The question, of course, is whether it’s true. Analog Devices’ whitepaper cites tests conducted in Creative’s, er, labs, but we’re not happy until we can explore an issue with our own systems in an environment we can control. So, we’ve rounded up a couple of nearly identical motherboards with Analog Devices and Realtek codecs to see how they handle EAX positional 3D audio, with conclusive results.
Much ado about EAX
EAX, otherwise known as Environmental Audio eXtensions, is a positional 3D audio standard with roots in Creative’s SoundBlaster Live. The standard tags in-game sounds with information about their position in the world, allowing for more realistic interactions with the player. Through EAX, positional audio is mapped to the appropriate speakers in a multi-channel setup, giving the player a sense of direction associated with each sound. EAX can also modify in-game sounds to take into account obstructions like walls and pillars, and the occlusion effect of different materials, such as wood and glass, between the player and a given sound.
The most recent incarnation of EAX is EAX Advanced HD 5.0, which can be found on Creative’s latest X-Fi audio cards. You don’t need a fancy audio card to enjoy EAX effects, though. Creative made EAX 2.0 a public spec—ostensibly to push its own 3D audio technology as the standard for positional audio in games—allowing other sound card and codec manufacturers to support it.
EAX 2.0 was originally handled in hardware by the SoundBlaster Live’s EMU10K1 audio chip. On motherboards, EAX 2.0’s positional audio calculations are now most often handled in software by the onboard audio codec’s driver. In fact, core logic chipset makers have largely washed their hands of 3D audio support, deferring to codec makers, whose sound drivers do the heavy lifting via Microsoft’s standard High Definition Audio bus.
Codec chips from Analog Devices and Realtek currently dominate the scene for integrated motherboard audio, and the latter is far and away more popular than the former. Fear the crab, baby.
Back in the day, it was nearly impossible to find a motherboard that didn’t use Realtek’s ALC650 codec. The company has done an admirable job of maintaining its strong market share since, becoming the de facto standard for eight-channel AC’97 codecs and now for the new wave of high-definition “Azalia” audio chips.
Analog Devices’ codecs haven’t been nearly as popular among motherboard makers as those from Realtek, but that may be starting to change. Asus recently switched from Realtek to Analog Devices codecs for its high-end motherboards, and since they sent us a copy of Analog Devices’ EAX whitepaper, we suspect it had something to do with their decision to buck the crab.
Our testing methods
To put each codec’s EAX implementation to the test, we assembled a couple of similar nForce 590 SLI motherboards from Asus and Foxconn, one with Analog Devices AD1988B codec, the other with Realtek’s ALC882D. Both chips claim to support EAX 2.0, and each was configured with the latest version of its drivers. Nvidia actually used to provide its own 3D audio drivers for nForce chipsets, but the company is now relying solely on codec makers for 3D audio support and has removed the 3D audio drivers from its ForceWare driver package.
We used the following test systems.
|Processor||Athlon 64 X2 5000+ 2.6GHz|
|System bus||HyperTransport 16-bit/1GHz|
|Motherboard||Asus M2N32-SLI Deluxe Wireless Edition||Foxconn C51XEM2AA-8EKRS2H|
|North bridge||nForce 590 SLI SPP||nForce 590 SLI SPP|
|South bridge||nForce 590 SLI MCP||nForce 590 SLI MCP|
|Chipset drivers||ForceWare 9.35||ForceWare 9.35|
|Memory size||2GB (2 DIMMs)||2GB (2 DIMMs)|
|Memory type||Corsair TWIN2X2048-6400PRO DDR2 SDRAM at 742MHz|
|CAS latency (CL)||5||5|
|RAS to CAS delay (tRCD)||5||5|
|RAS precharge (tRP)||5||5|
|Cycle time (tRAS)||12||12|
|Audio codec||Integrated nForce 590 SLI MCP/AD1988B with 188.8.131.5230 drivers||Integrated nForce 590 SLI/ALC882D with Realtek HD 1.47 drivers|
|Graphics||GeForce 7900 GTX 512MB PCI-E with ForceWare 91.31 drivers|
|Hard drive||Western Digital Caviar RE2 400GB|
|OS||Windows XP Professional|
|OS updates||Service Pack 2|
Our test systems were hooked up to 5.1-channel Logitech Z680 speakers using each motherboard’s analog audio outputs. The audio drivers were set to use 5.1-channel output.
Thanks to Corsair for providing us with memory for our testing. 2GB of RAM seems to be the new standard for most folks, and Corsair hooked us up with some of its 1GB DIMMs for testing.
Also, all of our test systems were powered by OCZ GameXStream 700W power supply units. Thanks to OCZ for providing these units for our use in testing.
We used the following versions of our test applications:
The test systems’ Windows desktop was set at 1280×1024 in 32-bit color at an 85Hz screen refresh rate.
All the tests and methods we employed are publicly available and reproducible. If you have questions about our methods, hit our forums to talk with us about them.
RightMark 3D Sound
We regularly cover the CPU utilization of integrated audio solutions in our chipset and motherboard reviews, so this time, we’ve narrowed our focus to positional audio accuracy. RightMark 3D Sound has a handy positioning accuracy test that lets one arbitrarily define a sound’s spatial position to see how each audio implementation handles it.
When dealing only with simple positional audio, we didn’t observe any differences between the Analog Devices and Realtek codecs. Each appeared to reproduce audio faithfully based on its position relative to the listener, taking both direction and distance into account.
Next, we fired up RightMark 3D Sound’s occlusion test, which lets the user observe how sounds interact with a pair of occlusion fields.
In the occlusion test, the Analog Devices codec smoothly muffled sounds between the listener and the occlusion fields. The Realtek codec did not, however. It behaved as if the occlusion didn’t exist, failing to modify sounds in a discernable way regardless of whether they were in front of or behind the occlusion field.
Moving to RightMark’s obstruction test, we observed similar behavior from the Realtek codec. It acted as if the obstructions simply weren’t there, while the Analog Devices codec modified sounds as they passed behind each obstruction. The difference between the two codecs couldn’t have been starker.
RightMark 3D Sound’s synthetic tests do a good job of highlighting the Realtek codec’s failure to support EAX occlusions and obstructions correctly, but do those features really matter? To find out, we fired up a couple of popular games that support EAX.
Battlefield 2 was configured with the game’s high-quality hardware audio option. We played through a 16-player game on the Strike at Karkand level.
We fired up Battlefield 2 on our Analog Devices-equipped system first, and didn’t encounter any problems with 3D audio. Sounds came from the appropriate directions, faded with distance, and were accurately affected by buildings, fences, and other in-game objects.
The Realtek codec, on the other hand, was a disaster. Sounds seemed to be coming from the right direction, but there were entirely too many at once. Further investigation revealed that in-game objects had little to no impact on in-game audio; tanks located on the other side of buildings—or even the other side of the map—were as loud as they were driving right next to you.
Battlefield 2 occludes sounds over distance to take air absorption into effect, but the Realtek codec didn’t seem to be taking that into account at all. Even when out in the open and free from obstructions, sounds didn’t correctly fade over distance.
This apparent lack of proper obstruction and occlusion support effectively made Battlefield 2 unplayable with EAX 3D audio on our Realtek-equipped system. You literally hear every sound on the battlefield as if it’s right next to you, plunging the player into aureal chaos.
Battlefield 2 is a chaotic multiplayer game with loads of in-game audio, so we scaled things back a little for another test, playing through early levels in single-player FEAR. The game was set to use hardware mixing and EAX 2.0 effects.
Again, we launched the game with our Analog Devices system first, and we failed to encounter any problems along the way. Bracing ourselves for the unplayable chaos we experienced in Battlefield 2, we ran the game on our Realtek-equipped system, and were pleasantly surprised to find a much more subdued environment. Obstructions and occlusions didn’t appear to be working correctly—we’d hear sounds through walls and windows that we didn’t with our Analog Devices system—but that didn’t affect the gameplay to a great degree.
Of course, FEAR is a very different game than Battlefield 2. Plodding through a largely linear single-player level doesn’t produce a lot of simultaneous sounds in different locations, especially if audio is triggered by scripts based on the player’s physical location.
There are clearly problems with the EAX implementation of Realtek’s HD Audio codec and its associated drivers. Despite the fact that Realtek claims to support EAX 2.0, obstructions and occlusions apparently haven’t been implemented correctly, if at all. For some games, that’s not a serious problem. For instance, FEAR plays fine, even if the audio isn’t exactly right. However, the lack of proper occlusion and obstruction support makes Battlefield 2 unplayable with EAX effects enabled.
We observed these problems on a Foxconn motherboard with an ALC882D codec, but any motherboard that uses a Realtek codec for High Definition Audio is likely to be affected. After all, Realtek handles 3D audio in the driver, which is shared among all of its HD Audio codecs. (Realtek’s AC97 audio drivers may have similar problems, but we haven’t tested those.) Fortunately, Analog Devices appears to be doing EAX right, but that doesn’t help if you already have a board based on a Realtek codec.
So we have a problem. If Realtek wants to tout EAX 2.0 support, that support needs to be complete and functional. Game developers need to be assured that if they devote the effort to supporting EAX, their users will experience the game as intended. As it stands, those efforts will be wasted on motherboards with Realtek integrated audio. Fixing the problem should be possible through a software update, and that needs to happen as soon as possible if Realtek wants to continue claiming support for EAX 2.0.