On Mar 31, 2007, at 5:26 AM,  Jens M Andreasen wrote:

What I wonder is; wouldn't it be possible to bypass this USB radio
transmitter/recaiver dongle-thingie and fast forward to the wireless
capabilities already built in to most of modern portable devices sold
today? Hey, even the OLPC aimed at third world children have this
capability ...


If you access to two OS X machines for a few minutes, run the
Network MIDI Driver over Airport (Network MIDI Driver described here:

http://www.soundonsound.com/sos/jul05/articles/tiger.htm#3

You'll see it works OK, even with the 1-2% packet loss that the IP
layer sees on a typical WiFi network.  Apple's implementation uses
IETF RFC 4695 (a.k.a. RTP MIDI) as the transport layer, which
includes a resiliency system (the recovery journal) for handling
an arbitrary number of lost packets with only transient artifacts.

More on RTP MIDI here:

http://www.cs.berkeley.edu/~lazzaro/rtpmidi/index.html

Note that there are actually third-party products now compatible
with Apple's protocol, although these products are for wired networks.
And, last time I checked didn't implement the recovery journal
part of the RTP MIDI (and so can't be trusted over switches that
lose packets -- although my info on the product might be out of date):

http://www.kiss-box.com/products.pdf


It kind of puzzles me that this is not happening already ...


I haven't looked into the issue in detail, but for my own use wireless
only makes sense if I don't have to be constantly worrying about
the battery life of the device.  At the moment, what I see in
the state of the art in portable WiFi chips -- the chip in the Sony
PSP and the chip in the Microsoft Zune -- is not too encouraging
in this regard.  But hopefully in a few generations of chip designs
this will get better.  This review:

http://emusician.com/midi/emusic_maudio_midair/

make the point well of how limiting the battery life can be for
a MIDI wireless controller right now.


---
John Lazzaro
http://www.cs.berkeley.edu/~lazzaro
lazzaro [at] cs [dot] berkeley [dot] edu
---


Reply via email to