4.3 KiB
Technical Notes
It took me a lot of time digging around the Internet to find some of this stuff. I'm writing down the highlights here, so that when original sources fall off the net, hopefully at least my notes will still be around.
Sample Rate
The number of samples taken since the last HID report is sent as the second Y axis, which doesn't appear to be used by anything else. You can use the included gamepad tester, to see the approximate number of samples as an integer.
This is approximate, because the browser encodes the value as a real number between -1 and 1. We convert it back, but may lose a little precision. It's close enough for me, hopefully it's close enough for you.
This number will not be very useful unless you are polling the wammy bar, since without that input, updates are only sent when a button state changes.
Debouncing
Using millis() time to debounce the switch
roughly halved my sample frequency.
So instead, I do some preprocessor arithmetic
to calculate how many samples to take after an edge,
in order to debounce switches.
The drum controller was a partcular pain: in addition to the switch bouncing, the stick was bouncing on the rubber pad. I settled on a 40ms silence window as feeling pretty good. You can adjust this if you want to.
USB Junk
Gamepad Buttons
The buttons struct member is a bitmap,
each bit mapping to one of the 12 buttons reported through HID.
Here's what each bit means:
- 0o00: Blue
- 0o01: Green
- 0o02: Red
- 0o03: Yellow
- 0o04: Orange
- 0o05: Tilt Switch / 2x kick
- 0o06: Solo modifier (second row of buttons on guitar)
- 0o07: ???
- 0o10: Minus
- 0o11: Plus
- 0o12: Drum pad modifier
- 0o13: Cymbal modifier
- 0o14: Select
hatAndConstant
The HAT switch reports its position like a clock.
7 0 1
6 8 2
5 4 3
0 is up, 2 is right, 4 is down, 6 is left, and 8 is centered.
velocity
I don't know how I can verify that I'm setting this right,
but the rbdrum2midi program looks at these bytes to detect hits.
I set them to 127 when a hit is detected on the digital pin.
Sending these values does not seem to cause problems with my Wii games.
Product ID (PID)
Nicholas, who did the initial work on the guitar, suggested that PID 0x0005 would get the sketch working as drums. And that was correct: this works great on Wii Rock Band 1 and Wii LEGO Rock Band.
But it fails in frustrating ways on Wii Rock Band 3: the yellow and blue pads don't navigate menus, and cymbals aren't detected.
The fix was setting the USB PID to 0x3110.
Drum Velocity
I split the 12 "reserved" bytes from Nicholas's
struct InstrumentButtonState
into 4 bytes of I-Don't-Know, 4 bytes of velocity,
and 4 more bytes of I-Don't-Know.
Whenever a pad is hit,
I send 127 on the corresponding velocity.
I did this because a program called rbdrum2midi
ignores the button press events,
and looks only at the velocity values.
None of the Wii games I have seem to care what these values are set to.
Clone Hero
This controller doesn't have a dpad (currently), so we're using 12 buttons to represent the state of 9 inputs. Sometimes software is just weird!
Clone Hero has two undocumented CymExt1 and CymExt2 mappings. It talks about what they should be set to (on Windows) in the wiki, but doesn't explain what they're used for.
It seems that Clone Hero expects the Yellow and Blue cymbals to send dpad inputs, in addition to the pad color and cymbal modifier. If the controller doesn't send these, then every drum pad hit plays that color's drum and cymbal at the same time, and every cymbal pad hit does the same thing.
I'm not exactly sure why Clone Hero does this, but in any case, Mockband sends the right things now, so the drum controller should work correctly after you map everything.
References
The most valuable sources of information I found were: