Axiom 49 MIDI-IN bug

[Pete]

I seem to have discovered a generic problem with
M-Audio's Axiom 49 v2 keyboard controller that you
might want to be aware of. I can't convince M-Audio
that it exists, but it is evident with all my computers
and with both the original keyboard I bought and the
replacement they sent me. So I thought I ought to
bring it to a wider audience [all three of you that
read these groups... (:-)]

I have to note that I don't use either Windows or Mac.
I mostly run the Haiku OS, but I also use Linux, and
the glitch is just as apparent there. In any case,
from the nature of the bug it seems impossible that it
would be affected by the OS. If anybody is using this
unit with either Windows or Mac and does not see the
effect, I'd love to hear.

The problem is with the external MIDI-In connection
on the back of the Axiom. This feeds MIDI data through
the single USB cable to a separate Input node in the
computer. As long as *only* this, or *only* the Axiom's
keys and controls, are sending data, everything is fine.
However, if the Axiom sends data while you are also
feeding the external MIDI-In, the latter can lock up.
The Axiom events continue to arrive as normal, but
the external line is totally frozen. Sometimes the
freeze lasts a second or two; sometimes it requires
a power-off on the Axiom. The usual trigger of the
freeze seems to be the stream of aftertouch events
when a key is pressed hard.

This has to be happening in the Axiom itself because
all MIDI events are sent as packets down a common USB
'wire'. The driver in the computer receives all the
packets (via the OS's "USB stack") into a common
buffer, and decodes them to determine which input
they are destined for. Using a debug version of the
(Haiku) driver, I can see all the packets as they
arrive, and during a freeze, ones from the external
MIDI-In just don't appear in the buffer, while the
Axiom's own continue uninterrupted.

As the external connector is a standard 5-pin MIDI,
it shouldn't be dependent on the actual source of
events, but just in case the piano I was using might
be deficient in some way (though it is absolutely
solid feeding other MIDI interfaces) I tried driving
it from metronome events generated by another computer.
The freezes were the same.

I'm pretty confident by now that this is a problem
in all Axiom 49 v2 units. No idea if it extends to
other similar products. I suspect people just don't
use that MIDI-In port, so it hasn't been noticed.
I was hoping to, but it's not fatal that I can't.
I have lots of available USB ports, so I just have
to run my piano through my MidiSport 2x2.

I should also report that in other respects I'm very
pleased with the Axiom. There are a few other little
things that don't match the description in the manual or
something, but overall it gives me a lot of flexibility
I never had before. If you were thinking of depending
on the MIDI-In feature, though, you may have to think
carefully.

-- Pete --
--
============================================================================
The address in the header is a Spam Bucket -- don't bother replying to it...
(If you do need to email, replace the account name with my true name.)