Roland VH-13 emits false bow hits on chicks

[Alek_A]

Hi! I've got Roland VH-13-MG hi-hat cymbals:

VH-13-MG.jpg

It is a pair of cymbals that mounts on ordinary hi-hat stand. When I press and release hi-hat pedal (do a chick), piezo trigger of the cymbal emits false hit.
It emits a hit before the chick and a hit or several hits after the chick. It's by design. When I set threshold higher the cymbal gets insensitive to low velocity hits.
Can we suppress these false hits somehow?
Maybe adding an option for chicks to take part in hi-hat suppression algorithm (retrigger mask) in settings of high hat pedal? (for false hits after a chick)
And also allowing to suppress bow hits when hi-hat pedal starts to get pressed quickly. (for false hits before a chick)
In example below I made four chicks.
Chick note - 57 (0x39)
Semi/half open bow hit note - 48 (0x30)
Open bow hit note - 51 (0x33)
First bow hit is played, just ignore it. Others are false hits. MIDI Log:

Chicks_Cause_False_Hits_Triggering.png

0.032 10 0x99 0x33 0x2c D# 3 Note On
0.198 10 0x99 0x33 0x00 D# 3 Note Off

1.639 10 0xb9 0x04 0x10 E -1 CC: Foot Controller
0.009 10 0xb9 0x04 0x20 E -1 CC: Foot Controller
0.006 10 0xb9 0x04 0x2f E -1 CC: Foot Controller
0.000 10 0x99 0x39 0x3a A 3 Note On
0.004 10 0x99 0x39 0x00 A 3 Note Off
0.000 10 0x99 0x30 0x05 C 3 Note On
0.139 10 0xb9 0x04 0x20 E -1 CC: Foot Controller
0.018 10 0xb9 0x04 0x10 E -1 CC: Foot Controller
0.015 10 0xb9 0x04 0x01 E -1 CC: Foot Controller
0.019 10 0x99 0x30 0x00 C 3 Note Off
0.026 10 0x99 0x33 0x01 D# 3 Note On
0.194 10 0x99 0x33 0x00 D# 3 Note Off

8.131 10 0xb9 0x04 0x10 E -1 CC: Foot Controller
0.001 10 0x99 0x30 0x02 C 3 Note On
0.003 10 0xb9 0x04 0x20 E -1 CC: Foot Controller
0.004 10 0xb9 0x04 0x2f E -1 CC: Foot Controller
0.000 10 0x99 0x30 0x00 C 3 Note Off
0.000 10 0x99 0x39 0x63 A 3 Note On
0.002 10 0x99 0x39 0x00 A 3 Note Off
0.000 10 0x99 0x30 0x06 C 3 Note On
0.000 10 0xb9 0x04 0x3e E -1 CC: Foot Controller
0.003 10 0xb9 0x04 0x4e E -1 CC: Foot Controller
0.126 10 0xb9 0x04 0x3e E -1 CC: Foot Controller
0.032 10 0xb9 0x04 0x2f E -1 CC: Foot Controller
0.019 10 0xb9 0x04 0x20 E -1 CC: Foot Controller
0.012 10 0xb9 0x04 0x10 E -1 CC: Foot Controller
0.005 10 0x99 0x30 0x00 C 3 Note Off
0.013 10 0xb9 0x04 0x01 E -1 CC: Foot Controller
0.044 10 0x99 0x33 0x01 D# 3 Note On
0.193 10 0x99 0x33 0x00 D# 3 Note Off

5.544 10 0xb9 0x04 0x10 E -1 CC: Foot Controller
0.003 10 0x99 0x30 0x03 C 3 Note On
0.002 10 0xb9 0x04 0x20 E -1 CC: Foot Controller
0.003 10 0xb9 0x04 0x2f E -1 CC: Foot Controller
0.000 10 0x99 0x30 0x00 C 3 Note Off
0.000 10 0x99 0x39 0x63 A 3 Note On
0.002 10 0xb9 0x04 0x3e E -1 CC: Foot Controller
0.001 10 0x99 0x39 0x00 A 3 Note Off
0.000 10 0x99 0x30 0x08 C 3 Note On
0.001 10 0xb9 0x04 0x4e E -1 CC: Foot Controller
0.003 10 0xb9 0x04 0x5d E -1 CC: Foot Controller
0.015 10 0xb9 0x04 0x4e E -1 CC: Foot Controller
0.020 10 0xb9 0x04 0x5d E -1 CC: Foot Controller
0.087 10 0xb9 0x04 0x4e E -1 CC: Foot Controller
0.007 10 0xb9 0x04 0x3e E -1 CC: Foot Controller
0.033 10 0xb9 0x04 0x2f E -1 CC: Foot Controller
0.024 10 0xb9 0x04 0x20 E -1 CC: Foot Controller
0.002 10 0x99 0x30 0x00 C 3 Note Off
0.011 10 0xb9 0x04 0x10 E -1 CC: Foot Controller
0.017 10 0xb9 0x04 0x01 E -1 CC: Foot Controller

4.784 10 0x99 0x33 0x03 D# 3 Note On
0.000 10 0xb9 0x04 0x10 E -1 CC: Foot Controller
0.002 10 0xb9 0x04 0x20 E -1 CC: Foot Controller
0.003 10 0xb9 0x04 0x2f E -1 CC: Foot Controller
0.000 10 0x99 0x33 0x00 D# 3 Note Off
0.000 10 0x99 0x39 0x73 A 3 Note On
0.001 10 0xb9 0x04 0x4e E -1 CC: Foot Controller
0.003 10 0xb9 0x04 0x6d E -1 CC: Foot Controller
0.001 10 0x99 0x39 0x00 A 3 Note Off
0.000 10 0x99 0x30 0x12 C 3 Note On
0.001 10 0xb9 0x04 0x7f E -1 CC: Foot Controller
0.012 10 0xb9 0x04 0x7b E -1 CC: Foot Controller
0.015 10 0xb9 0x04 0x7f E -1 CC: Foot Controller
0.076 10 0xb9 0x04 0x7b E -1 CC: Foot Controller
0.010 10 0xb9 0x04 0x6d E -1 CC: Foot Controller
0.006 10 0xb9 0x04 0x5d E -1 CC: Foot Controller
0.015 10 0xb9 0x04 0x4e E -1 CC: Foot Controller
0.008 10 0xb9 0x04 0x3e E -1 CC: Foot Controller
0.014 10 0xb9 0x04 0x2f E -1 CC: Foot Controller
0.014 10 0xb9 0x04 0x20 E -1 CC: Foot Controller
0.010 10 0xb9 0x04 0x10 E -1 CC: Foot Controller
0.011 10 0xb9 0x04 0x01 E -1 CC: Foot Controller
0.005 10 0x99 0x30 0x00 C 3 Note Off
0.015 10 0x99 0x33 0x01 D# 3 Note On
0.018 10 0x99 0x33 0x00 D# 3 Note Off
0.000 10 0x99 0x33 0x01 D# 3 Note On
0.197 10 0x99 0x33 0x00 D# 3 Note Off

[dmitri]

The process of chick/splash detection requires a delay. If cymbals hits to be filtered based on chick/splash detection then I will have to introduce a delay for hihat cymbal hits, i.e. latency. Do you think it is acceptable? If not, can you suggest an alternative?

[Alek_A]

Thanks for reply! I personally against delays! That's really what I wanted to ask next:)
I would suggest not to detect splashes at all. Instead when the pedal starts to get released quickly, we can send note off for chick and note on for splash.

The velocity of the splash will be the same as for chick if chick was made within short interval
and otherwise will be the velocity of chick multiplied by exp(-bt), where b - the damping coefficient, t - time elapsed after that interval. (damped oscillation)

Then we can add some sample offset for splash sound in the sampler. So the splash actually will be a chick + tail of splash sound.

[dmitri]

As far as I understand any subsequent cymbal (including HiHat cymbal) sound should cancel any previous sound from the same cymbal.
In this way, a chick or splash sound coming straight after a false hit sound should cancel this false (low velocity) hit sound immediately and will be masked by the much stronger (higher velocity) chick or splash sound.
Am I wrong?

[Alek_A]

You are right. The main problem is in the false hits after a chick. But I just thought that suppressing false hits makes recorded tracks clearer.

[dmitri]

So you want cymbal hits to be suppressed after a chick/splash?
I wonder why these false cymbal hits are happening AFTER a chick/splash. What is the MinScan on the cymbal? Aren't they from excessive stand vibrations?

[Alek_A]

That would be enough. They are happening firstly because upper cymbal hits lower cymbal. Probably chick detected before false hit. Also it happens when upper cymbal returns to its open position. Minscan is 20.

[Alek_A]

Dmitry, hi! How is it going? I've done some research and found out that there is a way for me to handle it in host! I tested one VST midi filter plugin and it appears that being a VST the plugin induces no latency! I downloaded SDK and developing midi plugin now. So unless this function is still important for someone or you about to finish it, it can be left as is, taking in account that the request is not trivial to fulfill. Thanks!

[dmitri]

It's good there is a workaround but I will try to fix it in firmware.

[Alek_A]

Great! Less I have to write myself!