Hi there, Having given Karplus-Strong algorithms plenty of thought, esp. for my "Superclav" project: I'd say that if you want increased realism, you want subtle noise at all levels of the signal, but, a "just right/Goldilocks" amount. Too much, and it's more chaotic than would be on an acoustic instrument. Too little, and it's not naturalistic in the way you're searching for.
There's audio-rate noise (using white or pink noise as an excitation signal for the delay line) and then there's control-rate noise (slightly offset the pitch or decays, etc., per performance note 'event'). Final thought - there's nothing better than an actual acoustic instrument for realism! At a certain point, I feel like, if you want 100% realism, use the universe itself as a simulator - pure analog hardware! What's interesting to me personally about software physical modeling is being slightly idealed and slightly removed from physical reality.... BTW, there are plenty of videos of me playing Bach on the Superclav on my youtube channel. Here's the latest, just released within the last 48 hours or so: https://www.youtube.com/watch?v=3xQ9VRNaSic Aaron Krister Johnson Music, etc.: https://soundcloud.com/aaron-krister-johnson https://soundcloud.com/filtercreed https://www.youtube.com/channel/UC_utjGYbSizWE0dNyr0Vdmg https://aaronkristerjohnson.bandcamp.com/ http://www.untwelve.org Code: https://github.com/akjmicro <http://www.untwelve.org> On Thu, Sep 17, 2026 at 5:31 PM Mouldy Soul <[email protected]> wrote: > Hello, > > Been wondering about realism with karplus strong. > > Mistakes are a big part of playing, I'm wondering if there's a way to > emulate the subtleties of missing an ideally played note on a fret. There's > nuances to playing, slightly miss-plays of notes, slightly deadening, or > slight changes based on where you're holding the note down on a fret... > > And also how long a note is held down (even when note value is supposed to > be consistent/regular). We're talking about microscopic details here, that > aren't that easy to recognise, but nonetheless are present in playing. KP > has fixed decay. Ridiculous idea, but is there a way to emulate the > microscope randomness of actual playing? For instance notes don't always > decay perfectly when playing, you can have mutes, accidental or not (from > hand interference), that add to a cohesive realism. > > https://www.youtube.com/shorts/ZeWUC68nR1M > > I think that video is a decent example, when he starts playing bach. If > you were to emulate with karplus strong algorithm, you'd have a consistent > decay of a note at each play, but that's not the case in actual guitar > playing. I also think there's some subtleties of fretting, whether your > finger is bang on the middle, or slightly off/lopsided, I think this > changes something, I can't put my finger on it (pardon the pun), but seems > like there's minutiae that exists in guitar playing that isn't perfectly > emulated in digital algorithms yet (unless I'm sorely mistaken). > > My question to the group is, how do you start even quantifying/analysing > this phenomenon so you can model it? Is slightly adapting decay enough, is > slightly or randomly modifying tiniest of tuning of notes to account for > slightly off balance fretting enough to add an extra dimension of realism? > _______________________________________________ > Faudiostream-users mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/faudiostream-users >
_______________________________________________ Faudiostream-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/faudiostream-users
