Midi Line Code Question (Low Level)

[Jeremy Brookes]

>>>I'm happily integrating my 256 pot midi controller UP however the midi
>>>output section is causing much grief. I have a series of parallel to
>>>seriel shift registers that I load then shift out at the proper rate
>>>howver the message is some how corrupted. I verified that the bits in
>>>the massage are correct:
>>>
>>>F0 7F AA DD F7
>>>AA is pot address
>>>DD is pot data
>>>|=start and stop bytes = binary one
>
>Jeremy,

Actually it wasn't me but someone else on the list (Philip Pilgrim) who 
posed the original query. But not to worry about that...

>See <http://www.hinton.demon.co.uk/midicode.html> for details on the MIDI
>protocol.
>
>The FO 7F sequence is for a special "universal real-time" Sysex message
>format. (You shouldn't use this ID unless you are intending to use MTC,
>MMC, or MIDI Show Control)
>
>I think you want to use the "non-commercial" Manufacturers' ID, which is 
7D.

Thanks for this info. I tried searching some of the literature I had but 
couldn't find it - it was quite old and included a whole host of now sadly 
defunct manufacturers.

>However, you probably want to put in some other format identifiers between
>the FO 7D and the actual data so you will be able to prevent conflicts on
>the 7D Mfrs ID, and to allow you to expand your protocol in the future.

This is probably the most MIDI-compliant approach to use but I fear it 
introduces more overhead in terms of extra bytes. I wonder why the guy who 
proposed this (Philip) doesn't send Controller Messages instead? (Just 
three bytes). Or even use NRPN messages (a few more bytes).

I've come across some PC based sequencers (Cakewalk) that don't allow 
real-time System Exclusive. So if you wanted to record these knob twiddles 
and play them back it might be necessary find the right sequencer or use 
Controllers or NRPNs.

Just my tuppence-worth...

mailto:jezz at enterprise.net
http://homepages.enterprise.net/jezz/