Hihat - dynamic chick - firmware testing

[gabriel1712]

Though I could imagine that simply using distance, the CC range traveled, would be the neatest implementation, disregarding speed completely.

I believe speed play a crucial part too, thou minor compared to distance. Pressure is imo best simulated by distance AND speed.


Long__(0)_1_2_3
_____________2
_____________1
Short__0_____(0)
_____Slow____Fast

Hope you can deciffer my home cooked graph
Speed horizontal, distance vertical.
0=zero vel
(0)= 1-10 vel
1-3 = 10-127 vel
Distance is the main factor determing MidiVelocity, with speed merely shifting the curve.

Without speed, a slow open HH stroke would end with an unnatural 80-127 vel chick, because you have to open HH passed the ½ and even the 3/4 distance mark to get the open stroke. On an acustic HH set you can choose the chick strenght by playing heal up or heal down foot on the open hat stroke. Only speed will pick that up, not distance.

Short distance (few mm) should return 0 to (0) depending on speed.

Would be great with different curves available and compression/shift as well. HH is by far the most personalized item in a drumset. It's the first thing drummers mess with on a foreign set.

Perhaps this discussion should live in a dedicated thread. Playing style/tecnique differs greatly so we may overlook somth obvious.

[userfriendly]

Hmm. Okay, you convinced me. Speed does matter too.

[dmitri]

I've just released a new firmware with changes which deal with HH pedal handling and the chick handling in particular. Descriptions for 4 new settings are at http://www.megadrum.info/content/hihat-pedal-settings

[gabriel1712]

Looking forward to testing those tonight

[gabriel1712]

Timewise, I'm a little stretched at the moment, so I wouldn't be able to fully test the new HH feature untill saturday. But from what I have observed so far, it looks VERY promising. There may be a point discussing the nature of splash triggering, since MD HH controller is apparently now becoming a precision instrument I'll give it a good work-out and get back with data and thoughts in the w/e.

Unrelated but not irrelevant. I can't seem to make MCT <-> MD send configs to each other. MCTs midi log is showing MD data (after using Gastrics methode of chosing a random IN/OUT + apply in MCT config, then chosing the correct IN/OUT + apply). I notice, the default SysEx ChainID in MD is 254 (should be 0, right?) but correcting it doesn't solve the config. communication prob. And after rebooting MD, SysEx ChainID is back at 254 even though 0 was saved w/ SaveConfig.

EDIT: Fixed the MD <-> MCT communication problem by reinstalling the firmware

[gabriel1712]

Testing the new Hihat controller options (from firmware 20090504 and forward)
OpenLvl/CloseLvl is an excellent addition to the MD. Mine is set to 32/103 and that providers exactly what I want from my HH in terms of movement.

Chick
Testing the chick I found two things.
a) I really like the idea of two scales for long and short chicks, but I think there's a need for a coordination between ShrtChckTh and LngChckTh. Otherwise two very similar input levels but on each side of the ShrtChckTh level will return very differently. Note how slightly less gets capped to far below at 64


b) My chick is now playable because it doesn't return max, but it's still very high on the scale and I have to close VERY slowly (½ sec) to get velocities below 60. A setting to scale time would be cool. I guess thats indirectly what Curves are doing, but those are for fine tuning and I'm apparently too far off the chart, for them to have an noticeble effect.

Splash
Another interesting topic. A splash is a chick replacement. You can't have both. Splash is not produced by opening the HH. Splashs is a chick, but with the shells reopen so fast after closing that it doesn't produce a dampened chick but a free sizling splash. In that understanding a splash replace a chick completly within a fraction of a second or it doesn't occure at all (chick instead).

I see two problems with the current splash generation.
a) The current splash can be generated upto 1½ sec after the chick it was supposed to replace.
Here one HH step produce a chick (37) AND a splash (38) about a sec later on release.


The current ChckDelay is good for cutting out chicks on splashs. But there's no user control doing the opposite. Preventing Splashs too long after the chicks it should replace. With ChckDelay already there, I think a user controlled setting to decide the latest a splash can replace a chick, would be enough to create a natural splash time slot.

Velocity of the splash should be equal to the chick it replaces, because it's the closing motion that gives the energy not the reopening time. At the moment is looks like the splash velocity is guided by how fast the reopening occurs.

Here's a very soft chick (37) but the splash (38) that should have replace it is loud.


Again thanks a lot for the 20090504 changes, which put a whole new dimension to the MD Hihat control.



As a final remark, I'm adding a how-to setting open and close for those who are interested.

Here a tip from FXpansion (BFD2) on how to identify Closed, 1/4 open and Open:

Heel up, no added pressure, just rest naturally on the palm of the foot = CLOSED


Heel down, no added pressure, just rest naturally on the whole foot = 1/4-OPEN


Heel down, toes up = OPEN


Watch MCT's Midi monitor and adjust OpenLvl and CloseLvl accordingly. Then fine tune to your preferences. Personally I prefer my Open to be a very short distance from the semi-opens.

[gabriel1712]

With ChckDelay already there, I think a user controlled setting to decide the latest a splash can replace a chick, would be enough to create a natural splash time slot.

I gave it another thought and think it could be even simpler. Today, correct me if I wrong, Splash is produced between ChckDelay time and the ~1.250 sec cap after pedal down. Splash velocity is determined by where in the time slot the reopening occurs. Sooner = faster. ChckDelay excludes later chicks.

I think this would work better:
No need for another user control. ChckDelay as today, except chicks can't occure if there's a splash and Splash can't occur after ChckDelay. Splash velocity is determined by the velocity of the chick it replaces.

[dmitri]

So you're saying that if there were a chick a splash must be forbidden to follow right after? IF others agree, I'll make changes in the next firmware.

[gabriel1712]

So you're saying that if there were a chick a splash must be forbidden to follow right after? IF others agree, I'll make changes in the next firmware.

One step on the pedal can only make a chick or a splash, yes.

I have a challenge for you:
Take two trash can lids.
a) Smash them together and keep them firmly together (chick)
b) Smash them together, quickly pull them apart and let them ring out (splash)
c) Smash them together, pull them parrallel apart and try to make them go chick+splash before the cows come in

Elrules wrote the same a while ago.

I think it was talked some time ago. When you step the pedal you only need to send one midi note: a chick OR a splash. You said that when you step the pedal, it should trigger a chick, and then, if you have raised the foot, trigger also a splash. But this is not the real feeling of this. This way, with Superior you get an overlapping chick+splash sound that is not very real. In fact, Superior, and many other samplers, provide a "complete" splash sound, I mean, with the small chick sound and the splashy sound, that an acoustic hat provides. If both chick, and splash, are triggered, the resulting sound is not realistic. Nevertheless, I must ask Gastric for a try on this matter. As he also owns Superior 2, he could check what I am saying (maybe I am wrong). I'll try to monitor Roland td-6 activity with hihat chick and splash playing to clarify which is the true behavior of a commercial pedal.

[dmitri]

I have a challenge for you:
Take two trash can lids.
a) Smash them together and keep them firmly together (chick)
b) Smash them together, quickly pull them apart and let them ring out (splash)
c) Smash them together, pull them parrallel apart and try to make them go chick+splash before the cows come in

I didn't quite get the option c.