One problem with using the Kaossilator Pro as a MIDI controller is that it only has very few controls: the X/Y pad, one slider, one potentiometer, and a few buttons - that's all. Of course, any self-respecting control freak would have something more like this in mind:
http://www.flickr.com/photos/apenny/2963893735/ Adding a USB-MIDI controller ---------------------------- One way to get more control(s) is by adding one of those fairly cheap USB MIDI controllers "for laptop DJs". Here's my current setup: http://downloads.qi-hardware.com/people/werner/m1/midi/vjlab.jpg Audio is played by the PC on its speakers and picked up by the M1 with the internal microphone. Mouse and keyboard connect to the M1's USB ports. (I need a separate mouse because the M1 firmware currently doesn't handle the pad built into my RF keyboard.) Labsw acts as a fancy power switch for the M1. Neither line in/out nor camera in are used. The Kaossilator Pro (KP) and the nanoKONTROL2 nK2) both connect to the PC with USB. The MIDI messages from both are mixed on the PC with Jack, and then sent with an USB-to-MIDI adapter to the M1. The connection routing in Jack's qjackctl ("Connect" panel) looks like this: http://downloads.qi-hardware.com/people/werner/m1/midi/qjackctl.png I first tried to keep things a little simpler and use the KP as USB-to-MIDI bridge, but found that the KP was very reluctant to have anything to do with MIDI input. I can get it to accept control input, but it refuses to play notes, let alone pass on MIDI messages from input to output. This needs more investigation. MIDI addressing --------------- Since M1 presently only listens on one MIDI channel, both KP and nK2 have to send on that channel as well. Note that MIDI channels are usually counted starting with 0 (e.g., M1, vkeyb), which is also what the MIDI specification says, but some count starting with 1 (e.g., Korg (KP, nK2), vmpk). With everyone sending on the same channel, controller numbers can clash. KP and nK2 indeed have some overlapping numbers (on KP, the "program memory" buttons with codes 49-56 vs. on nK2, the "M" buttons of groups 2 through 8). KP and nK2 both support reassigning of control numbers, but nK2 uses some proprietary Windows/Mac program and the KP2 has a rather awkward procedure and just happens to exclude these eight buttons from reassignment. I decided to save this battle for another day. Video ----- I made a little demo video: http://downloads.qi-hardware.com/people/werner/m1/demo/MVI_1747.MOV It's rotated again. To play un-rotated, you can use mplayer -vf rotate=2 The audio track is "Art Now" by Alex Beroza, featuring Snowflake, and licensed under CC-BY: http://ccmixter.org/files/AlexBeroza/30344 http://ccmixter.org/content/AlexBeroza/AlexBeroza_-_Art_Now.mp3 The M1 patch is still a hacked "Tornado", this time with the following controls: Variable Control midi1 12 KP Pad X rot = midi1*2 midi2 13 KP Pad Y zoom = 0.9+midi2/5 midi3 91 KP Slider wave_x = 0.5+midi3/5 midi4 94 KP Potentiometer wave_scale = midi4*20 midi5 4 nK2 Slider 5 wave_r = midi5* ... midi6 5 nK2 Slider 6 wave_g = midi6* ... midi7 6 nK2 Slider 7 wave_b = midi7* ... midi8 7 nK2 Slider 8 t = time*midi8*5 Color controls -------------- At the beginning of the video, I went through the new color controls, with roughly this timeline: 00:00 green = max, red and blue = 0 00:08 added blue 00:13 removed green, added red 00:23 replaced green with blue 00:31 brought up all channels and increased color change rate My color controls so far are more a proof of concept than anything useful. The very simple changes I made to the color algorithm in the patch don't provide good control over the color composition. Furthermore, changing the color change rate also changes the phase, i.e., the currently active color mix. I also don't know why green still keeps on dominating. - Werner _______________________________________________ http://lists.milkymist.org/listinfo.cgi/devel-milkymist.org IRC: #milkymist@Freenode
