On Wed, Aug 21, 2013 at 03:09:10AM -0000, steve_the_composer wrote: > Yeah--just a couple of footnotes: > > (2) I agree there may indeed be something going on with rate/speed. The only difference I see is that I don't think the problem is faster CPUs in as much as it is a failure of the OS or usb communication or interface drivers or something else that prevents steady and reliable 8-N-1 serial communication at 31250 bps. I don't disagree that adding in some extra time between either bits or bytes might solve the issue at hand. If it does, my gut says its more a matter of the nature of the serial communication process than the CPU speed. (Of course with GHz speeds, software and firmware designers may not give a hoot about serial data maintaining a reasonably reliable 31250 bps rate.) The 31250bps MIDI rate is locked down by the interface, and can't go anywhere. It is possible that faster processors are sending so much data that the interface is being overwhelmed, but a correctly-written USB driver *should* buffer the data in a FIFO and send as fast as the UART will take it. The CZs need the handshaking messages and it's really only a happy accident that slower computers give it enough time to get its daft wee silicon head together between messages without bothering with the handshaking. I don't think there are currently any applications out there that support CZ patch dumps correctly. It's probably time to break out the CZ programming manual... -- Gordonjcp MM0YEQ
Message
Re: [CZsynth] Re: Can't get sysex to work with a CZ-5000
2013-08-21 by Gordon JC Pearce