3D --- Media Creation --- Shared Space
Published by 4th WAVE, Inc.
Issue #707 4/14/97
CONTENTS
After years of development and anticipation, the Mpact media processor has finally arrived. Now available, the Mpact|M3000 is based on parallel processing and achieves over 3 billion operations per second dedicated to the acceleration of PC multimedia and communications functions. The processor is software-upgradable and offers acceleration and support for the concurrent operation of DVD, 2D & 3D video, MPEG-1 playback, SRS TruSurround and Dolby Digital Sound, videophone, fax/modem and full telephony features. Mpact chips are manufactured by Toshbia, LG Semicon and SGS-Thomson, while Chromatic Research develops the software.
Two Mpact boards are currently available:
$399 Mpact Blazer Pro: 4MB RDRAM, 33.6 modem
$449 Mpact Blazer Max: 4MB RDRAM, 33.6 modem, TV display, video capture, digital audio play
Datapath Port To Real3D Platform [April 1]
Datapath (UK-based) and Real3D (a Lockheed-Martin Company) have announced that they have ported Datapath's RealiMation to Real3D's hardware platforms. RealiMation, a real-time 3D development tool, combines powerful features with an easy-to-use interface, and provides menu-selectable renderers including OpenGL, Direct3D, Glide, RenderWare, and the Real3D Pro-1000 Series API.
3NAME3D has made available several new collections of 3D models in their Cyberprops library. The "High-Res Collections" feature models suitable for broadcast and film and include collections of cars, ships, buildings, electronics, sports and animals. The hi-res models are individually priced from $45 for animals etc. up to $175 for cars & coins and are also available as collections priced from $399 to $799. They have also announced "The Living Earth," high-resolution satellite composite images and texture maps, available as states, countries, continents and regions and with resolutions up to 1 km per pixel.
Online Gaming Report From Forrester [April 4]
A new market report from Forrester claims that the number of consumer playing games online will triple by the year 2001 - up from the current estimate of 6 million to 18 million. They also predict that online games will be supported by advertising and sponsorship rather than subscription-based, as well as the rise of online gaming centers. The report can be read for free on Forrester's Web site:
MetaTools Acquires Specular International [April 7]
MetaTools has announced a definitive non-dilutive agreement to acquire Specular International. MetaTools plans to issue the number of shares of MetaTools common stock equal to $5,451,000 divided by the average closing price of the MetaTools stock 20 days prior to the close of the transaction and $1,000,000 in exchange for all the outstanding shares of Specular. MetaTools will also issue 450,000 non-qualified stock options to purchase MetaTools common stock to Specular employees. MetaTools will continue to market Specular's professional 3D modeling software targeted at video professionals while expanding distribution of certain Specular consumer products. They also plan to close Specular's Amherst, Massachusetts headquarters.
3Dlabs Boosts GLINT OpenGL Performance [April 8]
3Dlabs have announced that they have again increased the OpenGL performance of their GLINT processors through software enhancements to their OpenGL Installable Client Driver (ICD). According to 3Dlabs, the improvements increase GLINT's Viewperf benchmarks scores by up to 30%, with a CDRS score 20.48 at true color depths (which they compare to an Intergraph Intense 3D-T Pro 1000 with a score of 19.59 and a Dynamic Pictures Oxygen 202 with a score of 14.94).
3Dlabs and Intervista Partner for VRML [April 8]
3Dlabs and Intervista Software have announced a partnership to "accelerate and promote" Intervista's WorldView VRML 2.0 browser on 3Dlabs hardware. In addition, Intervista is offering bundling deals to 3Dlabs OEM customers.
Caligari trueSpace3 Ships [April 7]
Caligari has announced that trueSpace3 will ship on April 15. New features include accurate modeling detection, live skin with metaballs modeling, the PlastiForm surface particle generator, improved forward dynamic and inverse kinematics and Direct3D and Internet support. They have also announced that a free improved key frame editor will be available within 60 days of the shipment of trueSpace3 to registered users.
Superscape and Black Sun Deliver Web Community [March 12]
Superscape and Black Sun Interactive have partnered to integrate Superscape's Viscape browser with Black Sun's online community technology. Via a new Java interface, the Viscape browser links with Black Sun's Passport client software to enable chat and interaction on the Web. The Viscape browser with integrated Black Sun Passport features is available for download from Superscape's Web site:
New ModelShop From Electric Café
Earlier this year, Electic Café released version 2.5 of ModelShop, a 3D modeling program for Macintosh and targetted at architects, engineers, and designers. New or improved features found in version 2.5 include Boolean modeling, floating-point accuracy, a hand tool for interactive viewing, QuickDraw 3D rendering, texture mapping, antialiasing, transparency, smooth shading, spot lights, animation timeline windows, keyframe-based camera controls and drag & drop support. ModelShop 2.5 retails for $349.
In addition, Electic Café has recently announced that ModelShop 3.0 (formerly 2.6) is currently in production and will be released this summer. Some new features scheduled for version 3.0 include photorealistic rendering, QuickTime VR export, the ability to access any QuickDraw 3D plug-in renderer, QuickDraw 3D 1.5.1 support, VRML import and export, and 3D TrueType and Postscript text. ModelShop 3.0 will retail for $399 with upgrades from previous versions for $99.
Intergraph Announces New 3D PCs [March 31]
Intergraph Computer Systems has announced new 3D PCs that feature Intel's MMX technology. The new models round out the TD PC product line targeted at technical and business professionals. The systems come standard with Microsoft Office 97, Intergraph's Intense3D 100 graphics processors, 16 MB RAM and 1.7 GB disk space. The new models include:
| TD-22 | 133-200 MHz Pentium | 16-64 MB RAM |
| TD-25 | 166 or 200 MHz Pentium w/ MMX | 16-64 MB RAM |
| TD-220 | 180 or 200 MHz Pentium Pro | 16-256 MB RAM |
For a limited time, promotion users can get a TD PC with a free copy of Macromedia's Exteme3D. Prices start at $1,185 (TD-22), $1,485 (TD-25) and $$1,680 (TD-220).
Diamond Ships 56K Modems [March 24]
Diamond Multimedia announced the SupraExpress 56 line of modems, with the initial models, the SupraExpress 56i and SupraExpress 56e, shipping now for $169.95 and $189.95, respectively. The modems incorporate the new Rockwell/Lucent K56flex technology, which has so far garnered the support of over 600 ISPs nation-wide.
1997 International Television Symposium
Where: Montreux, Switzerland When: June 12 - 17, 1997 Sponsor: Montreux Symposia Management (MSM)
The 20th anniversary of the ITVS (which is held every two years) is expected to be an important show, with growing support from major players including Sony, Panasonic, Philips BTS, Thomson, Avid, and Quantel, as well as many smaller companies. The Symposium will feature sessions on satellite broadcasting, cable, terrestrial broadcasting, future technologies and a large number of exhibitors from all aspects of broadcasting (more than 300 exhibitors were present at the last Symposium).
http://www.montreux.ch/symposia/
At WinHEC Microsoft announced that Meltdown 97 will run from July 9 - 11. This promises to be an important event for all developers using DirectX including Direct3D. DirectX 5.0 is to have shipped by then and the details of DirectX 6.0 should be surfaced.
http://www.microsoft.com/hwdev/meltdown.htm
707.3 Opinion: 3D Graphics and Its Mythical Killer Application by Brian Hook
Since the advent of the personal computer industry, analysts and historians have pointed to the numerous killer applications that have transformed platforms such as the Apple II, IBM PC, and Apple Macintosh from novelties of questionable value into massive successes, often in a very short time span. Because of this analysts have a natural tendency now to look at emerging technologies and attempt to identify their killer apps, the fundamental assumption being that new technologies MUST have a killer application to champion their merits and thus irresistibly force their acceptance in the market place.
So what exactly is a killer application? Since the term is not technical the answer largely depends on who you ask, but in general most people agree that a killer app is a single easily identifiable technology (not necessarily just a program) that makes another technology (CPU, operating system, computer architecture) so useful that it cannot be ignored, or at the very least makes a platform useful enough that it survives when the market makes a platform compelling enough to some target demographic that it manages to survive far longer than the market would otherwise allow without that killer app.
A famous and oft-cited example from history is Visicalc for the original Apple. We've all heard the legendary stories today of businessmen walking into local computer stores "to buy that VisiCalc thing" and finding out they needed an Apple to run it, so they bought that too. The appeal of Visicalc was so great that it was responsible for the widespread success of the platform that happened to run it.
Similar, albeit less dramatic, examples exist -- Aldus PageMaker for the Macintosh (some may say the Macintosh GUI was the killer app, but without applications a GUI is pretty useless), Lotus 1-2-3 or WordPerfect (take your pick) for the IBM PC, the World Wide Web and high-speed modems, and WordStar for the CP/M OS. From a slightly different perspective it is highly likely that Microsoft Windows made the mouse a requirement for all new computers.
It's easy to believe that to qualify as a killer application that a program or technology must transform a particular platform into a household appliance overnight. However, this isn't necessarily true -- many times a killer application is what allows a platform to survive far longer than what the market would have normally allowed. The development tools for NextStep allowed it to attain immense popularity in various vertical market industries such as the financial sector. The built-in MIDI capability made the Atari ST stay afloat in the face of an arguably vastly superior competitor, the Commodore Amiga.
We've established a mostly single-ended dependency for killer applications -- without Microsoft Windows, the mouse would not have likely achieved acceptance so rapidly. The same can be said for Visicalc and the Apple. But the relationship between a burgeoning technology and its killer application is often, but not always, symbiotic. To be viable, Microsoft Windows needed the mouse almost as much as the mouse needed Microsoft Windows.
In some cases the relationship can be entirely symbiotic and no clear distinction between killer app and emerging technology can be made. Which was more important to the other -- Microsoft Windows or the Windows graphics accelerator? Without Windows it's obvious that the 2D graphics accelerator would have questionable value (although 2D accelerator chips from Texas Instrument did, in fact, exist before Windows 3.x), but would Windows be as popular today if it did not have the dramatic, some would say ABSOLUTELY REQUIRED, performance boost that a 2D accelerator provides?
Finally, we have situations where an emerging technology succeeds without a single killer application. There were no specific killer applications for the 3.5" floppy disk, hard drive, CD-ROM drive, Creative SoundBlaster, or the LAN. These were all evolutionary technologies -- the 3.5" floppy held more data and more robustly than the 5.25" floppy; the hard drive was fast, big, and convenient; the CD- ROM was a better (and cheaper!) way of distributing large amounts of data; the SoundBlaster sounded better than a PC speaker; and networking just makes sense -- it's just much more convenient than throwing floppy discs around an office.
The point of all this is that many new technologies, while new and interesting, are NOT necessarily so radical or costly that they must have a champion in order to become successful. Or, as in the case of the CD-ROM, they are so obviously a part of the future that they are simply absorbed into the mainstream of computing quietly and without fanfare, just like new CPUs and larger displays. This seems to be common sense, yet we have many industry analysts lose sight of this and see each new technology in light of its as-yet-unknown killer application.
And this is the case with 3D computer graphics. It's obvious that 3D graphics, and by association 3D graphics acceleration, has its place in computing. It's useful for many different types of applications, including CAD/CAM, modeling and animation, home design, education, and, most importantly, games. No *single* application is going to make 3D graphics acceleration a part of everyday life -- THIS IS GOING TO HAPPEN ANYWAY.
The fact of the matter is that 3D graphics is an evolutionary step in computing, not a revolutionary new technology. The world of computing has simply been waiting for the cost to come down so that everyone can have fast and pretty 3D graphics on their desktop.
In our eagerness to identify new killer applications for each new technology that emerges we lose track of the fact that new technologies are often just necessary or desirable for computing. To truly qualify as "killer" an application or technology must save AN OTHERWISE DOOMED piece of technology from drifting into obscurity or obsolescence. If you can look at a technological advance and point out half a dozen or more "killer apps" for it, in all likelihood that technology doesn't have a single killer app -- because it never NEEDED a killer app.
So while we sit around waiting for a killer application for 3D graphics, fast 3D graphics is just going to become a part of day-to-day computing. Then one day we'll realize that 3D graphics is just a part of computing, and historians will try and figure out what its killer app was. And no matter what they say, they'll be wrong, because 3D graphics will never have a killer app.
Sometimes a new technology just makes sense.
But as Dennis Miller says -- that's just my opinion and I could be wrong.
----
Brian Hook is a 3D software and hardware consulting engineer specializing in real-time 3D graphics programming, training, and documentation. He is also a contributing editor for Game Developer magazine, the author of a book on programming 3D graphics for the PC, and writes the occasional article for The Wave Report in his copious spare time. He can be reached at http://www.wksoftware.com or bwh@wksoftware.com.
707.4 nvidia - RIVA 128 - Major Performance Improvements in 3D, 2D and Video by John Latta
nvidia has announced its 3rd generation product, the RIVA 128, for Real Time, Interactive and Video & Animation. WAVE has seen initial results from this chip and it raises the bar again in 3D and 2D performance. nvidia has done 180 degree turn away from a special implementation of the 3D pipeline with its nv1 and gone for the maximum performance it can get in a mass market part - $30 in quantity. At the time this report will be distributed the part will be only 2 weeks out of fab and early test results are significant - the fill rate is 100 Mpixels/sec with all features turned on.
nvidia took a risk in the design process and it appears to have paid off. That is, it was a flagship customer of the IKOS VirtuaLogic and Avatar hardware emulation. Using this technology in only 2 months nvidia was able to implement the complete silicon from the ASIC specification, do complete functional tests and begin the software driver tests before fab - what is called Virtual First Silicon. They claimed that Avatar allowed them to find every functional error on the chip.
The chip has a 128 bit pipeline throughout. nvidia claims that what they call fast and wide is critical to support not only 3D but simultaneous 2D and video. The chip contains 3.5m transistors, and executes at 20B operations/sec. There are 50 floating point processors on chip and all the set-up calculations are done in FP at a rate claimed to be 5 GFLOPS. The chip runs at 100 MHz, however, the design is scaled such that faster clock speeds are expected to be supported. The process technology is .35 micron with 5 layers. The RIVA 128 includes a texture cache and vertex cache. The expected performance is 5 M triangles/sec (peak with an average of 1.5M triangles - all features turned on) and fill rates at 100 Mpixels/sec for 25-pixel triangles. A 207 MHz RAMDAC is integrated into the chip. The chip also supports planar YUV formats. Capabilities include: perspective correction, bilinear filtering, lighting, 16 bit z-buffer and alpha. The chip will support AGP and PCI
Memory interface is to SGRAM up to 4MB with 128 bits data path and up to 1.5GB/sec bandwidth. What is important about this design is the handling of texture memory and caching. nvidia claims an intelligent caching algorithm which allows the chip to draw texture information from three sources - potentially all simultaneously. In order to achieve the fastest rate the first access is to the onboard texture cache. If accesses are required beyond the cache the chip can go to the SGRAM or system memory using either PCI or AGP. The SGRAM is used for z-buffer, frame memory and texture storage.
nvidia has potentially shortened the time-to-market for a critical component - the software drivers. Using the Virtual First Silicon nvidia was able to move very quickly from silicon to credible demonstrations of performance. On the show floor nvidia was showing WinMark 97 scores of 99.6.
At WinHEC it was announced that STB will use the chip in its Velocity 128 3D product.
Both Thompson and nvidia will market the chip. Thompson will focus on the overseas markets and nvidia will develop OEMs in the US.
707.5 Facility Review - Sega Gameworks , Seattle - Can it make money? by John Latta
With considerable fanfare Sega and its partners Dreamworks SKG and Universal Studios opened its first Gameworks installation in Seattle on March 12th. With the coin-op industry in difficult times, fresh approaches to out-of-home game play are examined with great expectation. Certainly Sega has led in its use of 3D and one expects that venues such as Gameworks would be a house full of 3D fun. In this respect one was certainly not disappointed. We counted 148 games and they ranged from the latest including Super GT based on the Lockheed Martin Model 3 platform to ancient titles such as Donkey Kong.
Located at the corner of 7th and Pine next to a Cineplex Oden multiplex theater and Niketown, Gameworks is at the center of entertainment in downtown Seattle. The inside was well-themed to be what looks like a friendly old warehouse in two levels. There is a retail center selling all forms of Gameworks clothing, a bar for adults only, a Fries stand, a Pizza counter, a place for drinks and the ever present Starbucks. It has been reported that $20m went into the facility but we came away with the impression that this is just a container for arcade games. In fact, at the recent ULI Urban Entertainment Development conference in Los Angeles, Gameworks Chairman and CEO Skip Hall stated that the problem with coin-op is not the product but the setting. This is was an apt description of Gameworks.
Players use smart cards which they buy from machines scattered throughout the floor. Each machine takes a smart card and deducts the play for each game. A small LED panel on the game provides the price. Just as the new Sega games are setting ever higher price levels for the coin-op operators to buy Gameworks did the same at the retail level. Here is a sample of the play prices:
Indy 500 $4.00 Daytona $2.00 Super GT $2.00 ManxTT $1.50 Alpine Race $1.50 Virtua Fighter 3 $1.25 Prop Cycle $1.00 Classic Games $0.50 PC Use $2.00/hour
A number of old coin-op games (Classic Games) were in a U-shaped area on the second floor and we were surprised at the number of players. Titles included: Missile Command, Galaxian, Popeye and Ms. Pac Man. Yet, to pay $0.50 for these games seemed high when the prices to buy many of the old games, at auction, run less than the cost of a home video game cartridge or PC game on CD-ROM. There were 11 portable PCs set up for game play and this area was only sparsely occupied.
We did a survey of the facility and found 305 present at 7 pm on a Friday night, of which 35 were in the bar. The demographics were diverse. The lower age of children was 12 with the majority of those present in the 16 - 24 age bracket. However, we saw a number of adult couples above 30. Game play was dominated by males but there were a number of female players. The visible staff on the floor was 52.
Given that this was only a few weeks after opening the level of excitement is still high. The key in the out-of-home market is sustainability. It is here were we have our greatest doubts. The question is simple: Will a fancy container sustain what is little more than a high priced coin-op? Granted that it is the social atmosphere which plays a major role in the success of a facility we question if Gameworks brings much to the social component. Dave & Busters certainly does much more than provide games and their success is a reflection of that. In other words, will the social context of the facility offset the high price of participation? Based on our initial review we are skeptical in its present form.
707.6 Pixel - Latest Computer Animation Report by John Latta
To be published on April 15th is the latest report on the Animation Industry - The Roncarelli Report on the Computer Animation Industry - 1996. The report highlights and documents important shifts in the 3D and animation industries. Commercial animation rose 33% from 1995 to 1996 to $11.6B and if the trend continues the animation industry could reach $30B by 2001. There are significant shifts taking place, as described in the report, from the SGI platform to the PC and in particular Windows NT. The report states that the PC is now the dominant platform in animation. The Roncarelli Report is the best source for information on the animation industry and the 1996 edition will continue to meet a critical need for information on this marketplace. Price - $1,395, additional copies $175.
Pixel - The Computer Animation News People
109 Vanderhoof Avenue Suite 2
Toronto, ON Canada M4G 2H7
(416)424-4657
(416)424-1812 FAX
e-mail: pixel@inforamp.net
707.7 Conference Report - WinHEC by John Latta
WinHEC exceeded expectations again this year with over 4,000 attendees. Microsoft Chairman Bill Gates again led the conference with a keynote talk on "The PC Industry - Opportunities for Growth." Two themes stood out: making the PC simpler to use and more manageable and taking the PC platform into new markets beyond personal productivity. In particular Microsoft wants to increase the penetration of the PC into the home with the Windows CE OS playing a critical role. Microsoft is also laying the groundwork for the potential integration of the PC and TV. Part of this was addressed in detail in a session on "Building a Broadcast PC."
Bill Gates stated that Windows 95 upgrades are selling at a rate 40% higher than a year ago. Since August 1995, 60 million users have purchased Windows 95, either as an upgrade or with a system. In the last 4 months the ship rate for Windows 95 has been 4M/month. There are over 3 million Windows NT users and this is climbing at the rate of 200,000 per month. Microsoft is investing $1 billion a year into Windows development. Its Internet browser, IE, now has from 25% to 30% market share. It was stated that the drive for simplicity is behind all Microsoft investments.
The schedule for the next major upgrades for Windows was only partially discussed. At WinHEC, a Developers Beta for Memphis, the next version of Windows 95, was released and is to be followed by Beta 1 in Q2. However, Microsoft was clear to state that a final version of Memphis would not be disclosed until the results of the Beta 1 feedback has been received. The Beta for Windows NT 5.0 would not be released until Q3 and no statement was made on its final release. There was also some discussion that the OEM release of Memphis might proceed the retail release.
A major thrust between Memphis and Windows NT 5.0 is to bring consistency between these two OS's. Key features to be shared include: the Win32 Driver Model (WDM), OnNow/ACPI (power management), Plug and Play, Streaming media, DVD and Direct X. Collectively these amount to a significant advancement for the PC platform. Although the move to a common driver model may seem esoteric, this will make it easier to upgrade PCs and provide support for a broader class of peripherals. Microsoft is pushing two new buses which will decrease the amount of internal upgrading which user must do: USB for low speed serial devices such as keyboards and 1394 for high speed devices such a video cameras.
Intel announced a 4X version of AGP but no details were provided. This would put the bandwidth at approximately 1GB/sec.
Jay Torborg, Director, Graphics and Multimedia, Windows Operating System Division, described significant extensions to Microsoft's approach to multimedia. With DirectX 5.0 to be shipped in June, Microsoft will redefine the low level DirectX API's as DirectX foundation. The next layer up, which includes ActiveMovie, ActiveX Animation, Direct3D Retained Mode and DirectPlay will now be called DirectX Media. The intent is that there will be a high level access to media services and there will be a unified component architecture for all media processing which is language independent. Further, this architecture is supported by the ASF storage format for streaming media. DirectX 5.0 will be released for Windows 95 redistribution in June and this will be fully supported in Memphis and Windows NT 5.0. An area which Jay stressed was the need to improve driver quality. Microsoft has done much to improve its documentation and it is looking for the developer community to do its part. Microsoft wants to engage the developer community more and is forming design review groups. Hints were given on DirectX 6.0 which will include full Talisman support, texture transforms, multiple textures per polygon and bump mapping. In order to give game developers more time before the holiday season Microsoft expects to ship DirectX 6.0 in the spring of 1998.
Ty Graham, Technical Evangelist, described how 2D-only graphics will be dead in 1998 as the mainstream will have shifted to include 3D. He stated that Microsoft wants to kill the use of the Tunnel example in Direct 3D as a benchmark. It is his view that the industry needs real applications to test 3D performance.
From 3Dfx, Scott Sellers, VP Engineering, characterized the minimum update rate is 30 fps and the goal for compelling content is 60 fps. He stated that 3D hardware performance is quadrupling every 12 months. His ideal 3D accelerator is one with multiple independent texture streams, "infinite" texture memory via fast downloads and compression and no bandwidth sharing with the host microprocessor. His road map for 3D game accelerators goes to 100 - 125 Mpixels/sec and 2 - 3 M triangles/sec in 1998 and to 200 - 250Mpixels/sec and 4 - 5 M triangles/sec in 1999.
At WinHEC a draft of the PC 98 requirements were released. This is a very significant document because it carries with it the requirements for the Microsoft Windows logo. Although in draft form the implications for the PC being an important 3D platform are clear.
The Basic PC will move up to a 200 MHz CPU with 32 MB of memory and a graphics display of 1024 X 768 X 16 bits. ISA will be eliminated. The buses included will be PCI, USB and IEEE 1394. The 3D performance guidelines include: 1 M triangles/sec with 40 M pixels/sec of bilinear filtering and a recommended rate to 60 M pixels/sec. A set-up engine is also recommended. These are required, as per the draft, for the entertainment PC and recommended for the Consumer and Office PC.
ATI announced Rage Pro which will ship in June with pricing TBD. The memory supported includes: DRAM, EDORAM, SDRAM, SGRAM and WRAM with bandwidths up to 800 MHz across a 64bit interface. The WRAM interface can be at 128 bits and an external DAC and memory configurations from 4 MB to 16 MB. The chip has 4 KB of texture cache. Included on the chip is: power management, a video engine, AGP & PCI interfaces, RAMDAC, 3D and setup engine. The RAMDAC operates at 230MHz which will support 85MHz refresh on a 1600 X 1200 display. The video features include 720 pixel wide horizontal lines with MPEG-2 motion compensation. The set up engine is rated at 1 Mpixels/sec. This is one of the first chips to use texture compression and their implementation is based on vector quantization methods. ATI claims 8:1 compression.
Three packages are supported in the PRO family: Pro-133 in a 256 pin BGA package, Pro-66 in the same package at the Pro-133 but for cost sensitive packages and Pro-PCI in a 208 PQFP package which is pin compatible with the 3D RAGE II and RAGE II+ packages and intended as an upgrade part.
NEC announced the PowerVR PCX2. This is an upgrade from their earlier chip and it now includes bilinear interpolation, however, it does not include 2D. The units will ship in May and pricing is $35 in quantities of 10,000. The memory technology is SDRAM. The process technology is .35 micron. It is expected that end-user cards will cost from $150 - $199.
Aureal - Bringing Spatial Sound to the PC
Aureal was showing at WinHEC its Aureal 3D sound. This is the first product to be compliant with Microsoft's DirectSound 5.0 standard. The technology allows developers to create compelling audio with only two speakers or headphones. The technology makes possible 3D sound placement which is fully spatial and interactive. Diamond Multimedia will be shipping the first card (Monster Sound) using the technology in 2 - 3 weeks. This card will cost from $149 - $199 and include 5 games. A number of game developers have announced support of the technology and they include Lucas Arts, Acclaim, Sierra On-Line , Spectrum Holobyte and Virgin Interactive.
WinHEC was the first stop in the race to win in Q4 1997 with compelling 3D on the Windows platform. The next two stops will be the Computer Game Developer's conference later this month and then E3 in June where entertainment content will surface. 1997 bears close watching to determine if this will be the year 3D comes alive as a market driver. 1995 and 1996 were 3D push years and we are waiting to see if 1997 will be a pull year. WinHEC was also a stimulant to examine the broader context of what is shaping the PC as a platform, how this is impacting the industry and the role 3D is playing.
In the next 90 days OEMs will make decisions on the 3D chips they will use for holiday shipments. At WinHEC we saw new offerings from nvidia, ATI and NEC. 3Dfx has disclosed it also has a new chip coming. Bill Gates was showing the new Rendition chip in his demo but the company has not publicly announced it. Intel was hardly heard from other than its normal statement that 3D is important and the platform design must be balanced. Yet, Intel has the most to lose in Q4 if the execution in 1997 is less than optimum. As with MMX in 1996, Intel proposed a major platform initiative which did not get delivered until 1997, the same is looking increasingly likely for AGP. AGP, which uses system memory for texture storage, requires Memphis and Windows NT 5.0. Yet, it remains unclear and, in fact, unlikely that Memphis will ship in time to have demonstrable impact on the market in Q4. Thus, other chip companies must rely on non-AGP solutions to bring their high performance parts to market. Intel is rumored to be working on a part code name Auburn in conjunction with Lockheed Martin's Real 3D. Thus, Intel not only supplies the motherboards which many OEMs are counting on, it is in direct competition with most of the chip companies. We see the major competitive battle for large OEM wins being between Intel and S3. In spite of the emergence of new 3D chips there exits considerable uncertainty on what platform will emerge to deliver 3D capabilities in 1997. It remains clouded if this will be the year 3D emerges.
There is a more global competition taking place - who shapes and controls the PC platform? Increasingly we see that this is between Intel and Microsoft. In the graphics space Intel proposed last year GC97 which has largely been ignored. The document with the most impact has clearly been PC 97 because it carries with it Microsoft's Windows logo requirement. Microsoft raised the barrier again at WinHEC with the release of PC 98. Intel spoke of AGP 4X, which we took as an indication that the current AGP has insufficient bandwidth, a view held in the industry since AGP was announced.
Microsoft has a close barometer on the PC industry because of the support burden it carries. This is driving it to make PCs simpler to use and to enhance the "user experience." Thus, the emphasis on driver quality and the move to closed-box PCs. Microsoft wants to significantly reduce its support costs and it is translating that into action to change the platform. In the highly fragmented PC industry there is a desire for leadership to set the platform standards and Microsoft has stepped up to this role. Its hardware quality labs also serve a valuable role in testing and this is now being extended to include drivers for 3D. Thus, even in hardware, Microsoft is the leader in defining the future of the PC. Microsoft also recognizes that if it is to continue to maintain or accelerate home adoption of the PC major platform changes lie ahead. Windows CE certainly plays a role here. This has the potential of being the foundation for many new PC based products and ones in which Intel could play a small or even minor role.
We believe that much can be surmised from the underlying motivations of these companies in this competitive environment. Intel has made it very clear it seeks to sell more microprocessors and driving demand for visually compelling PCs is one approach. In short, Intel must continually drive up the demand for MIPS. Any fall off in MIPS demand will limit the profits it generates and thus its ability to build new fab facilities. It is a circle of demand and profits which is only self sustaining as long as the demand for Intel microprocessors continues to grow. Yet, Intel only sets part of the hardware agenda and not the software standards. Its attempts in the past, such as NSP, have failed. In contrast, Microsoft must continue to drive PC penetration so that it gains revenues from its Windows products and applications. At WinHEC we saw Bill Gates's emphasis on moving the PC into new applications, especially in the home. Thus, only by sustaining demand, will Microsoft be able to see its stock price continue to rise. Microsoft does not have to be exclusively tied to Intel to achieve its objectives. Another way for Microsoft to improve its performance is to also decrease its support burden and this is fully consistent with the presentations at WinHEC. Microsoft has obvious motivations for driving the platform and in the process Intel has an important but potentially less critical role, especially with the Windows CE platform.
At WinHEC it was clear that 3D is integral to the evolution of the PC platform. These chips also fit within another larger whole. The display chip has become really a multimedia, graphics and display processor. In the process it is becoming the second most complex piece of silicon in the PC. This is a major market opportunity, just as PC logic chips were in the mid-80's. Intel cannot afford to ignore this opportunity because, if not controlled by it, its control of the platform could erode. By the same token Microsoft is driving 3D technology with Talisman, which also has multimedia capabilities. Again, we find that platform control is where the action will be as the PC industry seeks to maintain its 20% + growth rates into 2000 and beyond. 3D maybe a bit player but a very important one at that.
707.8 Conference Report - WDM Device Driver Conference by John Latta
Held on 11 April after WinHEC this was another packed conference. With over 2,000 attendees, 250 signed up for the sessions on Graphics Drivers. The conference was over-subscribed and the attendance had to be cut off on Thursday before the event. The WAVE Report attended the Graphics session and even here all those who wanted to attend could not fit into the one conference room. A separate video session was held simultaneously across the hall.
The presentations were technical and provided both details on driver development and access to key individuals at Microsoft. At the end of each session there were many questions. Microsoft has surfaced its concerns about the quality of drivers and this was an opportunity for the developer community to learn. At the same time a number of technical details came out of interest to WAVE Report readers.
Both Memphis and Windows NT 5.0 will support MultiMonitors. This will allow more than one graphics adapter to be used at one time. The ways in which this can be used include: multiple independent displays, large desktops and multiuser support (NT 5.0). There is still a requirement of one VGA capable adapter but there were hints that this may eventually go away.
Some hints were also made about future capabilities in Direct3D. These include richer geometries including other types of surface descriptions, progressive meshes and on the fly tessellation. Support for hardware transform and lighting was discussed. However, the lighting being considered would only be accomplished as multiple textures per surface and not traditional lighting done at the front end of the pipeline. All of this is in a very early state and attendees were encouraged to attend Meltdown 97 in July.
707.9 SGI Extends OpenGL by John Latta
SGI is making significant moves to establish OpenGL as the dominant imaging API on both workstations and the PC. Some early indications of this surfaced at SIGGRAPH 1996 and now a clearer picture is beginning to emerge.
As a major extension to OpenGL, SGI has proposed to the OpenGL ARB that a scene graph layer be added above OpenGL, currently called OpenGL++. This nomenclature is used to indicate an object-oriented approach with a higher-level description of a 3D API. Based on the experience learned from Performer this layer was described earlier at SIGGRAPH 1996 as Cosmo 3D whose terminology has since been dropped. At the February 17 - 19 ARB meeting it was agreed to move forward "on developing a scene graph API."
SGI is going beyond this and creating vertical market extensions to OpenGL. We first reported on this at Autofact with the Intelligent Simplification technology. The first implementation of this is known as the OpenGL Optimizer to improve rendering and interactivity of large CAD/CAM/CAE data sets. As a result of a meeting sponsored by SGI, IBM and Division it was announced that Dassault Systems, Division, Parametric Technology, Prosolvia Clarus AB, SGI and Structural Dynamics will work to define a standard API for CAD rendering. It is expected that this will result in a consortium to back an open API for this vertical market which is tentatively called the Large-Model Visualization API Consortium (LMVAC). We expect that SGI will announce further extensions in the Visual Simulation and Image Processing areas.
SGI has embarked on a three-layer approach to the OpenGL API that could have significant implications to the 3D industry. The lowest level is OpenGL as it exists today, the next level up is an object-oriented level which stresses ease of use and access and the highest level is an application-specific layer.
http://www.sgi.com/Technology/openGL/index.html
Copyright 1997 4th WAVE Inc.
May be redistributed in full for individual readership and posted to newsgroups, Web, and FTP sites. May not be reprinted or redistributed for profit. Short quotes are permitted but must be attributed to the WAVE Report on Digital Media.