Topic by lowendtheory

  • Avatar
  • lowendtheory
  • Member
  • Offline
  • Registered: 28-12-2012
  • Posts: 97

Topic: How to restore L-mode "brightness" at extreme string lengths (copper)

I've been playing with PT5 and like how the copper-wound strings behave when the length is changed - namely the L-mode changing pitch. Everything seems fine, except for one thing; the voicing for MIDI notes #30 to #21 change in a way that doesn't sound "right", as if someone put a notch filter at the 800-1500hz region (at the L-mode partial) when length exceeds 3.7m, and the attack also suffers heavily in that note range.

Would there be any way I can "restore" some of the brightness lost by increasing the string length, specifically the L-mode partial? I've heard recordings of the Rubenstein R-371 and have the Klavins 370 as a VST, and the L-modes there are just as audible in that note range (i.e. I can discern their pitch) as on a Steinway D. Note that at that string (case) length, the extra notes sound just fine, even scaling up to the case length of the longest piano (5.7m). However, that "color" the L-mode gives to that note range (#21-30) is almost nonexistent even at 3.7m, though the attack sounds okay.


Can someone help explain the unusually quick decay of the L-mode as described above? My suspicion is that it has something to do with actual spectral data being present, which causes the behavior above, but not for the notes which were interpolated or fabricated (below #21). Is this the case, or is it something different?

Last edited by lowendtheory (07-09-2014 03:53)

Reply by Philippe Guillaume

  • Avatar
  • Philippe Guillaume
  • Admin
  • Offline
  • Registered: 25-04-2007
  • Posts: 5,405

Re: How to restore L-mode "brightness" at extreme string lengths (copper)

Interesting question indeed. As you mentioned the extra notes, I guess you are speaking about the D4 or the K2 which are extended down to note #9 (one octave below the usual lowest A).

You are right: in the current version, the L-modes have been voluntarily tamed down when the length is not the default length (although there is no particular treatment of the L-modes for the extra notes). This was a pure aesthetic choice because we found that the L-modes sounded a bit weird when their pitch was "unusual" in the sense that our ears are not used to it.

It is worth noting that even at the default length, L-modes have often been judged undesirable: some piano manufacturers have tried to reduce them with more or less success, and when we released version 5, some people complained about the L-modes being too loud, even at the default length, so we had to reduce them slightly (see changes in version 5.0.1 following user requests - cf. Changes tab on page https://www.pianoteq.com/pianoteq5).

Nevertheless, we are currently preparing a Pianoteq update (version 5.1) and we could indeed increase the volume of the L-modes in the D4 if you think it's worth.

Philippe Guillaume's Website

Reply by lowendtheory

  • Avatar
  • lowendtheory
  • Member
  • Offline
  • Registered: 28-12-2012
  • Posts: 97

Re: How to restore L-mode "brightness" at extreme string lengths (copper)

As much as I'd like that request to be fulfilled, I find it too much of an imposition against the rest of the customer base changing the hardcoded values of the L-mode partials at the desire of only a few people. Maybe when you find a way to allow easy customization of that aspect (for PRO version), such a change will be easier to accept.

Reply by lowendtheory

  • Avatar
  • lowendtheory
  • Member
  • Offline
  • Registered: 28-12-2012
  • Posts: 97

Re: How to restore L-mode "brightness" at extreme string lengths (copper)

However, I will post a poll asking the community's input, so be sure to read the topic when it's posted.
There will be single note samples of the Klavins 370 to compare against the D4 at case length 3.7m.

Reply by Jake Johnson

  • Avatar
  • Jake Johnson
  • Member
  • Offline
  • From: Atlanta
  • Registered: 04-05-2007
  • Posts: 4,501

Re: How to restore L-mode "brightness" at extreme string lengths (copper)

lowendtheory wrote:

I've been playing with PT5 and like how the copper-wound strings behave when the length is changed - namely the L-mode changing pitch. Everything seems fine, except for one thing; the voicing for MIDI notes #30 to #21 change in a way that doesn't sound "right", as if someone put a notch filter at the 800-1500hz region (at the L-mode partial) when length exceeds 3.7m, and the attack also suffers heavily in that note range.

Would there be any way I can "restore" some of the brightness lost by increasing the string length, specifically the L-mode partial? I've heard recordings of the Rubenstein R-371 and have the Klavins 370 as a VST, and the L-modes there are just as audible in that note range (i.e. I can discern their pitch) as on a Steinway D. Note that at that string (case) length, the extra notes sound just fine, even scaling up to the case length of the longest piano (5.7m). However, that "color" the L-mode gives to that note range (#21-30) is almost nonexistent even at 3.7m, though the attack sounds okay.


Can someone help explain the unusually quick decay of the L-mode as described above? My suspicion is that it has something to do with actual spectral data being present, which causes the behavior above, but not for the notes which were interpolated or fabricated (below #21). Is this the case, or is it something different?

Well, I'm not sure about what aspect of the L-mode sound to worry about. Is it the amplitude of the LMode freqs or the degree to which they exist (if these are separate), or their relation to velocity, or the stages of their amp envelope--their amplitude at each stage? Or the way their amp stages react to velocity? Sometimes I think I hear that they should be loudish, but should die down a little earlier, so that they are more part of the attack and earliest decay than the later decays. I do want to be able to control them more. Not sure how.

Reply by lowendtheory

  • Avatar
  • lowendtheory
  • Member
  • Offline
  • Registered: 28-12-2012
  • Posts: 97

Re: How to restore L-mode "brightness" at extreme string lengths (copper)

Jake Johnson wrote:

Well, I'm not sure about what aspect of the L-mode sound to worry about. Is it the amplitude of the LMode freqs or the degree to which they exist (if these are separate), or their relation to velocity, or the stages of their amp envelope--their amplitude at each stage? Or the way their amp stages react to velocity? Sometimes I think I hear that they should be loudish, but should die down a little earlier, so that they are more part of the attack and earliest decay than the later decays. I do want to be able to control them more. Not sure how.


In this case, I'm asking the L-modes to behave as they do for the extra notes. Set the string (case) length to 3.71m and listen to the abrupt change in tonal quality going below MIDI #21.

To put a finer point on the issue, here is the D4 at 3.71m case length and the Klavins 370. MIDI #23 and MIDI #29 were used as control notes to best illustrate the differences between them. Note that even at maximum velocity, the L-mode cannot be heard clearly compared to the 370 at velocity 80 at MIDI #23. Both renders of MIDI #29 used 127 velocity.


The attack is a different issue and is best solved with a new piano model based on the Klavins 370.

Reply by Modellingoptimist

  • Avatar
  • Modellingoptimist
  • Member
  • Offline
  • Registered: 11-12-2013
  • Posts: 262

Re: How to restore L-mode "brightness" at extreme string lengths (copper)

lowendtheory wrote:

However, I will post a poll asking the community's input, so be sure to read the topic when it's posted.
There will be single note samples of the Klavins 370 to compare against the D4 at case length 3.7m.

Have I missed something?

formerly known as Notyetconvinced

Reply by lowendtheory

  • Avatar
  • lowendtheory
  • Member
  • Offline
  • Registered: 28-12-2012
  • Posts: 97

Re: How to restore L-mode "brightness" at extreme string lengths (copper)

Completely forgot! Post is up now.

Reply by Pianotrancer

  • Avatar
  • Pianotrancer
  • Member
  • Offline
  • From: Harrow, England
  • Registered: 07-08-2007
  • Posts: 138

Re: How to restore L-mode "brightness" at extreme string lengths (copper)

Philippe Guillaume wrote:

Nevertheless, we are currently preparing a Pianoteq update (version 5.1) and we could indeed increase the volume of the L-modes in the D4 if you think it's worth.


I would love to be able to increase the volume of the L-modes. Perhaps a graduated slider?

Reply by Jake Johnson

  • Avatar
  • Jake Johnson
  • Member
  • Offline
  • From: Atlanta
  • Registered: 04-05-2007
  • Posts: 4,501

Re: How to restore L-mode "brightness" at extreme string lengths (copper)

Pianotrancer wrote:
Philippe Guillaume wrote:

Nevertheless, we are currently preparing a Pianoteq update (version 5.1) and we could indeed increase the volume of the L-modes in the D4 if you think it's worth.


I would love to be able to increase the volume of the L-modes. Perhaps a graduated slider?

I would like this too, but I worry, as I think others have worried:

1. Would it be best to control the amplitude of the L-modes or to control what causes the increased amplitude of the L-modes, which might also have other effects? (The string material? I don't really know what controls the amplitude of the L-modes.) Does their amplitude have more to do with mic closeness--the mics should pick up more of the L-modes when closer (and when using the headphone mode, the head should hear the L-modes as louder when moved closer), but they may not be more audible than they are now from a distance?

2. Could their amplitude be controlled string by string\note by note? Seems as though that would be necessary, which makes it not so simple to do. Would their amp envelope differ on different strings, as well?

3. What happens with unisons? Would it be accurate to assume that the L-mode freqs are exactly the same on unison strings that are tuned to the same freq, or is there some slight variation simply because no two strings are exactly the same? Even if they are the same freq, would they change in a linear way as the unisons are detuned? And what would coupling do--force the l-modes into almost the same freqs? What would be the effect of changing the Unison balance?

4. Do they react linearly to force\velocity? Are there upper limits to how loud they are on each string?

Of course these are questions that can be answered, but only with time. But yes, I hope it can be done.

Last edited by Jake Johnson (22-09-2014 02:30)