NVIDIA PureVideo Technology
System Setup and Testing Methodology
AMD Test System Setup
AMD Athlon 64 FX-53
NVIDIA nForce4 Ultra Reference Board
Antec 480 watt
250 GB Maxtor 7200 RPM SATA
Creative Labs Live!
ATI Catalyst 4.12 & MMC 9.03
Windows XP w/ Service Pack 1
Our testing for this article focuses on two aspects: CPU utilization and image quality. CPU utilization was tested using the Windows Performance Monitor to gather the CPU utilization at each second for at least 60 seconds each run. The data was then graphed as well as filtered to get the minimum, maximum and average CPU utilizations. Lower CPU utilizations are better as they show less processor usage.
Image quality, as always, is much more difficult to define and compare. Comparing two images with technologies on and off display some very obvious changes, but in other cases (such as comparing the implementations of 3:2 pulldown) show similar images that are very subjective in nature. I'll give my thoughts on each of them as I experienced them.
Most of our testing was run in the Windows Media Player 10 application, including our WMV9 HD tests and the DivX HD tests. The exception to this is the Terminator movie since it uses its own player but the WMP10 codecs.
In our CPU utilization graphs, you'll notice there are two sets: Default and NV DLLs. The "NV DLLs" results were gathered after installing the DLL files that NVIDIA distributed to update WMP10. NVIDIA claims that these DLLs will be included in a near future update from Microsoft to the WMP10 application. They improve performance quite a bit, but I decided to include the default settings as well since ATI did not have similar DLLs ready for us at time of launch.