[linux-audio-dev] Fwd: OT -- USB History

From: Christoph Eckert <ce@email-addr-hidden>
Date: Thu Feb 16 2006 - 21:06:42 EET

Hi all,

sorry for cross posting, but I found the following info very interesting
when discussing about USB 1.1 vs. USB 2.0 stuff.

Some of us (including me) would like to see a USB 2.0 breakout box which
grants more than 2x2 channels while using a standard USB 2.0 protocol.

Bruce Wahler politely agreed to spread his info to our lists, so all
credits go to Bruce.


---------- ----------

Subject: OT -- USB History
Date: Donnerstag, 16. Februar 2006 18:12
From: Bruce Wahler <desp@email-addr-hidden>
To: access-list@email-addr-hidden

Hi All,

[Warning: This post has little to do with the Virus TI per se. It
 might be of interest to some of you, though.]

I was involved in the early USB efforts, working for a major PC
 manufacturer. The 3-tier speed approach of USB is a confusing -- and
 necessary -- part of the design. Early USB appealed to two groups:
 1) manufacturers who wanted to simple, cheap way to untangle the rat's
 nest of wires that were growing behind computers; and 2) developers
 who wanted a better, more flexible connection than serial and parallel
 ports provided. In addition, the creators of USB had this grand
 vision of "USB everything": kitchen appliances, phones, televisions,
 you name it.

USB attempts to satisfy all of these needs, but the goals of different
 markets are sometimes at odds with each other. Devices like mice
 can't afford to add even $1.00-2.00USD of product cost, because their
 customer base won't accept the price increase. On the other end,
 there's no such a thing as "too fast" for disk drives and networks.
 USB 1.0 (and 1.1) came out with Low- and Full-Speed specifications to
 try to bridge the needs of these two camps. Same connectors, same (or
 similar) cables, same hardware at the host (computer) end; all of the
 higher-speed functions had to be a layer on top of the basic ones.

At the time of USB 1.0 (1995), the practical limit for cables and such
 was considered to be somewhere in the range of 10-15Mbit/sec. This
 wasn't a PHYSICAL limitation; it was governed by the cost of hardware
 (cables, connectors, ICs, etc.) compared to the amount of data being
 sent (<1GB). Unfortunately, USB 1.x took several years to gain
 acceptance. (PCs had the USB ports back in 1995, but there were no
 real peripherals nor OS support for 3-4 more years. One of my bosses
 used to refer to it as the "Useless Serial Bus.") By the USB really
 took hold (2000? 2004?), there were enough advances in technology and
 manufacturing to up the speed a great deal. Add to that the need to
 transfer more data, and the fear that FireWire would eclipse USB, and
 "Hi-Speed USB" was born. Hi-Speed USB follows the same rules as USB
 1.0: faster protocols must work around the limits of slower ones, so
 nothing becomes truly obsolete. This is why a 12Mbit/sec Virus TI is
 still "USB 2 compliant."

Some important things to know about Hi-Speed USB:

1. The GUARANTEED cable length is shorter (5m vs. 2m). With a quality
 cable, you might run further, but there's no whining if it doesn't
 work. This certainly limits the ability to use the Virus TI as both a
 recording platform and a performance synth at the same time.

2. Raw bandwidth numbers of Hi-Speed USB are deceptive. (This is also
 true of FireWire.) While the cable and ICs can support 480Mbit/sec.,
 it takes great drivers, proper interrupt selection, and a relatively
 unused computer to use that bandwidth. Otherwise, it's a game of
 "hurry-up-and-wait." Sharing USB with slower devices also clouds the

3. USB 2.0 enhancements focused on data storage. There weren't any
 high-speed audio extensions added. If Access had wanted to use
 480Mbit USB audio, they would have had to develop and support it from
 scratch -- on both the Mac and PC. So, it's not just a case of adding
 a little product cost; it's a large development and testing challenge,
 too. Why weren't there audio extensions? Probably because the two
 "official" audio uses for USB -- Internet phones and digital USB audio
 -- didn't need them. The first one is fine with 12Mbit/sec, and the
 second one never really caught on.

4. The USB specifications were mostly written by big companies like
 Microsoft, IBM, Intel, and Compaq. They sunk a lot of resources into
 USB, and so their needs took top priority. None of those companies is
 known for professional audio gear -- they're computer companies, and
 USB audio was and still is a bit of an afterthought. (Quick: Name me
 one 'major' US PC manufacturer who sells a true MPC in their standard
 line? Anyone?)

So, why not add the hardware (ICs) now, and write the OS support later?
  The approach rarely works, IMHO. Even in the computer industry,
 known for technology advances, hardware that is unused at product
 launch often remains forever unused. Why? Remember the old adage,
 "If it ain't broke, don't fix it" ? Well, updating software or
 firmware requires breaking that rule. And anyone who's written
 software will tell you that bugs crop up in the strangest places.
 While the updates are cool, there's often very little evidence that
 the efforts resulted in big sales increases. Thus, a small company
 like Access must be choosy when planning product updates.



Bruce Wahler
Design Consultant
Ashby Solutions™   http://consult.ashbysolutions.com
978.386.7389  voice/fax
access-list mailing list
http://www.ampfea.org/mailman/listinfo/access-list   <---
Received on Fri Feb 17 00:15:06 2006

This archive was generated by hypermail 2.1.8 : Fri Feb 17 2006 - 00:15:07 EET