Halfway hihat pedal positions suppress hits of rolls

[anttipi]

Specifically this one: https://www.bitsbox.co.uk/index.php?main_page=product_info&cPath=302_305&products_id=2083&zenid=8aakgdr3oqmeln8ojvgo4243t1

Yes, that's the one. I just ordered a pair of those from Ebay.

Do I understand it right that you don't have a problem with lost notes(apparent rolls suppression) when the pedal is half pressed and you're not going to test it over standard MIDI?

Hmm, ok, so transferring via standard MIDI versus USB-MIDI from Megadrum to PC (and Megadrum Manager's MIDI log) might make a difference?

Well, I tested it and changed MDM's input to standard MIDI but the readings on the MIDI log (graphical and raw data) are similar. Hits get lost the same way because of the CCs bombarding Megadrum every 10-20 ms. But anyway, the missed hits were visible already on MD's own LCD graph too.

So to recap :
1. My hall sensor is likely too sensitive - better not buy 5 mV/g hall sensors (my sensor doesn't seem to be broken but is just very, uh, sensitive!)
2. Massive CC torrent overtaxed Megadrum, causing hits being missed in rolls and overall an occasional "tacky" playing feel, also probably was causing the spontaneous aftertouch messages (can't blame Megadrum because at worst there can be ~100 messages per second)
3. CC reduction behaves differently on old vs. new pedal algorithm. But that doesn't actually matter because the hihat controller shouldn't be so super sensitive in the first place. In other words, Megadrum's CC reduction level 3 is strong enough and if that's still not sufficient, you will have to swap the controller sensor. I also think that even though the CCs are reduced, it is still taxing Megadrum unnecessarily because filtering data a hundred times a second must require a bit of processing power (just a wild guess!)

[maru_engineering]

Hi all!

Allow me to jump in... From what I had seen it's a problem if you try to use the New Algorithm for the HiHat Pedal in conjunction with the Original Sampling Algorithm (AltSamplingAlg from the Misc section unchecked). In this configuration the HiHat Pedal singal from the Visual MIDI Log is "jumpy".

Could anybody confirm if this is correct?

PS
If you use AltSampling Alg with the New Algorithm it's ok...but maybe there is a problem with the AltSamplingAlg... with a real drummer I had experienced drop-outs on the snare when using positional sensing and a busy beat involving Kick, Hihat and Snare hits on the same time... I will try to look deepen into this but it's difficult because I cannot play drums on my own...

[anttipi]

Just an update: I'm totally at my wit's end! After waiting for over a month, I finally got the new less sensitive hall sensor(s) from Hong Kong. After soldering everything, to my surprise and horror the exact same problem persists!

I have NO idea what's happening here. Along the way I've tried re-soldering, changing the cable, changing the hall sensor to a model known to work, testing the hall sensor away from my hihat stand, swapping in different strength magnets... Still horrible jumping on both Atmega and ARM. Completely unusable with the new pedal algorithm even with CC suppresssion at 3.

Allow me to jump in... From what I had seen it's a problem if you try to use the New Algorithm for the HiHat Pedal in conjunction with the Original Sampling Algorithm (AltSamplingAlg from the Misc section unchecked). In this configuration the HiHat Pedal singal from the Visual MIDI Log is "jumpy".

It's jumpy with either algorithm. Just less jumpy with the old pedal algorithm (and CC reduction, without which the hihat is completely unusable).

[dmitri]

Do you have an voltmeter? Can you read voltage on the signal wire when the pedal is connected?

[anttipi]

As it's been a while, I re-tested stuff and it appears Atmega does work fine with the hall sensor, contrary to what I remembered and wrote above. Even at CC reduction 0, the CC messages stop as soon as the hihat pedal settles - occassionally it may take 1-2 seconds at max to settle.

As I also tried 2 different PCs and two different USB cables, it can only be the ARM Megadrum that is the cause of the issue.

Do you have an voltmeter? Can you read voltage on the signal wire when the pedal is connected?

I can borrow a high quality one from my brother (who is an electrical engineer) and also an oscilloscope, if needed.

[dmitri]

As it's been a while, I re-tested stuff and it appears Atmega does work fine with the hall sensor, contrary to what I remembered and wrote above. Even at CC reduction 0, the CC messages stop as soon as the hihat pedal settles - occassionally it may take 1-2 seconds at max to settle.

As I also tried 2 different PCs and two different USB cables, it can only be the ARM Megadrum that is the cause of the issue.

Do you have an voltmeter? Can you read voltage on the signal wire when the pedal is connected?

I can borrow a high quality one from my brother (who is an electrical engineer) and also an oscilloscope, if needed.

Yes, please do. And also do the tests with CC Reduction Level set to 3 and post screenshots of Raw MIDI as you did with CC Reduction at 0.

[dmitri]

To make sure MIDI CC causes the problem set HiHat pedal mode into Pot and see if you experience the same problem when the pedal is half way down.

Btw, did you test this?

[anttipi]

And also do the tests with CC Reduction Level set to 3 and post screenshots of Raw MIDI as you did with CC Reduction at 0.

Here are 3 screenshots, all with CC reduction at 3 and pedal almost fully open (124).

1. New algorithm enabled, you can see an example of the occasional note on/off messages too.

CCreduction3_NewAlg.png

2. New alg enabled, an example of occasional and spontaneous aftertouch messages included.

CCreduction3_NewAlg(2).png

3. New algorithm DISabled, CC reduction still at 3. CC suppression works a little better but is still "firing" pretty constantly.

CCreduction3_OldAlg.png

Btw, did you test this?

I did, and just now retested with the 2nd hall sensor. The hihat openness bar and the values on MD's LCD are just as jumpy as with the foot controller option but the snare rolls are triggered fine (because the CC bombardment is not happening when pot is selected).

My brother, without seeing it in person yet, suspected something is causing interference or something within MD (a soldering issue for instance?). I guess we'll just have to measure stuff. I don't have a clue how I could figure it out or solve it on my own so I'll just need to be patient and wait for my brother's help.

[dmitri]

How did you set Low and High values?
When slowly closing the pedal, does the pedal VU meter goes down smoothly and all the way from fully open to fully closed?

[anttipi]

How did you set Low and High values?
When slowly closing the pedal, does the pedal VU meter goes down smoothly and all the way from fully open to fully closed?

The thread title is a bit misleading: it doesn't really matter what position the pedal is in. The values jump up and down as soon as the pedal cable is connected, regardless of the high and low settings. But to answer your question: high is just below the max value when hihat is fully open (around 500) and low around the min value when pedal is pressed (about 350 with the new hall sensor, lower with the old one). In short, I'm following the standard procedure.

The pedal VU meter moves across the whole range and in pretty much a linear fashion as it should with a linear hall sensor. However, I'd hesitate to call it a "smooth" progression because of the interference or noise or whatever the correct term is (the up and down "spikes" that cause the CC flood).