!).
But my first love will always be the synth.

Different people perceive the same sound differently. The ultimate DP should be as this:

An individual strikes each key of a desired acoustic at different velocity, duration, etc. Record the corresponding electric signals his/her ears transmit to the brain through nerves. This profile defines the acoustic perceptual space of that individual. Based on this profile, adjust a DP, so that the DP arouses the same nerve signals when it is played at the same velocity/duration as on the acoustic. After the profile extraction and the DP adjustment, the DP will be the same as the acoustic for that individual.

The reason why some people like a DP while others don’t is because they have different acoustic perceptual profiles. DP should be adjusted to correct the profile difference.

That's right.

Now tell us the differences between how an acoustic piano projects it sound waves and how electronic speakers directionally project their sound waves. We will need a large number of computer controlled speakers to simulate the acoustics of a grand piano. It's not economically practical for home users in our lifetime.

How will they accurately manage to produce the hundreds of millions of combinations of potential sympathetic vibrations that modulate the waveforms when different combinations of notes are played together on a real piano? And factor in that the varying dynamic levels of every note within a group of notes (vertical or linear) produces different waveforms. Millions of possible interactions occur. As Mr. Spock would say these digital pianos are like trying "to construct a mnemonic circuit using stone knives and bearskins." It's like trying to get a robot to have feelings. It's way off in the distant future. we'll all be gone by that century.
It's not economically practical in our lifetimes.

Digital pianos are sort of like photographs, they capture a note in still life, it's very artificial, there is no felt hammer hitting strings strung across a metal bridge and a wooden sound board, AND the notes do not interact with each other the way they do in the acoustic world.

Nice analogy, let's develop it further from a slightly different perspective:
Sure silver photography offers greater resolution, dynamic range, colour reproduction, etc
No silicon detector arrays are there yet to "beat" it.
Yet, when was the last time you have used it instead of a digital camera?
Why? First of all - simplicity of use, price of the pictures and largely because digital cameras reproduced the image "pretty much good enough" for our eyes for 4x6, 5X7 and recently even larger picture sizes. (Note, not perfect but just sufficient for >99% consumers out there)

Surely, the sound of digital is not the same (and never will be the same) as of a nicely engineered acoustic piano.
Yet, what matters of all this rich and beautiful sound, generated near magically inside a resonator of a musical instrument, is only its tiny part which reaches our ears. The ears are definitely a pretty sophisticated sensing device but still one that is abiding by basic physical principles and having its clear limitations in terms of frequency and amplitude resolution.
So, very importantly, the goal of sound reproduction (similar to photography) is greatly simplified to deliver "good enough for our ears".
That is obviously why high quality headphones capture more and more ardent funs despite being just puny silly coils.
Furthermore, take the surround sound: by superimposing all the sound components right, we may hear the richnesses of the sound not attainable even by "putting you head inside the grand" (pardon my rude analogy).

Now returning again to the digital photography: about 10 years ago, several megapixel arrays seemed hardly a reality for consumer electronic. Yet here we are!
Now, there is NOTHING limiting superb sound reproduction by digital technology from physical, engineering and, I am sure with some time, also from financial perspectives.
32, 64, 128, etc mass produced tiny coils and diaphragms are not that costly after all! (Think about those ten million pixels, they are not produced by manual labour no matter how cheap it can be in some part of the World )

All in all, I am personally convinced that DPs will put upright acoustic pianos into oblivion.
Grands would hopefully remain nice niche musical instrument (similar to silver photography nowadays). I would certainly love to see grands around (after all some magic should remain in our robotic technological age) and one day possibly having one alongside with our DP.

Another analogy that I like is photograph (digital piano), vs painting (acoustic).

I like the analogy too. For slightly different reasons. Film photos and digital photos are both recordings. Neither one can compete with being there with the real thing. Same for movies and/or videos. In fact, if they can't get the sound right, they won't stand a chance for that either. That's why we have the concept of the willing suspension of disbelief.

No matter how adequate sound systems might get on basic specs like clean, flat, and natural frequency, dynamic, and/or impulse response, there's still no way existing sound systems can duplicate what you hear when you listen to a real piano. Just listen to how the sound changes as you walk into a room where a piano is being played. Or even when you sit down and turn your head ever so slightly. Then throw on a recording and do the same thing. The best you can hope for is to get a similar effect from one fixed location in the room. And that won't even happen. Because the sound from a grand piano sources from so many locations, starting with its array of 88 notes (composed of many double and triple strings), it's sound board with its gigantic surface area (compared to that of speakers), not to mention the cabinet and lid which dwarf most speaker systems. Certainly a puny little 2-speaker sound system isn't going to emulate that with any accuracy.

I doubt even a Cray running the most advanced convolution program in existence would be able to model that many sound sources in real time. Even if you put an independently wired 88-speaker linear array into an appropriate wooden case to play the output. Didn't someone here say something about a doctoral dissertation? Might I suggest they do it on analysis of a playback system consisting of 4 specially wired Bose PAS's inserted into a grand piano case... that could yield a 96-source array. Add a couple sub-woofers and maybe you could call it a Bose Imperial.

But if you're talking apples to apples... comparing recordings of real pianos with recordings of sampled and/or synthesized pianos, that's a totally different question. My personal opinion is that sampled piano recordings can compete quite effectively with acoustic piano recordings. If only for the practical reason that it's easier to accurately record each individual note and impulse response in a controlled environment. An environment you're not likely to tolerate for a live performance. And you can re-take for perfection till the cows come home. And samples are in fact just recordings anyway.

The only catch is that the performer and audience won't get the same feedback from a sampled piano performance. And the performance may be impacted as a result. But if you throw a tft recording strip into a real grand piano...

Howard

hv, definitely agree re comparing recordings of sampled and acoustic pianos. In an earlier thread, someone posted that David Neveu had been surprised to learn that a recording submitted to his site had been done with samples - Ivory, I think.

Tuan Vo, acoustics do not all sound alike. Their differences are a source of the now-standard advice on the acoustic forum that people should play as many pianos as they can, until they find one that sings to them.

Most piano stores welcome people playing, especially early in the week when the store is not busy. You could have an enjoyable day by visiting shops and listening to pianos' different personalities.

Including brands such as Knabe, Brodmann, Bohemia, Kawai, and Mason & Hamlin in your auditioning will help you to experience the broad range of tonal palettes available in the acoustic world and how much richer is the sound of an acoustic.

You'll also experience the sound of a piano in a completely different way. Samples gives us an audience perspective. Sitting at the keyboard of an acoustic will give you a player's perspective. (Some years back Vintaudio included a player perspective preset in its C-7 samples.) Could be fun.

You could do a short course re tonal palettes by comparing Art Vista's Virtual Grand - the one I play; it's a sampled Steinway B - Bardstown Audio's Bosendorfer Imperial, and Proaudiovault's company-authorized Bluthner Model One concert grand. Samples are on the sites.

It's more like NO maintenance for digitals (which for me is a big plus), studio quality recording is a breeze, and no problems switching back and forth with acoustics. I'm running over 8 different high end sampled pianos on the PC from my digital (the entire package for about 1200 bucks, not counting the PC say another 1500). I'd need half a million dollars if I wanted to get the real things

Do I still want a grand piano, yes, but I think the improvement would be marginal and hard to justify the price difference. I probably will upgrade in the next 5 years, but only because I know digitals are still evolving and a much better one will be out by then.

sid, over the last four years, I've had five different sampled pianos and my next buy will be Proaudiovault's Bluthner. Much as I love listening to the samples thru Senn's studio-quality phones, none of the samples provides the richness and complexity of an acoustic.

Upgrading to an acoustic would provide much more than a marginal improvement. Now if the issue is sonic bang for the buck, this Budgeteer agrees with you;-)

When you strike two notes on a real piano their waveforms modulate each other in complex ways. This does not happen in a digital piano.

From Wiki:

"String resonance occurs on string instruments.

Strings or parts of strings may resonate at their fundamental or overtone frequencies when other strings are sounded. For example, an A string at 440 Hz will cause an E string at 330 Hz to resonate, because they share an overtone of 1320 Hz (3rd overtone of A and 4th overtone of E)."


That was a simple example.

Now consider all the different combinations of notes that get played on a piano and the different dynamic ratios between those notes. The dynamic level of each note, or its vibrating "energy" level, effects the vibration of other strings in different ways. The result is a very large number of different waveform modulations, potentially millions more than chips can process. And every acoustic piano has different characteristics of resonance and will behave in its own unique tonal way. String resonance is only crudely mimicked on digital pianos with an artificial "on or off" sort of approach..

rintincop, wonderful post.

Both Kawai and Roland (the very least I know) offer user-adjustable string resonance for quite some time! Actually, that would be one of the most suitable for digital simulation effect, since, as you exemplified nicely, resonances are due to the overtones, which are very straightforward to calculate and emulate with sufficient variations.