Test Setup
| Hardware | |
|---|---|
| CPU | Pentium III 1 GHz Coppermine Socket 370 |
| Motherboard | Asus CUSL2-C-BPBlack Pearl Special EditionBIOS 1007 |
| Memory | 256 MB Micron PC133 RAM CL2 (OEM) |
| Graphics Card | Gainward GeForce2 Pro/450 64 MB DDR - CARDEXpert |
| Hard Disk | IBM 40GXP - IC35L040AVER0741.17 GB7200 RPMATA-100 |
| CD-ROM | Sony 52X CD-ROM Drive CDU5211(Replaced With Asus 12X DVD Rom For DVD tests) |
| CD-RW | Asus CDW-1210R - 12X/10X/32X CDRW |
| Network Card | 3C905C-TX Network Interface Card |
| Add On | Asus iPannel Basic (Removed From Our Final Testing Configuration - Read on to find out why.) |
| Case | Antec SX1030 Mid Tower Case With Antec PP-352X 350 watt Power Supply |
| Speakers | Altec Lansing ADA880W Dolby Digital Speaker System |
| Drivers | |
| Creative Labs Sound Blaster Live ! Platinum 5.1 | Version SBL5-WEB-W1-US |
| Hercules Gamesurround Fortissimo II | Version 1.0 (From Enclosed CD) |
| Hercules Game Theater XP | Version 2.02 |
| Philips Acoustic Edge | Version 2.59 |
| Turtle Beach Santa Cruz | Version 4081 |
| Aureal SQ2500 | Version 2048 |
So, How Did We Benchmark These Cards ?
When testing the sound cards, we tried to test them in three ways that we felt would yield the best gauge of the card’s true "real world" performance. For the first test, we used Quake III and UTBench with and without the 3D audio option turned on ; for the second test, we look at the CPU utilization while decoding MP3 files. After running the gaming tests, many felt that due to lack of 3D audio support in the newest point releases of the game, that another test was in order. Read on to find out why.
We thought about including Audio WinBench 99, but decided against it because it didn’t give us as clear a picture of real world performance as the Quake III and UTBench results. ZD Lab’s Audio WinBench 99 really only focuses on two aspects of the testing, which are described below :
Audio WinBench 99 uses eight automated tests in the Audio CPU Utilization Tests. Four of the tests use Direct Sound, while the other four tests use Direct Sound 3D. Each test measures the amount of CPU that the sound subsystem uses for common operations, such as playing sounds at 22 kHz and 44.1 kHz, 8 bit and 16 bit from both static buffer and streaming buffer.
The Audio CPU Utilization Tests allow us to determine how much work the sound card is offloading from the processor, so that you know how much CPU is available for other tasks. If the card you have is truly a hardware accelerated sound card, the card does most of the mixing. If you have a sound card that doesn’t support hardware mixing, the processor will do most of the mixing, which means the percentage of processor usage will be much higher. With almost all Audio WinBench tests the data on CPU use was inconsequential, never rising above 4% for the worst case scenario. As a result, we ditched our Audio WinBench 99 results and decided to focus on some real world performance data.
After loading the test system with our base line Windows 98SE Ghost Image without any sound card or sound card drivers loaded, we proceeded to install the sound card for this test. We used the most current drivers provided on each manufacturer’s web site, if they were different from the ones provided on the driver CD that is shipped with each card. (We note the driver revisions for each card below.) Once the card was installed, we rebooted the PC and installed DirectX Version 8.0a. Once the DirectX 8.0a install was complete, we rebooted the PC again.
