Re: piano touch

Hi

This relates to a topic in the fascinating archive, soon to be
on-line.

I came across across a discussion of piano touch in the archives,
and this is something I went into at one point.

I used to play acoustic piano a fair amount, improvising, and was
convinced there was more to its touch than just volume.

For instance, one can play a note ppp with a lot of attack, and
one can play a loud cantabile note with very little in the way
of attack.

The conventional explanation that one reads in books about
musical instruments is that this is just because of the context in which
the note is played, the rhythm and pacing of notes etc. However
it somehow never seemed totally convincing; didn't seem altogether
to match what I heard when playing the instrument.

Maybe the touch is more subtle than for other instrumetns in some
ways, but the acoustic piano definitely seemed to have something of
that sort, or so I felt, and so I'd also been told by pianists!

So did some investigation, and this is what I found (doesn't seem
to be mentioned in the thread, and seems to be little known as
yet generally, though I may easily have missed it!)

When you play a note, then first the hammer gets triggered to
move, and indeed, at the moment of impact the hammer is moving
freely, with no contact with the key. So as far as the hammer
is concerned, all that matters is the speed at which it hits
the string.

So this much of the conventional explanation of piano action
is correct.

However, the damper is another matter. The sequence of actions
as you hit a key is:

Hammer gets set in motion first

Then damper is released.

You can actually find a point in the action where the damper is
released. Then if you play to just depress the keys to that point,
and stop, you can play a totally damped note - you can do it
in such a way that the damper never leaves the string.

It is quite possible to play a tune like that, and get
a very soft totally muffled sound, unlike anything one usually
hears on the piano. It can be loud but softly muffled, doesn't have
to be quiet, just so long as you make sure that you stop pressing
the key at the exact point where the damper release is triggered.

Now, if you play hitting the key slowly, and accelerate into
the note, as one does for a note with a lot of attack, then
this means you lift the damper very promptly.

One can do that so that as the hammer goes free moving quite
slowly, and then you accelerate into the next part, to lift the
damper, so that when the hammer hits the string, the damper is
dotally raised from the string. Result is a note which can
be very quiet, with a lot of attack.

The other way, if you hit the key fast, and then decelerate
into the note, as one does when playing cantabile, then the
damper is only beginning to leave the string as the hammer
hits it. So you get a note with the beginning attack softened
out. That way you can get even a loud note with very little
in the way of attack, and it has a singing quality to it if
you get it right.

There may be more to it than that, but that at least is
enough to show that the piano touch is two dimensional,
with attack and volume varying independently.

I remember I contacted a researcher who confirmed that
modern analysis of piano touch was coming to this very
conclusion, though I'm afraid I forget the details.

Now for another interesting thing about it. It would be
perfectly possible to design an electronic keyboard to
vary in attack and volume in the same way, by using
the same technique - to have two sensors, one for the
release of the damper, and one for the velocity for
the hammer.

Would add a new dimension to touch, and could also be used
in a general way , not just to vary piano touch, but
touch for any MIDI instrument. One could make the touch
vary much more than it does for the piano if one wished,
or indeed, make it more subtle.

So, why doesn't someone build such an instrument?

It isn't my field at all, I have nothing to do with
building instruments, but I know many on this list do
so maybe it will give them some ideas. I hope so!

I suppose chances are, someone else has thought of it
before, but I've never come across it.

Has anyone else come across this?

It even has microtonal significance (to show this isn't
completely OT).

One could do it so that a note is made flatter or sharper
instead depending on the amount of "attack". Since pianists
get good at playing a note with just the exact degree of
cantabile or attack that they want, to a very fine degree,
then one would have equally fine control over the amount
of pitch bend, just by varying the way in which one
played the note, either by accelerating into it, or
decelerating into it.

Or indeed, one could vary any Midi controller the same
way, just assign one of them to the attack (e.g. if one is
happy to play in a fixed tuning).

So idea is, make a keyboard which outputs both velocity
and a controller message, depending on the acceleration
or deceleration into the note as one presses a key.

I think we'd have heard of it by now if it had been done,
as there are other designs for microtonal keyboards, but
never seen this.

I'm sure it will be done eventually, but perhaps I can
help accelerate the process by mentioning the idea here.

Robert

Hi,

Actually, come to think of it, I think the damper is connected to
the key directly by a string, or something of that sort. So one
is almost completely in control of the way it leaves the string
(though once it gets some impetus to leave the string, you can't
stop it short again as the string pulls only).

If I remember correctly anyway.

So that makes it more than two dimensional, as one can make it
so that the damper leaves the string slowly or quickly, and
accelerate or not too as it leaves the string. That is about
four dimensions already.

Might explain why pianists think they can also do many other kinds
of touch, e.g. to play to sound like various instruments, rather
than just attack and cantabile.

With the electronic keyboard, maybe just possibly this might mean
it would be feasible to vary two or more controllers, by measuring
the acceleration at various places in the key movement.

Robert

At 07:19 AM 4/9/01 +0100, you wrote:
>Hi,
>
>Actually, come to think of it, I think the damper is connected to
>the key directly by a string, or something of that sort. So one
>is almost completely in control of the way it leaves the string
Wtually, although this may be a minor point in terms of what touch
variations contribute, in grands the damper is lifted by a pad on the
back-end of the key and returns via gravity. In uprights the damper is
pulled away from the string by a spoon-shaped projection on the "wippen"
the chunk of wood on which is mounted the backcheck, jack and jack spring.
Since its movement is horizontal (well, okay, an arc), a spring returns it
to its position against the string.
Jay

Hi Jay,

Thanks for the information!

Robert