|
Fiber in the Whole (House): Cirrus Logic Buys Peak Audio
|
 |
|
by Christian Doering, Marketing Partner, Dynamic Market Systems
In case you didn't know, Cirrus Logic, maker of Crystal Semiconductor A/D and D/A converters (among many other things relating to digital signal processing), has acquired the assets of Peak Audio, the folks who created CobraNet. Peak will operate as a standalone business out of its current premises in Boulder, CO, but "all 18 Peak Audio technology and implementation experts will work for Cirrus Logic, including Charles Anderson, vice president of engineering, and Rich Zwiebel, vice president of marketing." So sayeth the official press release announcing the sale.
First off, why did Peak Audio sell when they were on a roll, signing up licenses by the Mustang-full while Digital Harmony's FireWire initiative languishes? We can assume that the return-to-risk ratio on Peak as business investment is, to coin a phrase, at a local peak - meaning it's been worse, and it will get worse before it gets better, if it gets better. Plus, the business risks ahead of Peak are a lot more manageable for a company with the resources of Cirrus Logic. That's because CobraNet (Peak's invention that makes Ethernet safe for digital audio) is a network technology.
|
A quick lesson in the economics of networks. The classic example is the fax: one's useless, two might be used once a day if their owners had to pass a lot of information back and forth, two billion are an essential part of any business, no matter how many other ways there are to communicate. In math terms, the cost of adding nodes (fax machines, CobraNet gozintaz or gozoutaz, what have you) rises linearly (3 + 3 = 6) while the benefits rise geometrically according the formula for the number of combinations p on a network of n nodes c = n! ÷ p!(n – p)!. That's the idea of "increasing returns to scale," which is the inverse of the familiar "law of diminishing returns" that governed the "old economy." The bigger you are the harder it is to grow - just ask Ford or GM. But the increasing returns to the scale of networks was going to be the engine of the "new economy," letting a thousand Microsofts bloom. Everyone's sales and profits were going to look like this:
|
Peak Audio CM1 CobraNetTM compliant interface
|
OK, we all know how that one played out in the stock market. Not everyone owns a network, not even Cisco. Peak does, however, which means that CobraNet becomes more valuable to Peavey, QSC, and the growing list of CobraNet licensees with every addition to the list. Every new licensee is a potential source of "fax machines" (CobraNet-enabled devices) that can talk to each other and pass signal over Category 5 cable, RF or any other transmission medium that reliably delivers data packets structured according to the Ethernet protocol. CobraNet also piggybacks Ethernet's improvements from 10MB to 100MB to Gigabit Ethernet, which are driven by Ethernet's status as an industry standard. You can't buy a computer today without an Ethernet port, except at a flea market or a users group swap meet. Ethernet enjoys increasing returns to network scale.
With new economy mojo like that, why would Peak accept an undisclosed sum of cash to begin reporting to Jason Carlson, VP/GM of Cirrus Logic’s Consumer Products Division? Two reasons: FireWire, and the digital home network market, which research firm In-Stat predicts will grow from 5 million to over 27 million digitally networked homes worldwide by 2004. FireWire, aka IEEE 1394, is a high-performance serial bus developed by Apple in the 90s. The IEEE 1394 Trade Association plans to increase the transfer rate to 3.2 Gbits/s. The current speed limit of 400 Mbits/s is already enough for full motion digital video. The drawback to FireWire's copper/electrical interface is the short cable run permitted: 4.5m before you need a repeater, for a maximum run of 72 m over the 16-device daisy-chain that the IEEE 1394 spec allows.
Japanese researchers have overcome that problem with high speed optical adapters that send digital video over 100 meters, (end zone to end zone and then some). That project, by the way, was headed by Keio University's Yasuhiro Koike and funded by the Electronic Industries Association of Japan on behalf of the Information Technology Promotion Agency, as part of the Ministry of International Trade and Industry's Project for Common Infrastructure Development of Data Systems in the Home. Companies participating in the trial included Matsushita Electric Industrial (Panasonic), Mitsubishi Electric, Hitachi, Sanyo Electric, Victor of Japan, and Seiko Epson. Once those electronics giants commercialize the technology, video cameras, TVs and PCs (FireWire is handled natively by both Windows and Mac OS) will be hot-pluggable anywhere on that 100 meter network. The applications both inside and outside your digitally networked home of tomorrow today include all the things for which our industry is now turning to CobraNet.
So the Keio University experiment is good news for Digital Harmony, whose DHIVA™-X (Digital Harmony Interface for Video and Audio) is "a complete low-cost reference design based on a Xilinx® SpartanTM-II FPGA." The Xilinx field programmable gate array is a "reference design" that enables OEMs to put FireWire ports on CD players, set-top boxes, audio video receivers, DVD Players, VCRs, surround sound decoders, residential gateways, and home and professional studio equipment. Digital Harmony showed a Mack truck and an SUV equipped with FireWire entertainment systems at CES, but so far their audio networking solutions have involved workarounds like RF hops for the long runs, or the hybrid IR and copper BiDAT system that helps JBL's EVO sound reinforcement systems adjust themselves to speaker placement and microphone position.
FireWire is still an "emerging standard" compared to the ubiquitous Ethernet, but don't forget that linear cost/geometric benefit thing: It means that FireWire could dominate media networking force quickly. Faster growth in market share (linear addition of installed nodes) means faster times faster (geometric) growth in value to the potential user trying to decide whether to stay with the current leader or bet on the challenger's becoming the next dominator. Welcome to the dark side of network economics: the leader in any given application, distributed digital audio for instance, has to keep accelerating just to stay in place. Acceleration requires investment, and that's where Cirrus Logic's resources come in handy. The company's been through some rough times recently, losing $6.83 per share in 1998 on sales of $628.2 million and $.83 per share on sales of $564.4 million in 1999. Cirrus closes its fiscal year in March, but analysts predict that 2000 sales will be up to $781 million, generating profits of $1.79 per share. That's enough good news for Salomon Smith Barney to have acquired 10.6% of the company recently. Citibank went in for 7.1%. Cirrus is 99% equity funded, so it has lots of room to finance a high-speed customer acquisition/market development program by borrowing, it if needs to.
Cirrus brings, uh, serious technology to the table as well as cash. They specialize in manufacturing DSP chips. Matt Czyzewski, Vice President of Technical Operations at Biamp Systems, says "the potential is there for the cost of a CobraNet node to come down by a factor of 5 to 10, if Cirrus Logic puts the whole thing on a chip. That would obviously be good for everybody, because CobraNet would be easier to afford, and real estate would go down, which helps the packaging." Meaning CobraNet products like Biamp's new Audia network, Peavey MediaMatrix or QSC Rave could be a lot cheaper in the future. "I'm sure they haven't started working on an actual chip yet," Czyzewski says, "so it could be six months to a year away."
Aside from the right to mention the United States Senate Chamber, the Sydney Opera House, Stadium Australia (site of the recent Olympics held in Australia), Skywalker Ranch, and the world’s largest theme parks in next year's annual report, what does Cirrus Logic get? This is no Mickey Mouse deal: Peak's technology is in line with the build-out of the Crystal A/D and D/A chipsets into other parts of the signal chain. Cirrus already makes PWM (Class D) amplifiers, and recently acquired some patents from Brit hifi mavens B+W, which "originated from research into digital room optimization" according to Dr. Peter Fryer, B+W’s Research Director. Cirrus just opened eaudio.com, "with the goal of becoming the Internet's most comprehensive source of audio products and information" by "carefully combining the best of traditional brick and mortar business methodologies with Internet efficiencies." And a better example of Wall Street-speak in the post-dot-bomb era you could not ask for.
"Cirrus Logic expects to leverage Peak Audio’s industry-leading skills and complementary technologies to develop a broad range of entertainment-centric audio solutions for networked environments in both the commercial and consumer audio markets," the company press release explains. Sounds like pedal to the metal time in Boulder: Cirrus has notified its shareholders that the Peak acquisition will be "accretive on a cash basis within six months." That's financialese for "profitable," so new hire Jimmy Kawalek has lots of selling to do.
Financial Magic with Network Math
Links
http://www.pcworld.com/news/article.asp?aid=15961
http://www.cirrus.com/press/news/index.cfm
http://www.cirrus.com/press/news/index.cfm?CategoryID=4
http://forum.swarthmore.edu/dr.math/problems/permform.html
http://www.prosoundnews.com/stories/2001/april/0423/0423_1.2.shtml
|