Piano World Home Page Forums Our Most Popular Forums Digital Pianos - Electronic Pianos - Synths & Keyboards New Yamaha Nu1X ‘loud note’ problem with Firmware 2.1

@MMM in the same way that in any upright action: upon repetition, if you not release the key high enough (almost completely), the hammer is not thrown far away and it does not hit the string.

Hammer sensors will notice this and and have a silent note, key sensors will just notice the fast key movement and trigger a loud note

I suppose the note to be loud because when the escapement is not armed again on a repetition, the key is less weighted and the note louder.

Using only key sensors can’t measure the actual status of the escapement, then it may be wrong sometimes.

Depending of the position of the sensors, we can have louder notes (note triggered with the escapement not armed) or silent notes (note not triggered be we feel the escapement armed which can surprise too).

Using rubber sensors would not change this behaviour. (But would alter the feeling).

Reading the Kawai brochure, (NV5)


Then, the main difference is that the NV5 measure hammer movement and the NU1X key movements.

Is this also true of the NV10 and N1X?

Yes.

The N1X also uses key sensors.

Kind regards,
James
x

I think the NV-10 uses the same hammer sensor system as the NV-5 (and Aures, ATX, etc.) so no loud note issues, it *knows* when a real hammer strike occurs.

The N1X and N3X use a hammer sensor system AND a key sensor system. I *think* the key sensor acts as the primary velocity detector, the hammer sensor verifies a true strike plus velocity, and the key sensor also acts as a continuous real-time damper sensor (whereas the NV-10 uses the hammer sensor to double as a 1/0 key-off sensor).

And of course there's Cybrid v1.0 which uses purely hammer shank measurements. CG has announced v1.1 with double the scanning speed and v2.0 with key sensors for release velocity to improve accuracy for the damper release (for the few VI which support it): https://github.com/ekumanov/cybrid

Hi U3Piano -

No. I took you seriously. Perhaps I should have studied the comment more and then maybe the penny would drop (as we say over here). But never mind, it's all clear now.

Best regards, John

IMO while a silent note is frustrating, yet (for me) it is not alarming. I would much rather be a bit annoyed than alarmed. Best, John

Ok, glad it is clear now. Sorry for the confusion.

The thing I'm not clear about for the fix is, I don't think it *solves* the loud note issue, in that I don't think it 100% detects a loud note and changes it into a silent note. What I've heard is that it reduces the volume of a loud note (if it suspects a strike may be a missed hammer, it doesn't play the note at the full velocity stuck, it dampens it a bit). This has the effect of making the loud note stand out less, and more likely to be what the pianist intended.

So that leaves two effects: 1) what if it gets the detection wrong? Then you had intended an ff strike and only get an mf even if you played it correctly; and 2) it's up to the player to decide whether the piano is at that point mitigating a limitation in an upright action, or masking a human error element that should really be addressed with changes to technique. I guess as 3) it's also kind of odd to clearly feel a missed hammer and feel the key sink without resistance to the keybed but to hear a normal-ish velocity strike result from that. Though as you mention, "odd" is not the same as "alarming."

As I wrote I did not install the fix because the problem is useful to improve my technique, but here is some speculation

Very unlikely. In my experience the loud note happens at extremely high velocity (I haven't checked, but I suspect it's always 127), which is seldom is ever reached in normal playing, and especially not in one note only and sudden. I am not a very strong pianist, but by comparison, in my fff I am never able to reach a loudness level (even with chords) as strong as those single notes.

But a simple test is possible: you play whatever at mf and an assistant strikes a note with all their strength with arm-weight. This will show what happens in your scenario. I'd do this test if I could by I can't since I haven't installed the fix.

Not sure what you mean exactly here. I guess the goal is emulating an acoustic upright as much as possible. There will always be difference, and actually there are different upright actions (including the ones with magnets) which behave differently.... so? Arguably some are better and some are worse, so in my opinion we should take the best that tech (purely mechanical or digital) has to offer. And best is obviously different for different people/context, as my non-installation of a fix which I like demonstrates (if there was a command to turn it on and off for different things, I'd have installed it)

Closer if not identical to acoustic upright, so this does not sound so bad to me.

Perhaps we're running into different issues, or experiencing the same issue differently, but what you describe isn't what I see. The loud note in my case is rarely ever at 127, though assuming the keybed is capable of producing such a strike at the given touch curve, I'm sure it can happen.

Rather, because the issue is that the hammer isn't caught by the action when the key is struck, so other than failing to "fly the hammer," the keyweight is simply dramatically reduced, resulting in a faster velocity keypress than intended, and sound generation that correctly interprets that velocity, but fails to properly predict the captured hammer that should really mute the sound.

In my case, this can be ANY velocity louder than intended, though it's clearly more audible and jarring when it is significantly louder than the pianist's intent. Cybergene posted a method to reliably trigger the loud note on the NU1/NU1X by playing a key and then returning it to within 1-0.5mm of the resting height before repeating, in this case I can trigger multiple "loud notes" at a number of different velocities (not just 120-127), the key factor being that the sounding of the note is not accompanied by hammer travel. I notice the same thing when loud note during actual play as well.


Yes, there are definitely differences between acoustics. Even discounting the uprights that have a true double-escapement implemented, I cannot easily replicate the mechanical action on several Yamaha U3s I play. If I had to guess based on the sample size available to me, it's something that happens more easily/frequently on the NU1/NU1X, for whatever reason (action geometry, keystick length, etc.).

My personal opinion of a hybrid with an acoustic action is that the electronics should behave exactly as the action it's designed around behaves. If you choose to build a hybrid around an action without a double escapement, then the sound generator should faithfully respond to the mechanical limitations of the action. If that behavior isn't desired, then a hybrid can be built around a different upright action with a proper double escapement, or around a grand action. Just my $0.02 of course.

I know other people prefer that the benefits of digital be used to the fullest extent possible to mitigate any acoustic limitations/failings. It's the old "most faithful reproduction" versus "most ideal reproduction" argument.

You are right, after I posted my answer I did check and it's rarely at 127 for me too. I'm sure it's the same thing.


I independently reliably found that method of triggering the loud note, and I totally agree with the "feel" you describe and (after my recent analysis of velocities of loud node) I realized that you are right.



Years ago I had one upright and I don't remember feeling a missed note, even though I do rarely experience a missed note in my current grand. So with my even smaller sample I agree with your opinion "this happens more often on the NU1(x) than on other instruments".


I understand your point of view. With that, you basically say that only hammer sensors are acceptable. And I agree that they would be best (that's why I unsuccessfully nudged CG to work with QRS or PianoDisc to provide his hammer-sensor based sensor to the mass). Yet, most hybrids and silent acoustic have key, not hammer sensors... so what? Let's do the best of what we have.

I think this can quickly go down a fractal path--do you NEED hammer sensors? Theoretically no, if you can have key sensors that accurately (ok, perfectly) predict/translate key position and movement to hammer movement (including changes due to changes in regulation/tolerances). Presumably, that's really hard to do, especially when it's also fairly easy to simply "go to the source" and measure the velocity and position of the hammer directly.

But, are hammer sensors enough? Kawai's implementation probably has fewer of these false positives and negatives, but without key sensors, you now have the opposite problem of using the hammer as a proxy for key/damper position. Yamaha's gradient key sensor tracks key position continuously, and that's used not only for real-time note-off velocity (which is not supported at all on Kawai's hybrids, and only after-the-fact on typical rubber-dome DP sensors), smooth release, partial-damper positioning and feathering, etc. Granted, all of these are way distant in priority to the fundamental piano function of getting the hammer-on-string keystrike right, but there's still a tradeoff. IMHO Yamaha really went the no-compromises route with their hybrid grands in utilizing both hammer sensors AND continuous key sensors.

I do agree that given key sensors are more commonplace and easier to implement, we/they should do the best they can. The question really is, is the "best" to most realistically mimic the behavior of the instrument (silent strike when the hammer is captured), or to best respond to the intent of the pianist (play a note at the intended velocity even if it defies the physical behavior of the instrument)? That requires taking a subjective position, IMO

I agree with everything you said, but this last "subjective position"
This is more than subjective! Even a single subject (like myself) would like to have it both ways: loud note when practicing, mitigated when "performing". Since I cannot have it both ways I decided to keep the loud note only for practicing, since I more often "perform" on the grand rather than digital....

Yamaha didn't play this smart. They could have made it a switchable user option, (loud or no note) and advertise it as an advanced practice function.

The problem here is, the Software doesn't decide playing a loud note just because it's build that way. The optical sensor is registering a faster keymovement than intended which produces the loud note. The Software cannot distinguish whether it's an intentional fff or not.

Written like this we can understand than Kawai have some false strikes. With an hammer sensor, we will have no loud notes, and silent notes would be silent too on an acoustic piano. We can’t name them false negative.

Ok, but then how did they (partly) fix this? I guess the software recognizes the unintended loud note now, somehow, most of the time.