Big Patch Series Prepares the Linux Audio Drivers for MIDI 2.0
phoronix.com
phoronix.com
All in SysEx format, I presume? :p
Sorry, that's some low hanging sound design/kernel programming intersectional humor. This is cool stuff though, I've been a happy camper using Linux for sound design ever since Bitwig picked up native PipeWire support. These patches will be nice to have for MPE setups in the future!
Lots of audio processing is already happening in userspace (think pulseaudio / pipewire), and putting USB drivers in userspace is also something that I think would simplify development. Or is there no ready-made infrastructure for that yet?
We could all be using user land networking stacks too over raw sockets but we don't do that either. Yeah, it too is simpler for the kernel.
One of the key roles of the OS is to abstract access to resources. Turns out that's kind of useful.
More layers, more complexity, let alone USB endpoints and software timers = this is not going to be the fun thing everyone thinks it will. Anyone played with a Waldorf product called the Blofeld? You feel latency that varies while you turn knobs in local-on mode! I’m not saying all implementations of hardware synths will feel bad in the future, but there is a giant ball of denial as to how newer not so embedded systems perform.
MIDI is not high / low but current on / current off.
But it is correct that the current which ends up flowing through the connection is what determines the bits sent (by virtue of the receiver being an optocoupler, not an op-amps).
[1] https://midi.org/specifications/midi1-specifications/midi-1-...
(thx, I didn't knwow about required independence from the ground)
(In this way, two connected MIDI devices can have different ideas of what signal ground is without causing issues.)