Looking backwards . . . and on to a great new year

Two thousand three was the watershed year for the PC and in some respects the multimedia and graphics market as well. It was the year we saw the leader of graphics unseated, and the number two supplier rocket to prominence in the eyes of the users and the stock market. XGI coalesced out of the remnants of SiS’s graphics group ...

Robert Dow


Two thousand three was the watershed year for the PC and in some
respects the multimedia and graphics market as well. It was the
year we saw the leader of graphics unseated, and the number two
supplier rocket to prominence in the eyes of the users and the
stock market. XGI coalesced out of the remnants of SiS’s graphics
group and what was left at Trident and declared it was going to
be the number three supplier. S3, however, actually shipped the
long-awaited and frequently spoken about DeltaChrome and announced
they were going to be the number three supplier—they can’t
both be number three, can they? 3Dlabs said nothing except about
how great OGL was and Matrox said something in Canadian that no
one understood, but it did have flair. Meanwhile in the hills
of eastern England, Imagination Technology readied a new killer
chip just to prove they could, and last, we bid adieu to Bitboys,
who became a captive supplier to NEC for cell phone display controllers.
MediaQ folded their tents and moved in with Nvidia, where jubilant
Jen-Hsun Huang declared it the initiation of Nvidia 2.0. A couple
of new IGCs came out and others were announced from SiS (yes,
they’re still in the graphics biz too and also have a discrete
part—things are complicated in Taiwan), ATI, and VIA, and
there were lots of whispers about Intel’s Grantsdale IGC.

The audio sector embraced 128-bit accuracy (independently of
what any speaker or ear can hear, or any DVD, CD, or MP3 can deliver),
hit 96 KHz sampling and demonstrated 196 kHz—enough samples
to make Nyquist think it was a pure analog signal. DTS and Dolby
Digital codecs appeared in several products, SRS kept making two
speakers sound like five, and Silicon Motion developed audio chips
for the automotive and entertainment PC market while Cirrus Logic
and others entered the theater in a box market with their chips.

IBM emerged as a major CPU player with a clean sweep of all the
next-generation game consoles and a powerful 64-bit engine that
put Apple back in the Spec game. Sixty-four bits was the talk
of the day and AMD rolled out a family of them with an entry for
every segment from servers to workstations and even gaming PCs.
Not to be left out of the game or stopped in its takeover of the
world, Intel sprung forth a 3.2-GHz fat-cached game processor
of its own and let us know, without telling us any details, that
next year there’d two CPUs inside a chip and a year or so later
16. Sun reminded everyone they announced such a multiprocessor
first and would be just as late as Intel delivering it. Transmeta
finally delivered their Astra and sadly renamed it the Efficeon,
or maybe it was a smart marketing move in that we even know its
name. In the meantime VIA jacked up the C3 to 1.25 GHz with more
to come and in classically VIA fashion announced without an announcement
that a C5 would soon be rolling out, on paper at least.

Linux lovers were quick to point out they already had a 64-bit
OS waiting for these brave new CISC/RISC CPUs from AMD, and Intel.
Sun said they had one first, and theirs was so robust it could
run on any RISC or X86 processors. Microsoft, of course, busy
of late with security issues, announced their regular bi-annual
delay of their 64-bit OS, their delay of Longhorn, and their delay
of SPOT. But they did show the next generation of the Media Center,
which looked a lot like the first generation, except this time
Microsoft had a lot of friends who gleefully were offering add-ons
to it. Naturally, all of IBM’s partners quickly announced their
Media Center PCs, hoping to catch some of the luster from mighty
Microsoft’s massive marketing machine.

2003 was the year the LCD TV came to life, much to the relief
of the giant LCD fab builders in Korea, Taiwan, Japan, and China.
With more generation-5 fabs putting out more panels than the fabs
know what to do with, the LCD TV was a very welcome new friend.
Having long ago given up on making any money from the monitor
suppliers, who still keep dropping prices in the hope that one
or two of the almost 100 suppliers will die, the fabs were wondering
if their only future lay in the growth of the more profitable
notebook market. OLEDs were demonstrated with production ready
screens as large as 4 inches, but they still have some problems
and so may be limited to watches, grocery store aisle price indicators,
and entertaining demonstrations at conferences and trade shows.

As the LCD TV emerged so did a handful of SoC TV processors from
ALi, ATI, ESS Genesis, NEC, Philips, and Zoran and its licensees.
Each one claiming to do it all and none of them doing it all at
all, they fought for design wins and declared everyone they got
as the road to riches and higher PE ratios for their share price.

And speaking of SoCs, NeoMagic’s handheld SoC, the M6, finally
showed up along with a couple of design wins for the M3—talk
about long sales cycles. Intel’s StrongArm proved not be as strong
as we were led to believe, and Intel made some major and painful
changes over it. Equator embedded WM9 into its list of codecs
on its SoC, and Sony used its own SoC for its handhelds.

But the handheld world is holding its breath waiting for the
rollout of higher bandwidth networks and the full deployment of
MPEG-4, which looks like it is happening everywhere in the world
except in the country that started the revolution and invented
most of the technology—the U.S where, technology leadership
has been traded off in favor of IP protection.

Meanwhile the HD DVD wars erupted and quickly died down as the
DVD community tried to quash yet another factionist battle over
standards, standards, who’s got the standards that kept the consumers
confused and away from the stores. One camp wanted to carry on
with the ancient MPEG-2 codec and make the disks amazingly dense
and fast enough to deliver a HD stream to the 135 ±2 HDTV sets
in the world, while the other camp wanted to use the more sophisticated
H.264 level of MPEG-4, or WMA, or DiVx—hmmm, can we have
a physical media standard and a zoo of codecs? Others still, like
Pixonics, wanted to stay with the existing DVD optics and motors
and just cram more stuff on the disk while keeping it compatible
with MPEG-2 players.

All of this, or a least a great deal of it, is in the name of
consumer entertainment in the home, and possibly on the road.
But it won’t be very entertaining if all you can watch are your
home movies and unprotected TV shows. CRM, or the lack of it,
is the dark thing that rattles under the beds of the Hollywood
moguls at night, who don’t want to fall like the former music
industry moguls have. The toothpaste was already out of the tube
for the music moguls who had the happy monopoly of selling you
the same song for a lot of money several times; and they still
haven’t realized those days are over. The movie moguls, not wanting
to repeat that process, have simply refused to let their content
out until the walls are unbreachable, impenetrable, and totally
bulletproof, as if any media ever could be.

Content—you want content, then go play a game. And although
this $65 billion market has exceeded the movies industry, and
game controllers have 13 different buttons, its basic use is still
mostly aiming at things and hitting them; it therefore seems undeniable
that video games, compared to other forms of entertainment, are
disproportionately concerned with violence—unlikely we’re
going to change that or global warming either.

How was it as a year? Pretty damn exciting. The industry and
the U.S. seem to have finally pulled out of the recession, and
the rest of the world is trying to do shake off their doldrums.
Technology development didn’t slow a bit, Intel led the world
with its “invest your way out of a recession” philosophy,
and either deliberately or subconsciously the industry seemed
to follow Intel’s guide—how’s that for thought leadership?

Next year? Well that’s the next edition, the first for 2004 when
we’ll give you all the predictions we can get plus our own. Want
a hint? It’s going to be great.