MIDI-piano performances and Yamaha E-competitions

[johnlewisgrant]

All studio recordings of solo piano repertoire are EDITED at the studio, mixing and audio splicing (for better or worse).

Why, then, don't jazz, pop, and classical recording pianists go ALL THE WAY: record the entire piece at a MIDI-equipped Bos or CFX or (after-market MIDI- equipped) Steinway.  Edit the midi file, if necessary.  Then have the midi-player REPLAY THE EXACT PERFORMANCE, perhaps making a few changes in microphone placement and recording levels, as well?

The advantages seem pretty clear.  If the pianist loves everything about his recording, except say one wrong note, one phrase, or one movement, he doesn't have to risk losing the recording altogether in a re-record.  He simply changes that ONE note, and KEEPS the performance he likes so much.

Why isn't this the new approach?   Or perhaps it is?

[ranjit]

As to why pianists don't use MIDI, it's because there is a clear difference in audio quality if you compare MIDI, even with really good sound libraries, and an actual piano performance. Someone who is totally unfamiliar with the piano may not be able to tell the difference, but it won't fool anyone who has a decent ear for that kind of thing.

As to why it's surprisingly hard to simulate a piano properly... Imagine playing a note on a piano. The corresponding string vibrates, but in addition, the rest of the strings resonate along with it. You will then have to account for the frequencies and intensities of 100+ strings, to simulate the sound produced in the next instant. And all of this is for one note. A note played in the bass can sustain for over a minute. It quickly gets overwhelmingly difficult to compute.

And, as you said, for better or worse, multiple takes are usually spliced together for piano recordings.

[johnlewisgrant]

Guess I wasn't clear enough.   I put up another post, elsewhere here, which probably isn't clear either:  I'm NOT referring to controllers, or Roland V pianos, or piano samples.

Nothing to do with that.

I'm referring to "Disklaviers" (Yamaha and now, Bosendorfer) and Steinway "Spirio" technology.

MODERN PLAYER PIANOS, in other words!   The technology has advanced to the point where Yamaha, Bosendorfer (now owned by Yamaha) and Steinway make full-sized concert grands equipped with super-precise player piano mechanisms.  The "language" these systems use is a 1056? velocity (like pppp to ffff) form of MIDI (not the usual 127 velocity layers.   

Among other things, these systems allow you to record and have the piano play back exactly---and I mean EXACTLY--what you played. 

Hence my question re: recording, splicing, fixing small errors.   Because, of course, one of the features of these pianos is that on hearing your "performance" replayed by the player system, you can make whatever alterations you wish to the corresponding midi file.

Think Yamaha E-competitions, etc....   

[j_tour]

Well, I'm not so sure that a recorded acoustic piano, which has implicit effects due to microphone placement, and most surely EQ and compression added at some point in at least the mastering stage, could be distinguished from a softsynth (either a many-GB set of samples or even a modeling technique).

It's been a long time since I've personally used this tech, but I'm almost 100% sure that a great many performances you'll hear on disc, at least in a band context with a good bit of "exposed" acoustic piano parts, are done in studio with modeled or sampled pianos.

Editing MIDI tracks?  Sure, not much different than splicing with tape or with a digital track.  It's not my idea of a fun way to spend an afternoon/evening, using a mouse or directly entering values via a keyboard.

The values the MIDI standard can encode are limited, granted, but I'm not sure it's a deal-breaker when considering the mixed and mastered final recording.

With all those parameters accounted for and recognized, I'm not so sure there's a big difference from recording and editing with tape in the studio.

What probably is of most interest is that most serious musicians these days have some cheap access to studio tools. 

So, who knows?

The short of my concern is that once an acoustic piano is mic'ed, recorded, mixed, and mastered, you already have significant processing of the recording.  Just the mechanics of splicing or altering performances is a bit more user-friendly than taking a blade to tape and performing splices that way.

[johnlewisgrant]

Well, I'm not so sure that a recorded acoustic piano, which has implicit effects due to microphone placement, and most surely EQ and compression added at some point in at least the mastering stage, could be distinguished from a softsynth (either a many-GB set of samples or even a modeling technique).

It's been a long time since I've personally used this tech, but I'm almost 100% sure that a great many performances you'll hear on disc, at least in a band context with a good bit of "exposed" acoustic piano parts, are done in studio with modeled or sampled pianos.

Editing MIDI tracks?  Sure, not much different than splicing with tape or with a digital track.  It's not my idea of a fun way to spend an afternoon/evening, using a mouse or directly entering values via a keyboard.

The values the MIDI standard can encode are limited, granted, but I'm not sure it's a deal-breaker when considering the mixed and mastered final recording.

With all those parameters accounted for and recognized, I'm not so sure there's a big difference from recording and editing with tape in the studio.

What probably is of most interest is that most serious musicians these days have some cheap access to studio tools. 

So, who knows?

The short of my concern is that once an acoustic piano is mic'ed, recorded, mixed, and mastered, you already have significant processing of the recording.  Just the mechanics of splicing or altering performances is a bit more user-friendly than taking a blade to tape and performing splices that way.

I should apologize, because I'm running 2 identical threads at the same time!  Anyhow, the Disklavier technology, now on concert (CFX) Yamahas, Bosendorfers, and Steinways could completely change the way recordings are done.   I've talked to piano techs who say that the current crop of "player pianos" play no differently than the real thing.

So imagine: Pollini (for example) is in the recording studio and gives what he thinks is a perfect performance of a difficult Chopin etude.  But there's ONE little wrong note.  Or one trichord has gone out of tune during the performance.   Or the microphone placement wasn't optimal.

With current technology he can KEEP the once-in-a-lifetime performance.  He doesn't have to re-record it, and lose the first take he liked so much.  The Disklavier MIDI RECORD of the performance is edited to fix the bad note.  The microphone placements are adjusted.  The out-of-tune note is re-tuned. …..   And then... important part.. the piano replays the EXACT performance (because the new technology is that good) with all the necessary changes.

Pretty amazing, if you ask me.

What you say is true, though.  The sampled pianos (not real) are now pretty hard, once recorded, to distinguish from a recorded REAL piano.   That's why listening to LIVE is the gold standard, I guess.

[j_tour]

With current technology [Pollini] can KEEP the once-in-a-lifetime performance.  He doesn't have to re-record it, and lose the first take he liked so much.  The Disklavier MIDI RECORD of the performance is edited to fix the bad note.  The microphone placements are adjusted.  The out-of-tune note is re-tuned. …..   And then... important part.. the piano replays the EXACT performance (because the new technology is that good) with all the necessary changes.

Yeah, I see your point now.

There've always been MIDI sensor strips you can retrofit keyboard instruments with, like acoustic pianos or Rhodes electric pianos and so forth.

But if the data received via MIDI is much greater than just a limited velocity, in addition to note on-off, duration, pedal on-off, then it's a sure thing that this process would simplify the workflow, in terms of time and simplicity, than manually editing MIDI data.

A continuous stream, sort of.

I don't disbelieve the claims at all.

Not really an order of magnitude greater than recent-past tech, but it's certainly simpler, and has the advantage of using tools and technology most people are already familiar with.

I like it!