In piano a sting is typically dampened (safe for pedal use) from one note to another that sound banks with a large number of individual recordings are used in commercial quality piano synths.

I do wonder however it what is a feasible approach for guitars though, particularly in cases where the strings are re-excited pretty quickly such as in a classical piece like Recuerdos de la Alhambra. As e.g. Wikipedia says about this piece:

The piece showcases a challenging guitar tremolo, wherein a single melody note is plucked consecutively by the ring, middle and index fingers in such rapid succession that the result is an illusion of one long sustained note. The thumb plays an arpeggio-pattern accompaniment simultaneously. Many who have heard the piece but not seen it performed mistake it for a duet.

Is this a negligible aspect, or does it have to be taken into account in a guitar synth that tries to close the physical model in sound reproduction?

I'd appreciate suggestions for further reading in the literature on this topic as well, e.g. research papers or book chapters dealing with this issue.

To clarify, I know something on a general level about Karpus-Strong, in the sense of "I heard of that". I was wondering about this specific aspect when the strings are struck again while still vibrating audibly enough. Is it a non-negligible thing to consider in a synthesizer?

I'm not entirely certain this is the right SE site to ask this on, by the way. If there's a more suitable one, please suggest it.

  • There have been [functional, real-time] mathematical models for strings since the mid 90s. Most large leaps in the modelling have been rapidly wallpapered in patents & kept close to the chest ever since. Are you just looking for the ones that were released into the wild of open source? [I used to work for one of the instrument majors & still under NDA, so there's a lot I'm not at liberty to say, even to this day.]
    – Tetsujin
    Dec 19, 2021 at 16:50
  • @Tetsujin: yes I know enough on a general level about Karpus-Strong etc. I was wondering about this specific aspect when the strings are struck again while still vibrating audibly enough.
    – Fizz
    Dec 19, 2021 at 16:58
  • 1
    tbh, though i worked 'in' this for a decade, I wasn't on the 'scientific' end of the structure, so my knowledge of actual methods & models is very limited. I mainly wanted to convey how secretive this entire genre can be. Re-excitement of a currently-playing string was originally done early 2000s, so it might be out there somewhere. Since I left the field, excitement of [unrelated string] sympathetic resonances was also added to the model. I know this can be done [these days in high-end consumer-grade products], I just don't know precisely how, nor the patent status on the various models.
    – Tetsujin
    Dec 19, 2021 at 17:02
  • Just for a bit of 'fun' historical background - I was actually in the meeting [5 of us] where the very first convolution software [non-realtime, we went for lunch whilst it calculated] was presented. Also, sadly for its inventor, the lawyers got straight onto it & decided [proven to be correct by history] it could not be patented, so they made their own. Such is business, I'm afraid. String/tube modelling was earlier, first worked on that in perhaps 92, but the first products weren't seen til a few years later.
    – Tetsujin
    Dec 19, 2021 at 17:06
  • I know of the even earlier Karpus Strong workings, but idk how theirs relates to what I worked on. Ours was entirely mathematical & didn't rely on any 'sampling' excitation method.
    – Tetsujin
    Dec 19, 2021 at 17:13


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Browse other questions tagged or ask your own question.