4 ms·

Karplus Strong sounds kind of bad though, if your goal is accurate sounds of instruments. The reason it works at all is that the noise generates the initial hig

Karplus Strong sounds kind of bad though, if your goal is accurate sounds of instruments. The reason it works at all is that the noise generates the initial high harmonics of a pluck, then a low-pass simulates the frequency-dependent decay of those harmonics. But this is very primitive compared to a massive body of research on better simulation. The delay line works great though, that part hasn't changed.

cp3io 2y ago
The algorithm was discovered in 1983, and for how computationally cheap it is, I'd say it's pretty convincing. Waveguide synthesis is a generalization of the same idea, often with more realistic results. Variations on Karplus-Strong synthesis can yield some pretty wacky, dry and plucky types of sounds. In the original paper, the authors propose a variant of the algorithm which involves randomly alternating the sign (+/-) of every sample in the feedback loop, which creates a lot more atonal information in the signal. There’s also a variation floating around which uses an allpass filter in the feedback loop to introduce inharmonicity in the signal, see for example https://www.youtube.com/watch?v=SpC3t1EVF_Q https://www.youtube.com/watch?v=SpC3t1EVF_Q.