Personally, I find that all of the current MIDI controllers are either a large compromise on functionality or prohibitively expense, often both.
With products like the FCB1010, they're designed so that you have 100-200 virtual buttons, with 10 or so available at a time through banks. The main problem with the FCB1010 design is that it's intended to show an LED for only the last button pressed, so it becomes less than useful for ON/OFF functions. The UNO chip deals with much of this with the "Stompbox" mode, but now you have 5 buttons permanently dedicated to the same ON/OFF functions across all banks - which can be an issue.
Ideally, each button should have a screen display associated with it that tells you what it does now, and operating firmware should allow buttons to be freely associated with any function dependent on the context of what you're doing with it.
Which is why I have a bag of 16x2 LCD screens, I2C adaptors for them, I2C multiplexer chips, A/D converters and RGB LED's, an Arduino board, and a Raspberry Pi sitting on a workbench waiting for me to build something with them. My design idea was to have 12 buttons and two rotay encoders (knobs), each with an LCD screen and an RGB LED associated with them.
The Arduino board is just to run the RGB LED's, as signal timing on them requires precision that the Pi cannot provide while multi-tasking. The Pi would run it all, with button presses doing one or more of the following:
- Send a MIDI command to the AxeFX
- Change 1 or more LCD displays
- Change 1 or more LED lights
- Redefine the actions associated with 1 or more buttons
So, as an example: Assume that when you boot the unit the first thing to do is to select a preset. So you might have 10 presets set up on ten buttons (all indicated with the associated display), and the last two buttons for bank up and down. Then when you press on one of the buttons, it sends the MIDI command to load that preset on the AxeFX and then 10 of the buttons switch over to functions directly related to that patch. Maybe you have two drives, each with X/Y defined on them, so that could be 4 buttons, perhaps multiple delays or reverbs, touch tempo, maybe a button that engages a delay and then lowers the mix of the reverb a little. Perhaps 4 buttons for scene selection. Everything with a relevant LCD display. Also, the rotary encoders would change functions accordingly. The last two buttons might be for "Utility" and "Preset Selection".
Also, the buttons are just momentary switches, so the software could easily detect double clicks and long clicks. So each button could easily house three functions at any one time. Maybe long click on all of the buttons could be back to the boot state, "Preset Selection" (or maybe just the button on the bottom right).
I got stuck when looking for a housing. Originally, I wanted a sloped cabinet with a flat section at the top. The rotary encoders, and main LCD panel would go on the top, and the footswitches and associated LCD displays and LED's would go on the sloped section. It was turning out to be very difficult to fine an off-the-shelf housing big enough in that layout. In the end, I decided a week or so ago that I'd go with simple sloped console design, with the top being 11" x 17". That will hold 3 rows of 4 switches/displays, and have room at the top for a row of LCD displays and rotary encoders. Hammond makes them for about $50, which is not to bad.
In terms of cost, a single set of footswitch/rotoray encoder + LED + display should cost about $6. The Pi and the Arduino are each about $30, and the housing is $50. There are some other incidentals, like multiplexer and Analog/Digital converter chips, but those are pretty cheap. So all told $200 worth of hardware, maybe a touch more.
Far more difficult, I need to find the time and energy to get out the soldering iron and build a prototype.
Some time ago I saw the Fxunits RAC12 MKII which looked like a really useful controller.
