Well maybe I read the post by "Anotherscott" incorrectly, but the way I read that post it was suggesting there was a difference between a note struck using the double escapement feature without letting the damper settle, and striking a key while holding the damper up. I can't see there is any difference at all. Also you said
Again, maybe I'm misreading your post, but SPEED is not a factor in the action, simply how far the key is allowed to travel. You can play the notes extremely slowly and softly and so long as you give the hammer enough inertia to actually strike the strings you can do it once a minute if you like, just so long as you don't release the key up past a certain point.

Pete

Pete(H),
When I tested a real grand, the soft notes were always heard as new, distinct notes. If you are saying that they will not be heard as new notes, then I'm not sure this is correct. Obviously if they are played so softly that the hammer barely even makes contact with the string, then yes, it might not be heard at all, or perhaps it will "upset" the currently sustaining note just a little, causing a tiny twang sound, or something like that.

Also, the double-escapement point is LESS than 50%. On my Casio, it's at 36%. (just measuring one note, once). I think this will be commensurate with a real grand. (it's such an easy thing for them to get right, I don't see why it would be any different)

PeteF,
Ah, yes, I see what you are saying now. No, that's not what I meant at all. A few days ago I gave a procedure for testing for the presence of double-escapement, and this procedure involves moving the key very slowly in fact. ;^) (this was in another thread). All I meant in that post that you referred to is that if the player is lifting off the key as he repeats, and then he plays new notes, if they are playing fast enough, we will see that the keys are only returning partially. I.e, when the finger makes contact with the key again, it will be at a point in the key's return that is substantially less than a full return. That's all.

Greg.

I've looked up my piano tech book, being Piano Servicing Tuning and Rebuilding, by Arthur A. Reblitz (2nd Edition).

I'll avoid quoting directly in case I'm violating copyright. Here are the main points, paraphrased:

1. A grand piano can repeat for lesser key returns than an upright piano. In an upright piano, the key has to return almost all the way to the top before a new note can be played. In a grand piano, the keys only have to return less than half way. This results in grand pianos being able to repeat faster than an upright piano.

2. The mechanism that allows the grand to repeat for these smaller key returns is called the Repetition Lever. This mechanism is not present in upright pianos.

3. The dampers are lowered to the strings when the keys are fully released.

4. Particular mention is made of soft repetitions, and trills.

5. If you watch the hammers whilst someone is slowly trilling louder and louder, you will see that the hammers start off being quite close to the strings, and then slowly move further away from the strings, as the striking distance is increased.

6. Regardless of how softly or loudly a trill is performed, a grand that is properly regulated will always repeat.

Greg.

Here's a post from Kawai about full-stroke repetition rates of real grand pianos:
Repetition rate of a grand piano

Now, looking at the sheet music for Billy Joel's "Angry Young Man", this calls for repetitions at a rate of 12 per second. (this is the rate I had earlier calculated by analysing a recording!).
So, going by Kawai's info, on an average grand that can do full-stroke repeats at 10 per second, this means that A.Y.M would indeed be relying on partial key repeats in order to be able to play at the right tempo. However, on a really good piano, that can do 12 per second (or 15 for the Kawai Millenium!!!!) partial repeats would not occur, because the keys would return further. As most of us would know, A.Y.M has quite forte repeats, so we can forget about subtle pianissimo repeats for this song. ;^) (the average 10-per-second grand would be producing softer repeats than the better ones for this song - that's worth noting)

Greg.

Using your example of Angry Young Man, even at that speed, it sounds to me like the strings are damped between each hit. Maybe it's only partial damping, at that speed it would probably be impossible to tell, but the notes are not "ringing" the way they would if they were fully undamped between strikes. They are clean percussive strikes that do not exhibit the characteristics of hitting an undamped string. Again, try it yourself... hit that middle C in rapid succession and have someone grab the damper so it can't damp between hits, I think you'll find that it doesn't sound "right," as it does if you allow the damper to fall.

Anotherscott,
Did you read the post I just made, just before yours? I think it answers your question to a large extent. On an average piano, that can only do about 10 full-stroke repeats per second, partial repeats would be required. (whether they are "partial" enough that the dampers do not make any contact at all is hard to say though).

But yes, my piano book does say that the repetition lever is particularly advantageous for soft repetitions.

It's very hard for me to do the test you suggested, because I don't have ready access to a real grand piano.

You may well be right about the A.Y.M recording though.

EDIT: Oh, you obviously have read that post. Well, I said in that post that it depends on how good the piano is.

Greg.

It is possible to tell that from physical laws:
If the string where completely undamped, the vibration energy would increase on every stroke until the string breaks...

Obviously this is not the case and so there must be some damping. I believe it is the hammer felt that dampens the string. The more intensive the contact between hammer felt and string is, the more damping will happen.

It does therefore not happen, that a soft stroke will dampen a previous harder stroke very much.

Peter

Btw, I'm just listening now. I really can't be sure either way.

We need to find a Disklavier, perhaps in average non-optimum condition, play A.Y.M on it, and inspect the special extended MIDI data that it records. ;^)

Anyway, all I'm saying is that given the following:

a) grand pianos repeat for relatively shallow key returns
b) the dampers do not come into contact with the strings unless the key has returned almost all the way to the top,

then,
it is logical that if the repetition is rapid enough, even if the fingers leave the keys, then undamped repetitions MAY result, simply because the rate is rapid enough for the keys to only partially return.

Greg.

Sorry if I was unclear. I will try to clarify that point.

I was responding to the assertion that, if you repeat a note fast enough on a grand piano, the dampers will not fall back to the strings between strikes.

In that case, IF rapid repetitions of a note (with pedal up) do not allow the damper to hit the strings in between strikes, THEN rapidly repeating a note should sound exactly the same as repeating the note with a friend holding the damper off the strings.

In my test, I found that a rapid repetition did NOT sound the same as it did when I manually held the damper off the string.

Since (as you point out as well) the two scenarios should sound the same, and yet in practice they don't sound the same, my conclusion was that, even in a rapid repetition, the dampers are falling back onto the strings between strikes.

(And in case I simply wasn't repeating fast enough to get the desired behavior, I pointed out that, in listening to Angry Young Man, it still sounds to me like there is string damping between strikes, since I don't hear the characteristics of re-hitting an undamped string.)

Anotherscott - yes - play faster, or find a faster player.

OR - find a grand piano that is not as fast. They are not all equal, as I said.

Greg.

(I'm reminded of the Nissan vs Porsche war, where Porsche said that they could not get within coo-ee of Nissan's lap time in Nissan's new GTR. ;^ ;^)

Greg.

I've just inspected the MIDI from my Casio PX-330, when playing in the A.Y.M alternating hands fashion. There are undamped repetitions almost immediately. (I was playing as fast as I could though, and it may well have been slightly faster than 12 per second)

Of course, my Casio may well have an action that is slower than most grand pianos. ;^) Maybe it's not even as fast as most uprights. I would not be at all surprised.

Greg.

Here's my MIDI file:
http://www.box.net/shared/nh6bzfl4ar

and the MP3:
http://www.box.net/shared/qmqke1rc22

Greg.

Look to this video, preferabbly with slow motion or manual frame selection:
http://material.miyazaki-c.ed.jp/ipa/piano_sangyo/oto_dukuri/action/s1ma21.mpg

It is visible that:
The damper lifts.
The hammer hits the string.
The string moves faster than the damper and thus the string hits the damper ;-)

The damper bounces back.
Unfortunately the video only shows a single stroke, not a repetition, but it can be imagined what complex events can possibly happen on a repetition....

Very interesting indeed!

RE: my repetition test, I see that despite some undamped ones early on, almost all of them are normal repeats. (just quickly scrolling down, I think there's only three occasions where they occur, and it's possible that they occur more due to sloppy technique than sheer speed)

Greg.

Bingo! (yes?)

http://www.youtube.com/watch?v=2mX3kiA-qrU

Greg.

That's an interesting test. But I'm not sure it really tells us what is going on internally. The MIDI stream may tell you that your fast key repetition is sending an additional NOTE ON without first sending a NOTE OFF (an undamped repetition), but what is the Casio doing internally when it receives that second NOTE ON? i.e. ---

* It is generating a second instance of triggering the same sample, in addition to allowing the first sample to continue to ring? or...

* It is generating a second instance of triggering the same sample but, in the process, it is silencing what would otherwise the be the continued sound of the first sample?

If it is the latter, it would match what I feel is "typical" piano behavior (I know, we disagree about that). If it is the former, it might more closely match the behavior you're looking for, but still wouldn't be quite authentic, since "overlaying" an additional sample is not quite the same, sonically, as re-exciting a string that is already vibrating. To duplicate that effect, Casio would need an additional entire set of samples. i.e. the sound of a key being struck with the string undamped. (Note that this is not the same as the sound of the key being struck with the pedal down, since that would product a whole bunch of other sympathic vibrations from strings of other undamped keys.)

I can't comment on your test nor what you hear, however this link may help explain things. http://www.piano.christophersmit.com/actionDetail.html

The link at the bottom provides an animation of the action, including the ability to repeat a note using the double escapement facility. There is absolutely no doubt the damper does not touch the strings if the note is repeated after being only partially released, there is no difference between repeating a note in this way versus either pressing the sustain pedal down or physically holding the damper up.

Pete

Anotherscott,
I agree with you that usually, when one is repeating rapidly, but forte (where the fingers leave the keys), the repetitions are probably USUALLY full-stroke, and damped. If one is playing like this, but right on the bleeding edge where the repetitions are partial (and either undamped, or partially damped), I suspect it's a very risky situation, and some repeats may be a bit too rapid, and simply not be heard at all. As I said, my piano book does seem to say that the double-escapement is more for soft repetitions & trills than forte repetitions. Looking at various live performances of Billy Joel playing AYM, I see that sometimes he does play much slower. Perhaps he only plays at a speed at which he is very confident that the piano he is playing at the time can perform FULL stroke repetitions.

Now, regarding the SOUND for partial repetitions, I can't hear the Casio doing anything fancy at all yet. (I'll have a closer listen later). When I tested a Clavinova, yes, I thought I could hear the slight chorussing sound of multiple, overlapped voices. Further, a Clavinova owner has reported that if multiple Note-Ons are sent to it, but only ONE Note-Off at the end of the repeat sequence, the note continues to ring. This strongly suggests that it does overlap voices, and needs a Note-Off for every sustaining voice. (this was in another forum) Both the Yamaha and the Casio do balance the Note-Ons and Note-Offs. I'm assuming Roland will be the same but I have yet to see any reports.

My understanding is that almost all DPs simply overlap voices for repeated notes with the sustain pedal down, and IMHO this actually sounds quite realistic. I'm not convinced that even Pianoteq does anything differently to this - it doesn't sound any more authentic than the results of overlapping. (moreover, I can see that Pianoteq does consume more voices for repeated notes, which is a bit of evidence to suggest that it does use the standard voice overlap method)

Greg.

Yes, and that is consistent with what I've said as well.

In addition, I agree with your earlier statement, "SPEED is not a factor in the action, simply how far the key is allowed to travel." Whether the damper falls between strikes is a function of how high you lift your finger before restriking, not a function of how fast you play, which is why I was taking issue with the talk about it being related to speed of repetition. Angry Young Man has fast repetition, but he is lifting his fingers so high off the keys between each strike that it sounds to me like he gets a full release rather than a partial release. I think keys can return very quickly indeed, and the only thing that prevents a key from always fully returning before a re-strike is the inability of the pianist to raise his finger high enough before the re-strike. (Billy Joel accomplishes it by using two hands for that part.) I suspect that having the string damped before re-strike is actually usually the desired behavior, and the fact that some trills or same-key-different-finger hits may actually re-trigger a note at such a low travel point that the damper doesn't fall might even be an unfortunate by-product of the mechanism that allows a key to be re-triggered at such a low point in the first place.

Also, just pointing out for accuracy's sake (and I'm sure you know this)... when you say "there is no difference between repeating a note in this way versus either pressing the sustain pedal down or physically holding the damper up" -- there is the difference in that, when the sustain pedal is down, you will get a sonically different result because of all the other undamped strings that will resonate.

I think you're right... I think most DPs (at least the lower cost ones that I'm familiar with) return more slowly than a typical grand.

Anotherscott,
I'm not yet convinced that it's impossible for a player to play rapidly enough such that he lifts his finger off completely, and then restrikes it, before the key has returned sufficiently to damp the strings. It MIGHT only be possible on an average piano, not a very good one, and it might take special skill.

I've already posted a YouTube clip of someone who I think, just maybe, is playing that fast. (it's a shame there is so much reverb though)

Anyway, on my little plastic Casio, I have already kind of proven that it's possible. ;^) (it's hard to prove that it wasn't just poor technique - I need a high speed camera, which by the way I see are becoming quite affordable)

Greg.

Also, remember that Billy Joel is using alternating hands. There is nothing whatsoever preventing undamped repetitions, because the two fingers can play the same note VERY VERY fast. I.e, maybe sometimes the second finger/hand comes down JUST AFTER the finger that played the previous note. Even if it's not often sustainable, I'd be surprised if it didn't happen sometimes, and when it does, it's a better result than it would be on an upright piano. When this "accident" occurs on an upright piano, no repetition will occur whatsoever.

GReg.

Even if it's possible that, on an inferior design, one could re-strike a key from that level of lift and manage to re-trigger it before the damper could fall, I'm sure you'd agree, there would be no point in building the duplication of such a flaw into a DP. And really, the more we've talked about this, the more I'm convinced that the ability of a DP to re-trigger a key without damping it is of pretty much no importance, even though, yes, it is an effect that can occur on a real grand in the right circumstances.

First, it seems to rarely come into play. Second, it's a very subtle effect, one you can't always even be sure whether or not you're hearing. Third, you suggest that the effect may be more or less pronounced on different grands... and moreover, upright pianos don't do it at all, so it's hardly a criteria for creating a "real" piano sound. (If I had a DP that sounded as good as some of the fine uprights I've played, I'd be thrilled!) There are just so many more substantial areas where the typical DP can be improved, that this is so minor in comparison.

I think the main issue is for keys to be able to be quickly repeated, and I think the "third sensor" is really a way to try to get around the fact that DP keys typically just don't come up as fast as real piano keys do, so that repetitions far slower than "Angry Young Man" have still been problematic on many DPs.

Anotherscott,
As we have seen from the text in my piano tech book, the double escapement mechanism is present to allow repeats with relatively small key returns. Even if you don't think this has any benefit for vigorous repetition, it does have a very definite benefit for subtle, pianissimo repetitions, and trills.

It's very simple. Without the double escapement, the key has to be lifted further, and the hammer is also released to a greater extent - i.e - the hammer is further away from the strings. This took more time than it would have on a grand piano. Now, the player wants to play another repetition, and again, softly. He has to press down on the key softly to create the soft note, and again, since the key, and hammer, have to move further, this takes more time than it would have on a grand piano. On a grand piano, the double escapement catches the hammer, and prevents it from being released very far away from the strings, at least until the player lifts off further.

Now, assuming that we are trying to make digital pianos behave as closely to a grand piano as possible, the addition of the third sensor makes OVERWHELMING sense to me. On a tri-sensor action, the player will be able to play in a similar fashion to the way they would on a grand piano. That's a plus, a benefit.

Again, I totally agree that it's subtle. I have never advised anyone to place a big importance on this feature. I merely mention the presence of this feature.

Greg.

Also, just to reiterate, the reason we can't simply use two sensors, and put the top sensor at a very low key release point, is presumably because we also have to take into account legato and staccato behaviour. If the top sensor is way down low, it will be too difficult to play legato, and staccato notes may be TOO staccato. So, a two-sensor action has to place the top sensor in a position of reasonable compromise. A tri-sensor action does not have to make this compromise.

Greg.

And regarding very high standard grands, with extremely fast actions, I suppose the reason it would still makes sense to have the double escapement is because it saves effort for the pianist, because he does not have to move his fingers as far as he does on an upright piano. (e.g for trilling)

Greg.

I agree with you that it is beneficial to permit repetitions with small-travel returns; I only differ in that I think the issue of such a repeated note being undamped vs. damped is not of great significance.

I understand the benefit of the three-sensor approach; I just don't assume that's necessarily the only way to achieve the desired results. Which gets back to what I said early on (in the other related thread, I believe)... that I can imagine that there may well be 2-sensor models from some manufacturers that permit a faster repeat than the 3-sensor Casio PX-130 or whatever, due to other design parameters. So if fast repeat was a big consideration for me, I would judge by which one actually had the better repeat capability, rather than by looking at the spec sheet to find out which had a third sensor. Just like you can't assume that a given 6-cylinder car will out-perform every 4-cylinder car, you have to look at the real world results, not just the technologies.

Anotherscott,
re: the Casio not necessarily being any faster than any other two-sensor model, I agree, and I had already stated this in an earlier reply in this thread:

earlier post

It's not just about raw speed though - it's overall behaviour.

Greg.