Yamaha NU1X "loud note" fix seems to cause another problem

My candidates would be the NV5 and the N1X. I think NV10 is just too expensive at more than twice the price of a NU1X.

Another NU1X user on PW has reported they cannot reproduce the problem!

RIP I have the NU1 too, but as I said I did not install the anti-loud-note firmware, so I can't test it, but I can speculate.

I think that the link to the loud note firmware is supposition only, so you can test it if so inclined.

spanishbuddha I think that the link to the loud note firmware is supposition only, so you can test it if so inclined.

You are quite right, my friend. I did that test, using PTQ as the MIDI velocity display and no sound (because my daughter was playing violin) and 🎉 I noticed the same problem that @porco_rosso described! I tried to play random notes with arm-weight technique, falling from about 4" above the keyboard and most of the time I got velocities around 120+ and occasionally around 105 and around 88. This is a NU1 with official firmware (no-loud-note-fix one).

Now, you may wonder why I never experienced that while playing and the answer is pretty easy. To achieve these 120+ velocities I had to do the technique described above, which after only a couple of minutes of doing made my fingertips hurts. I never play with those velocities, and certainly never for so long. During that time I must have played about 50 notes, and only about 5 showed the problem unambiguously, so if these statistics are correct, no wonder it went unnoticed for so long.

In any case, I think this is a shocking find! If you haven't already, @spanishbuddha I suggest you try it that way too and perhaps suggest to the PW thread also. You really have to "bang hard on the keyboard" to see this happening, your "banging" must be at a level of surely triggering 120+. As a speculation, the problem could be that on such velocities the downstroke is so fast that it misses the second trigger and reaches it only later after rebouncing from the keybed, resulting in a lower apparent value.

Now, as I said, I experienced no sound, so I don't know on that side, I'll give it a try another day.

PS: I also noticed that the NU1 sends aftertouch signals while doing partial note release, so maybe the loud note fix use those?

Thank you RIP delvento for testing it! Very interesting that it also appears on your NU1 without the lout note fix firmware. And yes, you really have to hit the keys with a lot a weight and speed to get these high velocities of 120+. But imo this corresponds well to get a real ff/fff on an acoustic piano. I would like to share these findings in the pianoworld forum, but can't send a link to this forum there (I think). Would you like share them there on the corresponding thread? 🙂 http://forum.pianoworld.com/ubbthreads.php/topics/3170250/casio-gp-310-510-vs-yamaha-nu1x.html#Post3170250

I was aware quite some time of the issues.
If I remember properly, the sudden 127 notes weren't present on the KAWAI Novus Device I had to mess for.

porco_rosso

To eddiepiano, delvento or anyone who has hands-on experience banging out these “ … lot a weight and speed to get these high velocities…” ffs and fffs on a real AP/GP compared to a midi/dp equivalent, does one really need to strike the keys with arm&hand blows delivered from a starting point of approx 4 inches above the keyboard?

Since there are these techy end-user tools that allow measurements that help’s one so minded to compare common characteristic results of DPs and APs - like tools/apps that allows one to measure midi/dp key velocity 0-127, and tools for measuring resultant SPL/dB values coming out of given AP/DP - is there a tool that measures the actual applied force of keystrokes so that it is known that a force of __ is required to be applied to an AP to produce ff or fff, etc ?

Its easy to presume there is such a (expensive) tool for designers in AP & DP laboratories, but how about a practical tool for end users? ( aside from something like MacMacMac’s “ my ears & hands are my tools “ 🙂 ) …. how do they know, with some certainty, that a respectable DP designed to emulate (or otherwise mimic) the AP experience is in fact performing comparably well regarding force ?

@Drew-r yes they have such machines. Quite expensive and complicated. Look at that video from the upright action improvement that got people excited, there's one in action.

porco_rosso Would you like share them there on the corresponding thread?

I wrote an answer there. But boy, PW is really tedious to utilize. I can't bear it anymore 😀

I did my best, and sorry if I sounded annoyed: I was (at the website, not at the people I was answering to). I don't think I'll have the stamina to continue the discussion there -- bring those people here instead…

RIP I don't think I'll have the stamina to continue the discussion there

That's a noteworthy phrase.

I think that here we have discussions. Over there we don't (or didn't). There were just threads.

Part of the difference is this newer platform. And part of it is the membership here.

Summary: I like it here. A lot.

Thank you for posting it delvento.

CyberGene I tend to the N1X as well. It's "only" 1200 € more expensive than the NV-5S, but

• it has a (short) grand action; I even liked the NU1X super short upright action way better then all regular digitalpiano actions
• I like the Yamaha samples much more than Kawais
• binaural audio samples for headphones
• integrated USB audio interface, that I found very useful with the NU1X for VSTs
• also sends note-off velocity iicr due to hammer and keysensor
• I like the simple but functional interface much better than the slow, low resolution display

The NV-10S has a wonderful action that I like even more than N1Xs from my last impression. But all the other pros of the N1X still apply. It's almost 2000 € less expensive than the NV-10S. And if the AvantGrand action is good enough for a pianist like him it will be way more than good enough for me for the rest of my life (it's the N3 in these videos):

hello, I have tried to replicate this problem cited in this thread. This is the second time I have tried. I have connected pianoteq and the highest values are around 110 (attached image), I have not noticed that they are filtered. To repeat the experiment it would be necessary to know all the configuration both in the NU1X and in pianoteq (which I requested in PW but I didn't get any answer). I am very interested in this topic because I don't like as an owner of this instrument that it is abandoned by Yamaha with unresolved problems. I will be the first, that I know if I detect it, I will send mails with complaints to Yamaha.

IMO I have the feeling that the instrument is being pushed to the limit with tests that I don't know if with a regular practice occur. It doesn't happen to me. I would like to know in which works the problem is apparent so I can try to reproduce it by playing normally. It has also been commented that the solution to the "load note" problem was a patch to get out of the trouble when, as far as I know, no one has obtained real information about how Yamaha solved the problem.

On the other hand I can't find any information about other series with the exact same laser system and the same vertical action (or similar but always vertical) in which the same problems have appeared (silent SG2 and SH2 series, transacoustic TA2, old hybrids DUP-8, DUP-22b, etc). If it was a problem of this sensor in combination with a vertical action it should have appeared beyond NU1 since it has been mounted for more than 20 years. In my opinion it is exclusively a software problem of the NU1 that the NU1X "inherited" and that in my experience no longer appears with the V2.1 correction.

In upright pianos there is no need for the hammer sensor because the key and hammer movement is synchronous. There is no double escapement or repetition lever. In the case of the NU1X not even the "hammer" has to produce a sound (as it happens with the silent and transacoustic); it would have the same behavior as the counterweight lever of the GH action or any other simulated one. With the sensor that these pianos have, they should collect all the information of the key movement (down and up) and transfer it correctly to the corresponding velocity values. In the grand piano action it is mandatory to have double sensor, it would be impossible to detect what is happening in the repetitions when the double escapement mechanism is acting with only the key sensor.

ediegus In the grand piano action it is mandatory to have double sensor, it would be impossible to detect what is happening in the repetitions when the double escapement mechanism is acting with only the key sensor.

I don't think this statement is entirely correct. Hammer sensors are beneficial in both grands or uprights, not because of the double escapement, but because the systems calculate key velocity based on timing between 2 points regardless of acceleration. So, using only the timing you can run into situations in which the same timing leads in reality to a ppp note or a silent strike (sudden acceleration in the last portion of key travel is enough to ppp). Of course you can measure acceleration continuously for all 88 keys, but that requires complex realtime computation far beyond what is implemmented in DPs. One additional simple hammer sensor solves this and guarantees correct detection of silent notes.
PnoScan and many other systems (including from yamaha) only use key sensors in grand pianos. And sometimes run into problems...

Hey ediegus, thank you for your participation! Sorry for not answering yet in the pianoworld thread.

First of all, as I wrote before, I noticed this problem more and more in regular playing and practicing - only after that I tested it systematically with Pianoteq to be certain about the nature of the problem. I am not a trained concert pianist who can play a bombastic ffff on a concert grand, but even me as a late beginner/early intermediate player can reach very high velocities - as it is supposed, thats why they are 128 MIDI velocity steps. Everyone playing classical music will get to literature that demands playing ff+. I really wouldn't say that I push the instruments artificially to its limits.

The settings for the NU1X and Pianoteq: default. Medium touch curve on the NU1X, perfectly straight line as the velocity curve in Pianoteq. But as I wrote before it also happens with the internal voice of the NU1(X).

As vagfilm wrote here and @RIP explained in the pianoworld forum, I do think there is an advantage using a hammer sensor instead/additionally to the key sensor. Kawai wouldn't add the hammer sensor just for fun to increase production costs in their upright silent and hybrid pianos.

ediegus I have connected pianoteq and the highest values are around 110

That's the "problem". You are not playing loud enough. As I said, to reproduce the problem I had to play with arm/body weight technique, and at such a level that my fingertips hurt after two minutes of that. Now, all human beings are different and I am on the frail side, so if @porco_rosso is on the normal-strong side in his upper body, I understand how the problem can be frequent for him -- it's just a non issue for me and from your test I think it's a non-issue for you. But if you try to play with the weight of your arms and body, falling from 4" (10cm) or a bit more, you should be able to get consistent values around 120+ and some of those (while you are still "banging" with the same intensity) will instead show the velocities in the 80s.

As I speculate in an earlier message of this conversation, perhaps this might be cause by the keys going too fast to be detected on the way down, and instead being detected on the slight bounce back from keybed, measuring an apparent slower velocity. To test this hypothesis, one would need a high resolution, fast speed (aka slow-mo) camera (perhaps a gopro would do it) but being a non-issue for me, I'm not going to bother (besides not having such a camera, even if my phone does have a slow-motion mode).

I agree with @vagfilm about the hammer sensor, and I'd add that the first escapement (which upright pianos do have) is sufficient for having hammers and keys moving in different ways.

porco_rosso I tend to the N1X as well. It's "only" 1200 € more expensive than the NV-5S

Check if you can bargain down the N1X.

When I walked into my local store last year the sales person knocked USD $1,800 off the price without me even asking for a discount (reducing it down to about USD $6,750 + tax down from the original price of USD $8,550 + tax).

This made the N1X almost the same price as the NV5S. I can see Kawai struggling to offload the NV5S in that situation.

RIP

I copy my answer from PW to the same discussion 🙂

Hello, just because there is a maximum value (the +120 -127 values) does not mean that it is necessary to reach it. It is a large range that allows a lot of nuances in the scale from 0 to 127 available. On my old clavinova clp-120 it was also not possible to get a +120 if you didn't want to break the keyboard and it was also not possible to get a dpi range below 10 really clear on the pianoteq velocity curve for this reason was the GH keyboard badly made or badly calibrated? I don't think so. On the NU1X I get pretty much the same thing and I find it easier to adjust the velocity curve to the expression the piano offers. Just because the wood action is hard and allows you to pound wildly on the keys doesn't mean it's something you have to do. Of course I'm not going to lend myself to hurting my fingers or pushing the keyboard to the limit by putting the action parts at risk or damaging the felts to find a problem I don't have now (and generate a new one).
I consider it an absolute excess to be playing random keys with force to see if a problem or a strange response from the software is found. So I ask for a piece of music where the NU1X action is not able to solve it.

In the case of the vertical action, there will never be a decoupling between the escapement and the key, the parts are unequivocally connected by the wippen.In addition, the flange is attached to the wippen, which pulls the hammer at the same time as the wippen is lowered in a single position to return the key to its initial position. This is the reason why by measuring just the key you would already have all the information of what the complete action is doing.

From my point of view if there are situations that give rise to speed measurement errors it is that the sensor is not well calibrated and there is a software failure, but not a lack of measurement hardware. Thanks for your comments and feedback.

burkey Check if you can bargain down the N1X.

I don't know if this is a cultural difference, but at least in mainly digital piano stores I have the strong feeling that there is no room for negotiation especially for high demanded hybrid pianos. That's different for acoustic pianos.

I strongly think about the MP11SE as a mid-term solution for a few years until I can more comfortably afford a hybrid grand or even silent piano. I like its action very much and the action alone is a significant upgrade to my FP-30 - only the "missing" MIDI velocities below 30 concern me as I really would like to use it as a MIDI Controller.

ediegus I consider it an absolute excess to be playing random keys with force to see if a problem or a strange response from the software is found. So I ask for a piece of music where the NU1X action is not able to solve it.

There are literally thousands of pieces requiring a ff+. And I don't like to repeat myself, but I noticed this problem in regular playing and practicing.

porco_rosso There are literally thousands of pieces requiring a ff+. And I don't like to repeat myself, but I noticed this problem in regular playing and practicing.

I do think it's shame that you never made serious attempts to discover whether this was a fault with your piano or whether it was representative of all NU1X pianos.

porco_rosso in mainly digital piano stores I have the strong feeling

That is self-defeating.

Don't think about negotiating. Instead, think about how much you want to pay ... and make an offer. Simple.

porco_rosso

Of course I know there are thousands of works with ff+. What I have requested is one or more of the pieces you have practiced regularly to see if I can find the problem.