Yop,

this topic http://forum.pdpatchrepo.info/topic/6559/samvada-ios-sympathetic-strings-app gave me the taste for a sympathetic sitar strings box.

I made a Samvada-like patch, with the info given in the post and in links pointed to by the post.
The seed frequency has been choosen to be similar to the one that can be seen in the demo video of the app.
The effect of the allpass dispersion filter is very subtle (I'm not sure to hear it).
I get many sigsend and receive~ errors due to the feedback loop under 64 samples, but it seems to do the job anyway.
The basic excitation method I used to "pluck" the strings doesn't work for highest notes, the resonance occurs normally though.
Sorry for the many multiplication constants used between "stages", it just came like that . It works well with my violin, would be glad to hear if other instruments need improvement in the levels in the patch.
The "jawari" feedback between output and delay length was fixed to 4%.

Thank you again for the inspiration, Iotic !

Nau

http://www.pdpatchrepo.info/hurleur/dronestring_testbench_lehurleur.pd

This is so much fun!

Thanks

I'm getting pairs of error messages as soon as I turn the DSP on. The message is:
receive~ 1019-fb: vector size mismatch
sigsend 1019-fb: unexpected vector size

Cmd clicking on the first a [r~ $0-fb] object inside the dronestring~.pd abstraction is highlighted, but doing the same on the other error message nothing really happens...

Thanks Alexandros,

I already mentionned these errors, I don't think there's a workaround but the patch still works...

Nau

Hi there,

this testbench needed a decent gui. Now notes can be selected in a graphical way, strings can be muted separately, 10 presets can be saved and recalled.

This is a testbench, and is made for testing a lightweighted implementation of the resonators, with a portable device state of mind. Many problems are still there, amongst them

• string plucking still doesn't work for high pitches, but these strings seem to resonate properly with the input sound anyway
• I have the intuition that my way of implementing the string is not optimal, and globally the sound of these strings taken could be better, and control parameters could act in a more efficient way with an better implementation (still lightweigthted)
• every string resonates only with the input sound, but they aren't linked together, so they don't resonate with each other. I could have a look at this, but I'd preferably try to get a better string implementation before coupling them
• low string notes have very long decays with respect to high pitched strings. In a certain way this is logical, but the effect is as strong that I don't use notes within the two first octaves allowed in the gui
• I do perceive the dispersion effect now (educated my ears, didn't change the implementation at all ), but I wonder if the coefficients used to make the computations are really adapted to sitar strings, as the article they come from tries to emulate piano strings
• the abovementionned errors are still there, but the patch works anyway

So... comments welcome.

Nau

http://www.pdpatchrepo.info/hurleur/sympathetic_strings_testbench.zip