First Virtual Piano Competition
bluegreenone writes: "The New York Times has an article on what may be the first 'virtual' piano competition. One of the judges for the contest being held in St. Paul will actually be in Japan. He will evaluate the performances as relayed by Yamaha's Disklavier system. This has some interest from a technical standpoint, and also raises new questions about what a "live" performance is."
There are a few things that might not be captured, such as the speed at which a damper is replaced on the strings when a note or the pedal is released. I don't know if the system accurately reproduces this. Certainly MIDI wouldn't.
I suspect the main difference between a live performance and a performance from a Disklavier is that a live performer is constantly adjusting his touch to account for the individual characteristics of the piano, whereas IFAIK the Disklavier system does not have this feedback loop.
So it would only sound exactly the same if the piano used to play back the performance was identical in touch and tone to the piano that recorded the performance.
Mind you, I doubt if I or many others would notice the difference.
In the case of the other piano, elsewhere in the world, there will be subtle differences in the instrument itself. Even if you can (and this is arguable) reproduce 99% of the nuance of the actual key velocity, you're not taking into account the fact that the musician is feeling and reacting to a separate instrument with a key action which must be different, affecting his touch, and with possible harmonic and amplitude differences across the piano keyboard range even given the same key velocity.
But first and foremost this is most certainly NOT live music. It's reproduced mechanically, and that is no different from playing a CD (reproduced optically). Just because it's a real piano you're hearing, it's not the same piano the artist was playing. And he/she's not even there.
Conversion Rate Optimisation French / English consultant
But the pianist will be responding to the touch and feel of the piano (s)he's actually playing: I know, I'm a pianist. So unless they have an *identical* piano, in *identical* humidity, etc (which is impossible, given the subtleties a really good piano contains), they can't possibly have a 100% accurate reproduction.
The question is, is it accurate *enough* for this purpose? I would claim "no", but I've never seen the system in action.
- Oliver
The right to bear arms is only slightly less stupid than the right to arm bears...
...in Player Piano, in 1955 or so. A recording, no matter how faithful, is a recording, it captures a performance, not a performer. The Disklavier is a player piano, a really good one. It makes the hammers hit the strings in (insert meaningless technical quibble here) the same way the pianist's fingers did. But until it captures Glenn Gould's humming, Stevie Wonder's head-bobbing, or Tori Amos' (essentially) fucking the piano stool, I will not be fooled into thinking that I am experiencing a live performance.
(Been playing since I was four, and I prefer Steinways to Yamahas, but that's another matter.)
The early player pianos were simple mechanisms. There was no loud and soft controls other than the pedals, so the only way of varying the intensity of the sound was by playing the notes more often. You could not repeat notes too quickly or the roll might tear along the dotted lines, so the players used an octave tremolo style that gave these performances a very distinctive sound. Plus, the machines used to live in bars, so the tuning was sometimes rough, and beer got spilled inside.
Forget them. The Ampico series B used to have 16 levels of force behind the hammers, with separate settings for the 'left hand' and 'right hand' (not individual key control, but not bad for the time). The speed of the hammers was recorded using the spark-gap timing techniques used for measuring bullet velocities, a spin-off from the armament industry for WW1. Stick a roll in one of these beasts, and close your eyes, and it's just like being at a performance. Even a CD player and hedphones has trouble sounding this good. The downside was they cost a few thousand pounds, which in its day would buy you a street of houses.
Recording was not fully automatic. People needed to exercise judgement over how to convert things like the key velocities into the 16 pressure settings. There were also some sequences of rapid notes that could not be reproduced accurately. However, they could play the roll and log the timings, and edit it until the timings got as close as possible to the original performance.
So, is it live? Well, back then they decided there was no risk of duff notes, and you don't have the actual performer present, so it was definately not live, but in some respects it was better. Same would be true today, I guess.
Beware, I'm a jazz musician, so I hold no credibility among people who work for a living...
The claim that 127 bits is enough, or that any digital device 5000 miles away can qualify as a live performance is pure bunk, and Yamaha is notorious for this kind of garbage (Am I the only one who remember's their jazz band full of WX-7 wind synths that they said would be "revolutionary?")
Simply put, the energy one puts forth when playing is not there when a computer is shoving down the hammers. I will admit:
1: 127 bits will get a pretty good velocity vector for the hammers. I'm sure whatever checks they have to determine dampers coming back on,etc are sufficient to not make it sound comical.
2: From a technical standpoint, it's a great achievement to do what Yamaha has done. It is really leaps and bounds ahead of most things out there.
BUT
that being said, where's the energy? where's the breath of life that you put into the instrument every time you play. Where's Vladmir Horowitz playing a sold out concert in moscow looking like he's calmy sitting and waiting for a bus while lambasting an opressive communist regime through the music? Where's keith jarret groaning and Philly Joe Jones responding when he belts out a solo? Allow me to indulge in an Anecdote (Courtesy of Kenny Werner's excellent book, Effortless Mastery)
"I went to Bill Evans' 50th birthday party. So many pianists were in attendance, it looked like a dictators convention. Many people played for Bill, at a piano that will remain nameless. This brand of piano has a tendency to sound bright (pop-ish is the easiest def. i can give... Paul simons electric pianos are bright... most acoustic jazz stuff (herbie hancock...) is not). All the pianists who played said piano sounded that way. Then Bill sat down to play, and he sounded dark, rich and full, on the exact same piano. Looking at his hands, the wrists were like shock absorbers. when he "dropped his fingers" (Dont worry about the def. unless you play), he had a special way of accelerating them so full yet rich force was achieved, so his whole arm / hand weight would keep the hammers where they needed to be."
Now, does the disklavier have that enrgy, that intensity? I don't think so. The point is that it's not a digital thing, playing an instrument. Trying to quantize "Soul" of music is counterproductive, and although being able to reproduce sounds in the way yamaha has been working is a great step, calling it a "live performance," and having a competition where the MIDI (sorry, disklavier...) interface records the velocities (Even if it is not recorded sound, in a way, it is a recording), is not under any definiton a live performance.
No really good-quality audio recordings of his best work exist, he was born too soon. But we do have digital recordings: piano rolls he cut for player piano while he was at his prime. Five years ago or so, some smart people found some of those original piano rolls, scanned them into the computer, and converted them to MIDI files. Any adulterated roll-holes that the publisher might have added were removed -- at the time, player-piano publishers often took a razor blade and cut extra holes to make it sound like their artists had more hands. And subtle dynamic variations were added by hand to each note, since a player-piano roll has only one note attack volume (which at the time was often crudely modulated on playback anyway). As the liner notes say: "Converting Morton's old 78 recordings to computer data, we were able to study them from myriad standpoints of tempo, melodic shaping, accentuation, swung rhythms, chord voicing, and pedaling."
Then they played the files on a Disklavier in a concert hall.
It's eerie to listen to. It's this guy who was born in 1885, actual recordings he made in 1920, and it sounds... brilliant. You're not used to hearing jazz pianists from that era in CD quality on a great piano. Suddenly you realize that the 1920s did not sound like the 1920s to the 1920s. It's like seeing photographs a hundred years ago in color -- your mind knows better, but your senses go: whoa.