@nicnut said:
But like I said, i don't have this problem in Max or any other software, it's a unique problem with Pd.
If you have multiple MIDI devices, there are two ways to model the channels.
-
One way is to model separate devices, and within each device, model channels 1-16.
-
Another way is to model the first device as channels 1-16, the second device as channels 17-32, the third device as 33-48 and so on. (A weakness here is: For MIDI messages that don't carry a channel number, you can no longer disambiguate devices.)
What you cannot do is to model a single "channel number" space and have all devices occupy 1-16. Then you have no idea which device sent a given message.
It seems like your point is that the Pd way is flatly incorrect. It isn't incorrect exactly: It provides enough information to tell devices apart (except e.g. sysex). It doesn't do it in the way that you're accustomed to, but that doesn't make it wrong.
If Pd assigns channel ranges based on the order of devices reported by the OS, and the OS lists the devices in an inconsistent order, then it is as much the fault of the OS as of Pd. The "save MIDI configuration" approach might be a workaround for confusing behavior when (I'm guessing here) the OS is confronted by multiple devices with identical ID information.
hjh
