Topic by daniel_r328

  • Avatar
  • daniel_r328
  • Member
  • Offline
  • Registered: 27-07-2025
  • Posts: 64

Topic: Metallic rattle on some Steinway D NY notes: bug or feature?

Good morning Modartt team and community,

I've noticed something in the v9 Steinway D NY that I wanted to get some clarification on. Playing the notes around A5 (aka A'') at fortissimo, I hear a distinct, separate, dry, metallic rattle at high pitch - similar to what a vibrating string hitting a hard surface would sound like (sounding more like a multi-frequency noise than a clean resonance). I reproduced this on different audio chains and different presets, so am confident that it's in the model rather than a digital artefact. It's also present in the HB models (albeit fainter), and absent in the other models I've tried.

To Pianoteq's credit, this sounds very much like something an acoustic instrument would produce, so I wouldn't call it a simulation flaw. But I would call it a potentially undesirable characteristic in an acoustic instrument and am curious if it's intentional, and/or if I can tweak this effect away somehow?

I'm not clear on what would happen in an acoustic instrument to cause that sound, so unsure which part of the virtual design influences it.

Grateful for your thoughts!

Reply by Philippe Guillaume

  • Avatar
  • Philippe Guillaume
  • Admin
  • Offline
  • Registered: 25-04-2007
  • Posts: 5,460

Re: Metallic rattle on some Steinway D NY notes: bug or feature?

daniel_r328 wrote:

Good morning Modartt team and community,

I've noticed something in the v9 Steinway D NY that I wanted to get some clarification on. Playing the notes around A5 (aka A'') at fortissimo, I hear a distinct, separate, dry, metallic rattle at high pitch - similar to what a vibrating string hitting a hard surface would sound like (sounding more like a multi-frequency noise than a clean resonance). I reproduced this on different audio chains and different presets, so am confident that it's in the model rather than a digital artefact. It's also present in the HB models (albeit fainter), and absent in the other models I've tried.

To Pianoteq's credit, this sounds very much like something an acoustic instrument would produce, so I wouldn't call it a simulation flaw. But I would call it a potentially undesirable characteristic in an acoustic instrument and am curious if it's intentional, and/or if I can tweak this effect away somehow?

I'm not clear on what would happen in an acoustic instrument to cause that sound, so unsure which part of the virtual design influences it.

Grateful for your thoughts!

That may come from the duplex scale (which is indeed quite present at fff), have a try reducing it.

Philippe Guillaume's Website

Reply by daniel_r328

  • Avatar
  • daniel_r328
  • Member
  • Offline
  • Registered: 27-07-2025
  • Posts: 64

Re: Metallic rattle on some Steinway D NY notes: bug or feature?

Philippe Guillaume wrote:

That may come from the duplex scale (which is indeed quite present at fff), have a try reducing it.

Thanks! I've used this as a starting point to dig deeper:

Setting sympathetic resonance and duplex to zero, reduce the general shimmer of the note a bit, but the specific effect I hear is still there. However, Setting the impedance slope to a flat (0.2) value and shortening the string length to its minimum strengthens the effect - it's now auditable as a bell-like sound in all notes but particularly strongly emphasised around A5. So it looks like it's effectively a hump in the frequency response of the soundboard: the short strings excite a wider range of frequencies and the soundboard emphasises (overemphasises?) some of them.

This may very well be an accurate reflection of how the reference instrument behaves, so it's really down to personal preference. It makes me wish for a few more user-accessible parameters for the excellent new soundboard model, though. I'm thinking that something that compresses the peaks and dips of the response curve towards a floating average, could act as a parameter to make the model more neutral-vs-characterful (I'm making a lot of inferences here, I appreciate).

Anyway, for my needs I should be able to tweak string length and maybe impedance to bring this under control - thank you again

Reply by Philippe Guillaume

  • Avatar
  • Philippe Guillaume
  • Admin
  • Offline
  • Registered: 25-04-2007
  • Posts: 5,460

Re: Metallic rattle on some Steinway D NY notes: bug or feature?

daniel_r328 wrote:
Philippe Guillaume wrote:

That may come from the duplex scale (which is indeed quite present at fff), have a try reducing it.

Thanks! I've used this as a starting point to dig deeper:

Setting sympathetic resonance and duplex to zero, reduce the general shimmer of the note a bit, but the specific effect I hear is still there. However, Setting the impedance slope to a flat (0.2) value and shortening the string length to its minimum strengthens the effect - it's now auditable as a bell-like sound in all notes but particularly strongly emphasised around A5. So it looks like it's effectively a hump in the frequency response of the soundboard: the short strings excite a wider range of frequencies and the soundboard emphasises (overemphasises?) some of them.

This may very well be an accurate reflection of how the reference instrument behaves, so it's really down to personal preference. It makes me wish for a few more user-accessible parameters for the excellent new soundboard model, though. I'm thinking that something that compresses the peaks and dips of the response curve towards a floating average, could act as a parameter to make the model more neutral-vs-characterful (I'm making a lot of inferences here, I appreciate).

Anyway, for my needs I should be able to tweak string length and maybe impedance to bring this under control - thank you again

Yes you can proceed as you describe, there's just one thing I would not change: the string length, as it really provides the signature of the original instrument. Of course if you don't care about that, then it is not an issue changing it, to another piano hence.

Philippe Guillaume's Website