Beta testing for a totally new sampling/triggering alg

[ignotus]

There were some issues with one of the previous beta versions but now it should work perfectly. Are you sure you've loaded the latest beta firmware?

Yes, I can confirm that it's the latest FW and I have that problem - chokes work fine but hits on the edge switch produce nothing, not even a bow note. As soon as I disable the new algorithm, it's back to normal.

[dmitri]

Can you please post your full config (zipped) here and tell me which inputs you're having the problem on?

[ignotus]

Here's my config file. The problem is with:

- Hi hat and ride cymbals (piezo/switch/switch 3 zone): can't get choke aftertouch, edge or bell note when hit. Edge and bell hits produce a bow note.
- Crash 1 and 2 (piezo/switch 2 zone): aftertouch works, but an edge hit produces nothing, not even a bow note. Bow works normally when hit.

[dmitri]

1. HiHat and Ride
The Threshold on the Edge should be set between ~60 and ~88. I set it to 74 for both and they both worked perfectly for all 3 zone with choking.
In the new algorithm Gain on Switch inputs has no affect as it is fixed to maximum gain level to have widest Threshold range on Edge Yamaha type switches (with 10k resistor). I may re-think if using hard coded max gain or revert it to as it used to be in the old algorithm

2. Crash1
Worked perfectly without modification (provided the switch on the edge is a pure switch without a 10k resistor)

3. Crash2
Edge hits were not producing any notes due to a combination of things (see below). After I set Threshold in the 3rd Zone to 0 (it was 122) it worked the same as Crash1.

4. Both Crash1 and Crash2.
For now the new algorithm doesn't process 'Disabled' in the 3rd Zone properly (bug, will fix it) so if it is a dual piezo/switch cymbal it is better to set all Notes there equal to Notes on Edge (except DampenedNote) and make sure Threshold on the 3rd is lower than Threshold on Edge.

[dmitri]

Update beta firmware in the first post.
Corrected behaviour when 3rd Zone is Disabled. Instead of incorrectly suppressing Edge hits it now sends Edge zone notes.

[ignotus]

I spent some time with the latest beta. Now all edge switches work without problems, thanks for that! After a fair few hours of tinkering, I've (just about) managed to get the kit performing the way I like it. I have to say that with the "normal" firmware, I had it dialled in really well and didn't have any annoyances like double triggering or crosstalk, so I suppose the most I could hope for was to get back to that performance, albeit with different settings.

What I've found much more difficult with the beta is zone separation and 3rd zone hits on piezo/piezo pads. This is the one thing I can't get working as well as before - I'm having a hard time finding the sweet spot to get well-separated head and rim, and also produce rimshots relatively easily hitting both at the same time. I used to be able to dial this in within 5 minutes. Yesterday I spent at least an hour on the snare and still couldn't get it right. It's either the head triggering 3rd zone notes or the rim triggering the head. I'll have another go at it today but any pointers are welcome.

Other than that just a couple of things I noticed:

- Upon booting, the MIDI led stays on until a pad is triggered. After that it behaves normally.
- 3rd zone hits produce green bars on the module's TFT screen instead of orange.

[dmitri]

Thank you very much for spending your time and testing it thoroghly!

My main focus so far has been on handling false double triggering without effecting fast rolls.
The zone separation for piezo/piezo pads in the new alg is preliminary and rudimentary yet. I will now work on improving it (and fixing the bugs you noticed).

[ignotus]

It's very good at getting rid of false triggering, I was able to lower all the related settings drastically. A question about the waveform graph on page 2: if we see a peak at, say 0.7 ms, does that mean minscan could, or should be lowered to around 10? I ask because my mesh snare and toms seem to consistently peak at below 1 ms on the head, and at around 3 ms on the rim. Would that result in better latency (at least on the head)?

[dmitri]

On a single zone pad you can lower MinScan to the peak time plus ~30%.
On dual piezo/piezo pads if both zones triggered (which is the case most of the time) it will continue to sample both inputs until the maximum MinScan of either of the zones reached. So most of the time the maximum MinScan of either of the zones will be used.

[dmitri]

Update beta firmware in the first post.
Dual piezo/piezo pads some bugs fixes and handling logic optimisation.
ignotus, can you please test it specifically with your piezo/piezo pads?