piano samples vs. real piano

[sign]

Originally posted by Raymar:

Can't do pitch bends with a real piano. When is somebody going to get around to implementing that feature?

 

steve

What's the use of pitch bending a piano?

 

Have you noticed the misery with pitch bending guitars by Kahler and other tremolo devices that were popular in the early nineties.

 

People were tuning their guitar most of the time and still is was out of tune.

 

Besides that, a grand piano has a tension of between 15.000 and 32.000 kilogram on the strings (Wewus, am I right?)

[Duddits]

These things we know:

 

Real pianos do things fake pianos don't.

Pianos are hard to mic.

Not every "keyboardist" is a piano player, but real piano players play better on real pianos.

Not every piano part in a multitrack mix works best with a real piano, but if it is a traditional or featured track, then real pianos are better than fake ones.

If the track is part of a complex mix, a minimal riff or color, or part of a dance or pop track, then a fake piano might be better than a real piano.

 

Thus:

 

Great piano, great mic, real piano player, featured piano part = use real piano.

 

Average piano, average mic, real piano player, featured piano part = use real piano.

 

Average piano, bad mic, real piano player, featured piano part = go either way.

 

Average piano, average mic, "keyboard player", featured piano part = go either way.

 

Average piano, average mic, real piano player, piano part one of many = use fake piano.

 

Average piano, average mic, "keyboard player," piano part only one of many = use fake piano.

[EmptinesOf Youth]

when all else fails bust out the mellotron...we have our tapes play with people screaming

[fantasticsound]

First of all, I need to reiterate that I enjoy using keyboards, including piano samples. But that doesn't make it the same thing. Like you said, Miro, if you can record a pianist on the real thing, it's definitely worth considering.

 

However...

 

Originally posted by miroslav:

...One thing to remember when you argue "sampled piano" vs. "recording a real piano"...

...a sampled pianoISa recording of a real piano!...

This is analagous to saying a video of a beach in Hawaii, a Spring-time mountain meadow in the Rockies or the Sahara desert will give you an accurate perception of what it is to be in those places.

 

As previously mentioned, the "recording a real piano" has to ignore many of the interactions that occur when 200 or so strings are physically connected to the same sound board even though only a few are deliberately activated at once. The two are absolutely not equivalent, nor is the experience of watching video, no matter how big, of an environment compares to being in the sun, wind, space and smells of that environment.

 

jmitch - The action of the controller is only a small (yet important) part of the experience that is playing a real piano. The environment of a piano includes space. That is something tha can be approximated by samples and reverbs, but my experience listening to high quality recordings of acoustic instruments tells me there is still a wide gap between a tweaked sample track and a great acoustic recording.

 

For all our technological magnificence, great recording engineers still hold Belafonte at Carnegie up as one of the greatest recordings of all time. IIRC, it was recorded with 3 microphones. Absolutely incredible what can be done with a great acoustic space, fantastic musicians, and a few microphones.

 

And, I think it goes without saying that nobody expects a piano that is going to be buried in a mix to be recorded (rather than a sample) unless the piano and mics are readily available and the player requests it.

 

Oddly enough, I had a keyboard player (who is a pianist) lament that he needs a controller with keyboard (rather than piano) action for his organ samples. Anyone who plays organ knows that everything that makes a piano's action part of the experience is not what's expected or needed for organ.

[koolkid]

i defy anyone here to tell the difference between a recording of a steinway and a recording of a roland fantom or a korg triton extreme grand piano sample with the same room sounds

[MusicWorkz]

It's funny for me. In a live situation, a good ampled piano will cut thru the clutter without all the hassle. In a recording situation, on those songs where the piano is NOT the main focus and is really playing a well-mixed supporting role, a good sample USUALLY works.

 

When the music requires the piano to really carry the tune, however, NOTHING irks me more than hearing a sampled piano, no matter how good, trying to pass. The attention to that kind of detail will make or break even the best written song.

 

If you want the track to feature a piano, go out and record a good REAL piano. If you do the preproduction right, you won't have to spend thrt much to get it done properly.

[MusicWorkz]

Originally posted by koolkid:

i defy anyone here to tell the difference between a recording of a steinway and a recording of a roland fantom or a korg triton extreme grand piano sample with the same room sounds

That would be EASY to do. A real piano will always have sympathetic resonance of the other strings and sound board as well as an articulation you will NEVER be able to get on a Triton or Motif (I have both in my studio and can tell the difference even between the two with my eyes closed and one ear covered...)

[jmitch]

Originally posted by Duddits:

These things we know:

 

Real pianos do things fake pianos don't.

Pianos are hard to mic.

Not every "keyboardist" is a piano player, but real piano players play better on real pianos.

Not every piano part in a multitrack mix works best with a real piano, but if it is a traditional or featured track, then real pianos are better than fake ones.

If the track is part of a complex mix, a minimal riff or color, or part of a dance or pop track, then a fake piano might be better than a real piano.

 

Thus:

 

Great piano, great mic, real piano player, featured piano part = use real piano.

 

Average piano, average mic, real piano player, featured piano part = use real piano.

 

Average piano, bad mic, real piano player, featured piano part = go either way.

 

Average piano, average mic, "keyboard player", featured piano part = go either way.

 

Average piano, average mic, real piano player, piano part one of many = use fake piano.

 

Average piano, average mic, "keyboard player," piano part only one of many = use fake piano.

I don't think it's that simple and straightforward. Period.

[fantasticsound]

Why do you disagree with Duddits, jmitch?

 

I think Duds hit the head on the nail with that list. Which part do you disagree with? The funny thing is this set of guidelines works in any situation. It even leaves room for either a piano or sample when certain aspects tend towards the real thing and others suggest a sample would be adequate.

[jmitch]

Originally posted by fantasticsound:

Why do you disagree with Duddits, jmitch

I don't disagree. I never said that. I just think it's not that simple.

[Tedly Nightshade]

OK, so when electric pianos first came out, some jazz fans noted, you could tell all the players apart easily acoustically, but they all sound the same on the electric piano.

 

Now we can tell some distinctive players by their touch on the electric piano- they are electric piano players, possibly as well as acoustic piano players, but they have a whole different bag on the electric piano. Herbie Hancock comes to mind.

 

So now we have sampled pianos. I think a lot of my dissatisfaction with those is simply that I have a piano player on hand whose specific needs and voice I am trying to complement. Her touch is unmistakeable- a lot of "personal tone", as she puts it, and she can pull some personal tone out of a casio. And she has a specific repetoire (always expanding) that she wants to play. So, she'll play it on an instrument and it will or won't work especially well. Sometimes it's magic.

 

And some real pianos work far better for her purposes than others. We have not found anything to beat the 1928 upright we have here. 9' American Steinway is the best contender, but ultimately it just does not have the "age of jazz" sound, and it's kinda spendy!

 

She can pull her tone out of the rather musical and not-completely-pretending-to-be-a-real-piano Kawai MP9000, but it's been some years- not really that great.

 

So I think it has a lot to do with what you want the piano or fake piano to do. I'm thinking, convey the musical personality and vision of a performer- that may be an assumption that doesn't really apply in all cases! But if it does, take a look- how effectively can the performer get the music across?

[koolkid]

Originally posted by MusicWorkz:

Originally posted by koolkid:

i defy anyone here to tell the difference between a recording of a steinway and a recording of a roland fantom or a korg triton extreme grand piano sample with the same room sounds

That would be EASY to do. A real piano will always have sympathetic resonance of the other strings and sound board as well as an articulation you will NEVER be able to get on a Triton or Motif (I have both in my studio and can tell the difference even between the two with my eyes closed and one ear covered...)

i didnt say a motif, i mentioned a triton extreme or a roland phantom,both have new digital samples that defy recognition as a sample

 

there is really no difference ((from a physics standpoint)) between a digital recording of a steinway and a digital sample of a steinway

 

i understand the whole ' i can hear the overtones at 900 hz when you play a cmaj7th and its not right' thing,but thats just not feasible... unless you have a monster ear,which most people do not

 

i will have to email you some samples and put you to the test

[fantasticsound]

Originally posted by koolkid:

...i spoke to the rep at roland and they supposedly took that all into account when they did the fantom samples

 

i will have to email you some samples and put you to the test

Who was the rep? I want to hear their explanation. Despite innovations in digital recording, in order to take these varying sympathetic vibrations into account across the keyboard would take several hundred meg of sounds, by my estimation. Remember, those resonances are different when different chords are played and when similar chords are played in different registers. I just want their explanation from a Roland rep rather than second hand.

[jmitch]

Originally posted by Ted Nightshade:

...take a look- how effectively can the performer get the music across?

This is definitely something I agree with. Some performances may require a real piano to get the music across, whereas others may require a certain sound only achieved by sampled pianos. How many different pianos does a producer usually have at their disposal. Maybe two or three at most. But pianos are hugely expensive instruments. That's where I think samples really come into play. If you need that certain sound, that certain tone. Kind of like guitar distortion. There's so much variation and subtlety to the kind of tone a guitar can produce when you bring distortion into the picture. It's the same thing with pianos. The tone variation is so wide, there's no way you can get it all in acoustic pianos. That's where samples come into play.

 

Also, I'll take a great sample anyday over a shitty sounding piano. The only reason I would ever take a real piano over a quality sample of a multi-thousand dollar instrument, miked by professionals is if it was a really, really good. -like a new Steinway Grand. But, you don't get much opportunity to have access to those kind of instruments. Samples save the day.

[koolkid]

Originally posted by fantasticsound:

Originally posted by koolkid:

...i spoke to the rep at roland and they supposedly took that all into account when they did the fantom samples

 

i will have to email you some samples and put you to the test

Who was the rep? I want to hear their explanation. Despite innovations in digital recording, in order to take these varying sympathetic vibrations into account across the keyboard would take several hundred meg of sounds, by my estimation. Remember, those resonances are different when different chords are played and when similar chords are played in different registers. I just want their explanation from a Roland rep rather than second hand.

scott,he would come in every week while i was working at guitar center,forgot his last name

 

great demonstration on the fantom though,phenomenal keyboard

[Duddits]

KOOLKID:

 

As you know from my posts, my position is not "use a real piano for everything, no matter what." Sometimes I use a real one, sometimes I use a fake one. Both have their place.

 

While it may be debatable when to use a real piano and when to use a fake one, if there's one thing that's not debatable (and every kool kid should know this), it's DON'T BELIEVE THE COMPANY REP! If anyone has an agenda, it's them. No matter what they're selling, acording to them it's better, more affordable, hipper and real-er than anything else on the market!

 

Piano emulation has a rich history, and every year or so it seems like another piano-emulation technology is billed as "this is the one that finally will convince you it is the real thing." But what real pianos do is far more complex than anything that has yet to emulate them. There are a tremendous number of simultaneous processes -- physical, mechanical, accoustic, psychoacoustic -- that transpire every time you play a note on a real piano. Fake pianos have a come a long way, but they are a far cry from the real ones, no matter what the Roland rep may say.

 

"Don't .... Don't ... Don't believe the hype...." (Public Enemy)

 

I'm guessing that the action alone -- just the action -- on my piano probably costs around $15,000 to build. That's just the little hammer thingies under the strings that transfer the notes I play on the keyboard to the strings that are hit on the harp. I don't think the action on the Roland synthesizer (and I have one of those too, not the Fantom though), is quite the same. And that's just the mechanical, not sound-generating portion of the piano.

 

As far as the sound itself is concerned, a number of people have brought up "sympathetic vibrations." Do you know what that is? It is the complex mechanism through which the direct ringing of one piano string or strings (what you actually play) sets into motion the indirect ringing of other strings. There have been a few attempts to emulate this through samples, but none has come anywhere close to the real thing.

 

Every time you play something on the piano, there are millions of different combinations of sympathetic vibrations (not exagerating), a number that is further increased by the use of pedal (which is not an on-off choice, but has multiple steps in between). Futhermore, the indirect ringing of a string is not simply a lower-volume of its direct ringing. It's an entirely different sound, with different upper harmonics set into motion in a variety of ways based on the residual vibrations it receives from the directly played string(s).

 

Sympathetic vibrations are one of the main ways in which a pianist can create the illusion of legato (and it is an illusion). As compared to a trumpet, violin, or voice, in which a melodic phrase moves smoothly from note to note in a single continuous stream of air or on a single string, a piano is a set of separate sounds, seperate notes, seperate strings, seperate buttons. Pianists create the illusion that the notes are connected, as part of a legato phrase, melody, chord, or sonority, largely through sympathetic vibrations -- enhanced through the use of pedal (which enables "sympathetic strings" to more fully vibrate and for longer amounts of time).

 

Whether you're talking about a $60,000 piano or a $600 piano, the phenomenon of sympathetic vibrations that occurs in the real thing would require a supercomputer and software that has yet to be written in order to emulate it through what would be an unprecedented library of samples.

 

I'm sure the Fantom is a phenomenol keyboard, and I'm sure the Roland rep is a great salesperson. But ...

 

"Dont .... Dont ... Don't believe the hype....

Dont .... Dont ... Don't believe the hype...."

[koolkid]

Originally posted by Duddits:

KOOLKID:

 

As you know from my posts, my position is not "use a real piano for everything, no matter what." Sometimes I use a real one, sometimes I use a fake one. Both have their place.

 

While it may be debatable when to use a real piano and when to use a fake one, if there's one thing that's not debatable (and every kool kid should know this), it's DON'T BELIEVE THE COMPANY REP! If anyone has an agenda, it's them. No matter what they're selling, acording to them it's better, more affordable, hipper and real-er than anything else on the market!

 

Piano emulation has a rich history, and every year or so it seems like another piano-emulation technology is billed as "this is the one that finally will convince you it is the real thing." But what real pianos do is far more complex than anything that has yet to emulate them. There are a tremendous number of simultaneous processes -- physical, mechanical, accoustic, psychoacoustic -- that transpire every time you play a note on a real piano. Fake pianos have a come a long way, but they are a far cry from the real ones, no matter what the Roland rep may say.

 

"Don't .... Don't ... Don't believe the hype...." (Public Enemy)

 

I'm guessing that the action alone -- just the action -- on my piano probably costs around $15,000 to build. That's just the little hammer thingies under the strings that transfer the notes I play on the keyboard to the strings that are hit on the harp. I don't think the action on the Roland synthesizer (and I have one of those too, not the Fantom though), is quite the same. And that's just the mechanical, not sound-generating portion of the piano.

 

As far as the sound itself is concerned, a number of people have brought up "sympathetic vibrations." Do you know what that is? It is the complex mechanism through which the direct ringing of one piano string or strings (what you actually play) sets into motion the indirect ringing of other strings. There have been a few attempts to emulate this through samples, but none has come anywhere close to the real thing.

 

Every time you play something on the piano, there are millions of different combinations of sympathetic vibrations (not exagerating), a number that is further increased by the use of pedal (which is not an on-off choice, but has multiple steps in between). Futhermore, the indirect ringing of a string is not simply a lower-volume of its direct ringing. It's an entirely different sound, with different upper harmonics set into motion in a variety of ways based on the residual vibrations it receives from the directly played string(s).

 

Sympathetic vibrations are one of the main ways in which a pianist can create the illusion of legato (and it is an illusion). As compared to a trumpet, violin, or voice, in which a melodic phrase moves smoothly from note to note in a single continuous stream of air or on a single string, a piano is a set of separate sounds, seperate notes, seperate strings, seperate buttons. Pianists create the illusion that the notes are connected, as part of a legato phrase, melody, chord, or sonority, largely through sympathetic vibrations -- enhanced through the use of pedal (which enables "sympathetic strings" to more fully vibrate and for longer amounts of time).

 

Whether you're talking about a $60,000 piano or a $600 piano, the phenomenon of sympathetic vibrations that occurs in the real thing would require a supercomputer and software that has yet to be written in order to emulate it through what would be an unprecedented library of samples.

 

I'm sure the Fantom is a phenomenol keyboard, and I'm sure the Roland rep is a great salesperson. But ...

 

"Dont .... Dont ... Don't believe the hype....

Dont .... Dont ... Don't believe the hype...."

i actually feel to the contrary,it will take a super computer to hear the sympathetic vibrations that the human ear simply cant distinguish

 

as much as we would like to believe we have this great ability to distinguish overtones,its simply not the case from a scientific standpoint,most samples nowadays can have more tonal parts than a human ear can distinguish

 

combine that with the fact that when you record it digitally you are going to 'sample ' the sound anyway .... i really dont see the need in the near future to have a steinway around to mic

[Duddits]

Kool Kid --

 

Try playing a few notes on a piano. Then, holding the pedal down, play the same few notes again. You won't need a supercomputer to hear a difference, but you'd need a supercomputer to be able to produce similar results with samples alone.

[jmitch]

This debate is a long debate, that will be debated for years to come. I think there will always be a place for natural acoustic instruments. But, the gap between original acoustic instruments, and electronic emulation of acoustic instruments is always shrinking. There will come a time where acoustic instruments will be a novelty I believe. The ongoing advancement of technology cannot be stopped and will never stop. So, in evaluation, it is just a matter of time before that gap is brought to a pinpoint.

[Duddits]

Jmitch -- Of course the increasing sophistication of electronics supplants accoustic instruments to some extent, but you are ignoring something fundamental that goes beyond a discussion of "what should I use to record on the cheap?"

 

Humans, and even the occassional cat, have a relationship with the physical world around them, and a need to make sounds, clanking rocks together, blowing through shells, singing. Simulated music production can never replace that underlying need. People won't stop making music in the physical world simply because simulations become more and more believable.

 

The most expensive part of a synthesizer workstation has changed over the past two decades. It used to be the RAM, but now it's the keyboard itself, the physical part. This is why companies like Roland offer several models of the same synth at radically different prices, based on no more than the quality of the physical keyboard.

 

The mechanical sophistication of an average piano goes way beyond what is offered in the most sophisticated synthesizer, and serves a different purpose. The mechanical action of my piano is exquisite, and putting it into a synthesizer, even one that could realistically simulate sympathetic vibrations when that becomes available, would not decrease the cost -- 15k just for the action! Not to mention that speakers, as good as they have gotten, cannot yet reproduce reality by a long shot.

 

Again, this is not to say that "pianos are better than synthesizers" -- both have their place -- and obviously this discussion goes beyond the example that initiated it -- for what you seek to do, it may not make one bit of difference whether it's a real or fake piano.

 

But to look at the discussion in a broader sense, I don't think you can conclude that fake pianos are good enough at this point in time to replace real ones, and I'm pretty certain that it will be well over one hundred years into the future before the discussion becomes genuinely controversial!

 

If you're involved in producing house or techno music, there's a certain type of piano effect -- minor chords jumping around -- that can only be achieved on a synthesizer. The richness of a real piano makes it sound corny, and can't be quanitzed robotically, as required.

 

But if you're Ray Charles, Emanual Ax, or even a lowly piano-playing cat like myself, what you can do with a real piano is as far from a synthetic simulation as a caveman clanking rocks together or blowing through a conch is as far from a synthetic simulation.

[jmitch]

You misunderstood me Duddits. I wasn't saying that any "simulation" would replace human live play. I was saying that the instruments which humans would use to create the music would become more and more simulations and less the real acoustic counterpart, simply because of the technology advancement improving the similarties to the acoustic instrument. No where did I say humans would not longer play instruments and it would all be "simulation".

 

Like I said, there will always be a place for acoustic instruments and something about them that synthesizers can never touch. But, the fact that synths of all kinds of instruments are coming closer to the real thing cannot be denied.

[Hendmik]

I just had to chime in. Check out MP3 comparissons of many digital pianos and an original recording of a Steinway B here: http://www.purgatorycreek.com/pianocompare.html

 

Here's my feelings:

 

If I have my druthers: well tuned and maintained Steinway or Bosendorfer.

 

If the client can't afford it, or the piano isn't well maintained (you'd be surprised how many cracked soundboards are in Denver churches/schools), or if all they have is a Yamaha C7, I'll use Steinberg's The Grand.

 

If I'm playing with a live band: My Kurzweil PC88 - decent piano, great touch keyboard, only 50lbs.

 

If I'm seriously drunk and/or interested in making a fool of myself (I assure you that more often than you think), I'll break out the old Proteus 1.

 

All in all, I love a real piano. But if the owner refuses to properly care for it, there is really no point. Samples are always in tune and in most cases, can fool most listeners especially if other instruments are in the song, masking the sublties.

[jmitch]

Excellent listening test Hendmik. Thank you for directing us. I will take some tests and post my conclusions.

[MusicWorkz]

Originally posted by koolkid:

 

there is really no difference ((from a physics standpoint)) between a digital recording of a steinway and a digital sample of a steinway t[/QB]

I disagree.

 

It doesn't matter if it is Roland or Kurzweil, Yamaha or Korg. I have Fantom rack in my arsenal for a few weeks and it is not too much different. I know enough about its sound to recognize it already and know when to chose its piano over one of the others at my disposal, including my 400 MB sample of a Bosendorfer or a real yamaha C7.

 

When a piano is sampled, it is really just a snapshot of it at the moment. Once the snapshot is taken, that is it.

I did not mention a few other factors, humindity and temperature being to critical ones as well. That is the physics that cannot be sampled. That resonaance is more than just simple vibrations, but harmonics that no $3000 keyboard or even hot-rodded PC can capture and emulate.

 

Even still, there are a ton of factors, only a few of which I can even recall, that make the difference between recording a real piano and using a well sampled one.

 

There is a time and place for it, but when given the choice, at the end of the day, there is NO substitution.

[jmitch]

Okay, I sampled each and every last sample. First of all, the real Steinway is amazing. I was blown away right from the getgo. The samples immediately following were rather unimpressive- far below the musical capabilities of the real thing. However, as I got further down, there were a few that sounded quite similair to original Steinway. I was quite amazed at the realism many achieved.

 

My original perspective on this matter has not changed. Digital recreations will continue to come ever closer the real thing, and take more and more from it. Nevertheless, there will always be a place for real, acoustic, traditional instruments.

[MusicWorkz]

Oops...

[fantasticsound]

The Bardstown Imperial Grand samples pale in comparison to his recording of the real thing. And the recording, while done well, was a rather pedestrian, vanilla recording. Still, it was much richer and coherent than the samples. I'll have to listen to others, later.

[gtrmac]

I think that in some contexts a well recorded good acoustic piano is needed. A Jazz trio recording for example will benefit from the dynamic possibilities that a good player can exploit and the acoustic interaction of the strings on close harmonies is thrilling. As far as I know sampled pianos use crossfaded samples of only a few different key velocities and the real piano has an infinite variety of dynamics for each note.

 

On the other hand in the context of a pop song where the piano is in a mix of a lot of instruments and highly compressed the samples could be virtually indistinguishable from the real thing.

 

The point that the player may perform better on a real piano is certainly valid also.

[koolkid]

thank you hendmik for the excellent piano comparison page

 

i think we are pretty much in agreement that the average listener will not be able to distinguish between great samples and a miked steinway

 

especially in a digital recording situation where the steinway is going to be sampled at the same rate anyway

 

the sound samples you provided dramatically prove that point

[koolkid]

Originally posted by MusicWorkz:

Originally posted by koolkid:

 

there is really no difference ((from a physics standpoint)) between a digital recording of a steinway and a digital sample of a steinway t

I disagree.

 

It doesn't matter if it is Roland or Kurzweil, Yamaha or Korg. I have Fantom rack in my arsenal for a few weeks and it is not too much different. I know enough about its sound to recognize it already and know when to chose its piano over one of the others at my disposal, including my 400 MB sample of a Bosendorfer or a real yamaha C7.

 

When a piano is sampled, it is really just a snapshot of it at the moment. Once the snapshot is taken, that is it.

I did not mention a few other factors, humindity and temperature being to critical ones as well. That is the physics that cannot be sampled. That resonaance is more than just simple vibrations, but harmonics that no $3000 keyboard or even hot-rodded PC can capture and emulate.

 

Even still, there are a ton of factors, only a few of which I can even recall, that make the difference between recording a real piano and using a well sampled one.

 

There is a time and place for it, but when given the choice, at the end of the day, there is NO substitution.[/QB]

can you really hear the difference between the steinway and the 4front module in hendmiks page??

 

and if you do,do you think its of any significance ,,soundwise??

 

honestly??