 |
 |
|
Midi Express XT Parallel Driver |
|
|
|
|
|
|
| MIDI interface progress |
|
|
By Cyan - Posted on November 16, 2007 - 21:25:55 (#22101)
Current version when comment was posted: 0.2 |
|
|
The schematics are now up:
http://littlebluerodent.tripod.com/MIDI/Midi.htm
No idea if it'll work yet though -- probably worth holding on until I've actually assembled something and tested it, not to mention improving the documentation.
Drivers should be coming shortly!
|
|
| I think I could swing that.. |
|
|
By dokV - Posted on October 22, 2007 - 19:57:00 (#22036)
Current version when comment was posted: 0.2 |
|
|
Sure, that sounds relatively easy. I've got plenty of soldering experience and I've done a bit of PC board etching as well if that's an option. It may take be a bit of time to cobble the whole thing together but I'm certainly interested.
|
|
| Custom MIDI interface |
|
|
By Cyan - Posted on October 18, 2007 - 16:14:29 (#22021)
Current version when comment was posted: 0.2 |
|
|
The design I've got at the moment involves hooking some 16C550A UART chips up to the parallel port, with some basic glue logic. It can be built from 1x0 up to 32x32 ports (!) pretty easily.
A stand-alone switching matrix can be implemented using mechanical switches (though it'd only switch 16 channels at a time with no MIDI merge).
This design would require roughly five chips (DIP), plus a further two chips for every MIDI I/O pair. It can be built using point-to-point wiring (no PCB, minimal or no soldering), though it may become a bit of a "rat's nest" if built that way.
Would you be able to construct such a device yourself?
|
|
| I like your idea |
|
|
By dokV - Posted on October 17, 2007 - 01:37:58 (#22020)
Current version when comment was posted: 0.2 |
|
|
I agree that a new ground-up parallel port interface is the way to go (just don't use hyper-proprietary MOTU's non-standard pinout and it should work quite well). I'm aware of the reverse engineering challenges with the commercial interfaces and the Opcode is serial anyway. The only feature I would miss from the commercial units is the ability to work as a network-able stand-alone routing matrix.
I'm interested in your design. Let me know if I can help out.
|
|
| dokV -- |
|
|
By Cyan - Posted on October 16, 2007 - 19:00:40 (#22018)
Current version when comment was posted: 0.2 |
|
|
One of my current projects is developing a new MIDI sequencer for BeOS / Haiku, so I'm very interested in getting a parallel-port-based 8x8 MIDI interface working ASAP (I prefer parallel over serial/USB because the timing is so much better).
So as it stands, I can see two options: 1) Write a driver for a commercial parallel-port MIDI interface, or 2) Develop a completely new MIDI interface specifically for BeOS.
(2) is actually the easiest option -- I've already got the rough design complete, but you'd need to either build the device yourself, or find someone who can build it for you. Also, if you want it supported under Linux/Windows, you'd have to assist with porting the drivers.
The problem with (1) is the lack of specs. If you could find some detailed specs for a relatively common 8x8 parallel-port MIDI interface (e.g., the Opcode interfaces?), I'd definitely consider writing drivers for it -- in fact, it'd go straight to the top of the priority list.
|
|
| Thanks for trying |
|
|
By dokV - Posted on October 15, 2007 - 11:37:13 (#22013)
Current version when comment was posted: 0.2 |
|
|
I have an MTP serial/parallel and have had the best luck using it with BeOS MAX 3 so far (only limited use on other versions but it does seem to make a difference which version you use) and an old Dell 7500.
Sending midi from the computer to the interface seems to work well enough, but midi from the interface too the computer has problems with both discerning which input the midi is coming from (it seems to pick the first midi input it gets info from and then gets "stuck" there) as well as issues of buffer overflow (large amounts of midi info can and will cause a system crash, esp from multiple midi inputs).
Of course this is all a bit redundant since this is without much doubt the worst implemented PC to midi interface ever created for reasons made abundantly clear in the driver files' included documentation. Still I thank the authors for trying as BeOS could really use all the drivers for multi I/O midi interfaces it can get. Now if only there was a driver for my old Opcode Studio 128 (or any Opcode, or any non-motu unit of this type..) That would be a lot of fun to run with Sequitor... Ah well.
|
|
|
|
|
