For those who are interested in further reading - there was a good
article titled 'A Musician's Guide to USB' by Scott Wilkinson in the
"Desktop Music Production Guide 2000" published by Electronic Musician
about a year ago. If you can find a copy of this article you'll be able
to read about the differing methods USB can use to transmit data and
see why Gary's explanation is accurate - and parallesl so many of our
experiences.
You might also have noticed Win users appear to have much better luck
with USB + SysEx then Mac users - I have no scientific proof this is
only my general impression. If time permits, I create some more SysEx
dumps via USB - I still some old ones from the MIDIMAN 2x2 too.
Jon
article titled 'A Musician's Guide to USB' by Scott Wilkinson in the
"Desktop Music Production Guide 2000" published by Electronic Musician
about a year ago. If you can find a copy of this article you'll be able
to read about the differing methods USB can use to transmit data and
see why Gary's explanation is accurate - and parallesl so many of our
experiences.
You might also have noticed Win users appear to have much better luck
with USB + SysEx then Mac users - I have no scientific proof this is
only my general impression. If time permits, I create some more SysEx
dumps via USB - I still some old ones from the MIDIMAN 2x2 too.
Jon
--- In AN1x-list@y..., "Gary Gregson" <gary@y...> wrote:
> Paul,
>
> My understanding is that its not a case of bandwidth (MIDI only needs
> 31.25KBits/s)...its a case of the underlying protocol used to deliver the
> data. Both Serial and Parallel interfaces have handshake protocols
> (implemented in hardware). This means if you try to overrun the receiving
> device it can hold off further transmission until its ready.
>
> Many USB MIDI devices appear to operate on a 'fire and forget' scheme...they
> expect the receiver always to be in a state where it is ready to accept
> data. However, in the case of USB MIDI there is a step transition of
> bandwidth. ie. at some point you have to go from the 100Mbps of USB, to the
> 31.25KBits/s of MIDI. Unless the interface designers have done some clever
> stuff, this can cause big problems.
>
> Typically I have found most USB devices work fine for simple MIDI
> playback/record. However when you start putting heavy demands on
> them...particularly involving large amounts of sysex....then problems occur.
> The net result is often that data is lost due to receiver overrun (or other
> glitches).
>
> The An1x uses big sysex packets for transferring patch data and is very
> sensitive to the packet timing. Consequently it tends to show up problems in
> USB interfaces:-(
>
> Perhaps if people with working USB interfaces (that can reliably bulk dump
> AN1x data to/form the computer) would care to post a message...then we could
> weed out the good ones from the bad.
>
> Regards
>
> Gary
> Email:
> gary@y...
> http://www.yme.co.uk/yme
>
> -----Original Message-----
> From: dkblade@b... [mailto:dkblade@b...]
> Sent: 15 February 2001 10:18
> To: AN1x-list@y...
> Subject: [AN1x-list] Re: MIDI Interfaces - avoid USB?
>
>
> Hey there,
>
> I really don't see how the interface between the PC and the MIDI port
> could make a difference. USB supports over 100Mbps transfer rate,
> compared to 112Kbps of serial and 5Mbps of parallel. Er...greater
> bandwith = higher timing resolutions possible = more accurate. But
> it's besides the point. The timing is determined by either your
> sequencing software or your master midi clock, not by interface.
>
> The only reason not to get USB midi port is possible
> incompatibilities with software, but mine works great :)
>
> Cheers,
>
> Paul Kabzinski