MPS 750x

[3hhh]

Some test files as discussed.

[3hhh]

This should be ready now.

[3hhh]

On 11/13/23 19:50, Volker Fischer wrote: If you can set the rim shot threshold so that it only triggers (i.e., returns a "is a rim shot") for one of the instrument parts (rim or bell) and if you find another rim shot threshold where both instrument parts are detected correctly, then you can distinguish between them without having to use the MIDI velocity at all.

I tried to find such a threshold, but unfortunately both zones still triggered at a rim threshold of 31.

[corrados]

Hm, you could try to change the following line:

edrumulus/edrumulus.cpp

Line 512 in fe6b175

rim_switch_treshold = -pow ( 10.0f, pad_settings.rim_shot_treshold / 10.0f ); // rim switch linear threshold, where 10^(31/10)=1259 which is approx 4096/3 (10 bit ADC)

and introduce some offset:

rim_switch_treshold = -pow ( 10.0f, ( pad_settings.rim_shot_treshold - 15 ) / 10.0f );

which is introducing an offset of 15. If that is not enough, simply increase that offset until you are in range.

[corrados]

Sorry, you have to add the offset, not subtract it:

rim_switch_treshold = -pow ( 10.0f, ( pad_settings.rim_shot_treshold + 15 ) / 10.0f );

[3hhh]

I'll test that even if I still don't understand your idea. Anyway since the voltage divider current should hopefully be near-constant during the time of an active rim switch (and most importantly the rim switch causes a low input impedance which effectively creates a simple circuit with Rs, Rp and the input impedance shorted to GND in parallel to the piezo), the average voltage seen during the switching/"DC offset" should be slightly different for the different input impedances (edge vs bell). Possibly one could detect that.

[corrados]

the average voltage seen during the switching/"DC offset" should be slightly different for the different input impedances (edge vs bell). Possibly one could detect that.

That's exactly what I would like to find out. And I hope we can do that by simply introducing a second threshold. But first we (i.e. actually you :-) ) have to find out if it works by changing the one threshold we currently have available to see if it could work.

[corrados]

Thanks, I'll merge it now.