www.megadrum.info

Hi

I'm following these steps (https://www.megadrum.info/forums/viewtopic.php?p=20543#p20543) to configure a DIY 14" snare pad. The result is satysfying (good balance between false notes suppression and good fast rolls) for me, but only if I configure it as a single zone pad. Problems emerge when I want to incorporate more advanced features: trigger side stick with the second zone and rimshots with a 3rd zone controls.

I feel like changing DualMidWidth affects dynamic threshold and breaks my false notes suppression. Manual doesn't explain relations between two zones of the same pad so I'm using the trial and error method but it feels like "voodoo programming". If the result is a bit better - I still don't know why, so I can't take it to the next level.

dmitri, could you write at least some pseudocode algorithm of MegaDrum triggering procedure for dual zone pads (with positional sensing and 3rd zone enabled)?

If not, these are some questions I'd like to ask, containing my educated guesses about this procedure.

1. AltSamplingAlg for STM32 reduces digital noise and allows lower threshold values. Does it have any side effects affecting triggering? Should I use it by default on STM32?
2. Should MinScan be a multiply of Latency?
3. Should I use only HighLevelAuto to determine HighLevel value, or is it better to increase it manually to the real limit.
4. "DynTime, DynLevel, MinScan should be set to close values between Head and Rim on dual zone pads" (https://www.megadrum.info/forums/viewtopic.php?f=4&t=3012#p30739) - i set the same values on both zones, but I feel like I could variate it a bit in my favor if I understood how they relate to each other. I believe Retrigger value should also be the same.
5. Which is better: set up zones separately and then join them with DualMidPoint and DualMidWidth, or join them first and then try to eliminate false notes?
6. What is the real difference between 1, 2 and 3 positional sensing algorithms, apart of having to set PositionalLow and PositionalHi higher values?

1. AltSamplingAlg is not tested as much as the standard sampling algorithm and people reported issues with it.
2. MinScan should be identical between both zones. The bigger mesh pad the bigger MinScan should be. For a 14" snare it should be at least 60.
3. HighLevelAuto is used to get rough HighLevel values. You should disable it after getting a rough HighLevel value and then adjust HighLevel so that only hardest hits produce maximum velocity with a Linear Curve.
4. DynTime, DynLevel, MinScan should really be close between zones but they of course allow variation.
5. For piezo/piezo pads set the zones separately for best results on each zone individually and then join them together. For piezo/switch pads you of course have to join zones first.
6. Different positional sensing algorithm use slightly different maths so the results are slightly different. For some pads one algorithm is best - for other pads another algorithm is better.

Hi!

Thank you for your guidlines. I think increasing threshold and tension of the head helped as well. I'm satisfied with my configuration. It gets like 0.1% false notes (mostly with strong flammed rimshots) and it doesn't trigger very soft hits in the area outside of the center of the drum, but tackling these issues makes it unplayable. Maybe it's related to the mesh head itself (2-ply), which rattles at least two times longer than a drum-tec real feel (but I don't have a 14" one yet to compare directly).

Still, I'd like to know does it matter for MinScan to be a multiple of Latency value.

And for the Retrigger setting - I use a value of 2, but in my experience setting it higher (to lower DynLevel a little bit more) wouldn't make a difference or made it even worse. I remember you wrote you set it 4 to 12 on DIY pads, but I can't find this post at the moment. I believe MegaDrum works as presented in the picture and Dynamic Threshold starts working after Retrigger Mask time passed, but I experience a real difference only between 1 and 2, and maybe 2 and 20, when fast rolls are impossible.

MinScan does not have to be multiple of Latency at all.

The picture is mostly correct provided it is understood that neither DynLevel nor DynTime are fixed - they depend on the strength of a hit and DynLevel/TynTime settings.