Rattling sound at high G sharp only - on most pianos in ver 6.0

Thank you, I will try to make something later today. With the exact midi note, number and as a clear sound files as possible. Maybe it's a "ghost" note mentally, or just a perception from me. You know, some people thinks all F sharps on any instrument sounds buzzy, and has some built in "unpleasantness" to it. It's mostly heard and noticed when I whack it, i e full velocity at 127. Both on headphones (AKG 701) and coaxial studio monitors (Tannoy). Soundcard is RME babyface, 44.1 Khz, at buffer 128, 32 voice polyphony. I tried changes of rate and buffers, and even changed out to older soundcards, with firewire, usb, just to make sure it wasn't that.

But I'll return. Thank you for all the tips and tricks.

Fabulous Mats, Pianos do interesting things around 127 velocity

From what you mention, have you tried 48k instead of 44.1kHz?

Or higher rates like 96 or 192 kHz if your RME version has it.

That's an interesting thing, it's probably even more advanced than we could imagine (without a whole lot of years or decades up close to the mathematics) but I do believe it is both from string to string plus the whole radiating modelling also.

Can't wait for the new string work introduced in the Karsten collection of instruments to roll out to all the pianos.

At the shared files it is uploadeed now, it seems out of concept as there exist too many beautiful files of mp3.

It's called Ab5 092 note rattle Steinway Prelude D.mp3

And its the Ab5 midi note 092 played manually on the keyboard. It's the first note heard. Then I just  test plays other notes around it, to make a comparison. This is at the stock preset Steinway D Prelude but exists also in K2 and others. All EQ and effects section turned off. I have a wav file too of it, if this isn't clear enough.

Turn it up and you can almost hear that there's someone in the background shaking a tambourine at the same time....some rattling. I can remove the file later on, I don't know how to attach mp3 files to individual posts.

I hear nothing (but I am old!)
A suggestion: could it be that in the room where you listen to this something like a music instrument or any object which can resonate is sympathetically resonating with this very frequency?

What you are hearing is exactly the duplex scale resonances. It is interesting to observe that you pinpointed one precise place where there is a jump in the duplex scale design, between G5 and Ab5. The duplex resonance on G5 appears at more or less 3.5 times the fundamental of G5 (hence around 6000 Hz) because the duplex string piece is 3.5 times shorter than the speaking part of the string, whereas on Ab5 it appears more or less at 2.5 times the fundamental of Ab5 (hence around 4300 Hz) because the duplex string piece is 2.5 times shorter. Hence it is more audible on Ab5 than on G5 (our ears having their maximal sensitivity in the 3000-4000 Hz range).

You can reduce the duplex scale resonances by simply moving its slider to the left in the Design panel (Standard/PRO).

Aaaahh, so you actually heard it? It wasn't that my hearing's playing games with me then. As opposed to the other one? OK, then I now know how to remedy this. Duplex scale slider to the left. Check!
I think it's an occupational hazard of me (sound engineering) for pinpointing out all those ugly sample jumps and sample points in old piano samplers... hang on...trying it now...

Yes, I can hear it clearly now, if dragged the slider of duplex scale to the far right, the notes adjacent to Ab5, (G and A) on either side, sounds the same, but when hitting that Ab5 again, the rattle is diferent and jumps out like a sore thumb, and anyone can hear it magnified. Now, dragging it to the left it disappears. I knew it was way more advanced than just to diss it some kind of "sample" artifact... thank you for solving this. But funny no one else detected this before as I heard it immediately. I do play that note/tone from time to time... :-D People seems to stay in C and white keys way too often..;-)

Sorry for my prickly ears, I hereby apologize... :-)

I can very well hear people go "Damm it Mats, now I hear it too just because of you...." :-D

There's nothing to apologize, you are perfectly right .

Note that this problem occurs also in the real life, and it is not rare to see concert piano tuners wedging a few felt pieces in the duplex scale to reduce some too pronounced resonance (following e.g. a request of the pianist). To be complete, I must add that these resonances which can be annoying when zooming on individually, are part of the global beauty and richness of the sound, particularly for the concert listener perspective (that is, not too close to the piano and with some nice acoustic reverb).

Hehe, that's the risk indeed

The duplex scales can be very different from one piano to another, particularly with the historical pianos which don't have a proper duplex scale, and where what is shown as duplex resonance in the UI are in fact global instrument resonances.

Exemplary hunch, question, suggestions, evidence and results Mats

Although I like pianos to exhibit realistic stresses, esp. maximum velocity. To me it's not a negative thing to be fixed per se - but depending always on intended output.

In the past my work was cut out, to make unrealistic MIDI piano sounds more real - it feels it's the other way around these days with Pianoteq. Pro gives us ways to surgically alter particulars of single notes or small ranges.

Finding I rarely hit or program 127 velocity now. With many General MIDI tools during the 90s I was sitting a lot of phrasing between 100 and 127 (bad piano sounds in a mix - dolling it up to work).. these days with Pianoteq, I can genuinely go way lower more often (sonic depth in the range) and only higher than 112 much more rarely. For a real pianist to hit a theoretical maximum velocity, Philippe and Pianoteq has this more in line to reality than other MIDI instruments I feel.

Good to see good conversations and ideas - thanks for that Mats.

@sigasa - Chris, that's really interesting - I do like how your exploration does give a good natural sense. Those edits do elicit a lot of great string sounds, shown to great effect. Fabulous!

Yes, I read that the duplex scale was invented or patended by Steinways in 1872 in connection with Helmholtz, so that one makes sense.

The duplex was originally more of an aliquot-like system of shorter strings between the pins and the agraffe that weren't muted with felt (though front and rear duplexes are now possible and rear are more common).  There's an 1872 patent for the non-speaking length of the string to be an octave higher than speaking length that is credited to Theodore Steinway and still used agraffes in the treble.  This was tuneable but a bit unwieldy and time-consuming (which is a lot of the instability and inconsistency that Philippe is talking about).  An aliquot stringing (used in Bluethners) was invented in 1873 which uses an additional aliquot string tuned higher to create different overtones, so there are four strings with only three struck by the hammer and still uses agraffes in the treble: it remains in use in their pianos today.  The capo d'astro bar was invented to replace agraffes and help regulate the tuning of the duplex scale and stabilize it, but on Steinways that largely forces it to stay the same tuning and basically eliminates the option of the front duplex, which was a trade-off of stability versus control.  The trend has also moved towards rear duplexes and muting the length between the pin block and the capo d'astro (a front duplex) and using adjustable (or non-adjustable) wedges near the static pins in the bridge.

Fazioli has an adjustable capo d'astro bar that allows a technician to retune the entire duplex scale to a pianists taste.  Mason and Hamelin returned to the original 1872 patent with a few changes to allow maximal control of duplex overtones--without the addition of the fourth string.  Borgato uses a fourth string, but it is struck by the hammer instead of sympathetically resonating.  And, last but not least, Stuart and Sons claim to have two duplex scales, one on the front end of the string and another at the opposite end, but they don't offer much explanation of how it works or how well it works.  (It would explain why the overtones of their pianos sound radically different from any other maker, and I'm not altogether sure that's a good thing.)

It's interesting that nobody has created a good solution for a rear bass duplex or--better yet--a controllable rear bass duplex.  I grant that the treble range of the piano offers a sympathetic bass duplex when the dampers are raised, but I'm not aware of any maker that has found a way (without making the piano 13 feet long) that doesn't require the rear duplex of bass strings to be muted--particularly as you could easily design something that utilized the second octave overtone instead of the first octave.  It might not sound good (which would be a good reason not to do it) or it creates greater harp tension, but it'd be intrigued about the results.  I should experiment in Pro with spectrum profiler to simulate a stronger bass duplex.

Well, I can make more mileage out of that note now, since X-mas's soon coming up around the corner. Isolated it sounds like one "jingle bell" if one can use cunning EQ-ing out the root note... so that will run in the background, and I'll play "Jingle Bells" in Ab all of the time.

There's always a use for some idiosyncrasy or anomaly...

thanks everyone....