Encoding Performance Benchmarks
We're going to start off the benchmarks with what we think is the most important statistic: encoding speed. Using five CPUs with and without the assistance of the WinFast PxVC1100 card, here are the results:
There are two main things we can see very quickly. First, the TMPGEnc 4.0 XPress application takes good advantage of threading. See how poorly the Athlon II X2 250 is performing, despite its high 2.9 GHz clock speed? Secondly, note just how quickly the WinFast PxVC1100 encodes video, even when paired to the Athlon II X2 processor.
The lesson here is that if you're planning a PC build with a primary function of video encoding, the speed advantage of the WinFast PxVC1100 is very hard to ignore, regardless of the CPU that you choose to deploy in your machine. Even the slowest Athlon II X2 250 managed to encode the video in about seven and a half minutes using the WinFast card, compared to about five and a half minutes with the Core i5-750/WinFast PxVC1100 combo. For comparison purposes, the Core i7-920 completed the task in just over 11 minutes without the SpursEngine helping it along.
We'll explore the cost/benefit ratio below, but for now let's have a look at the video up-scaling performance that the SpursEngine processor offers. We encoded a short 720x480 video sample into the higher HD 1920x1080 resolution using the CPU, the WinFast card, and then the WinFast card again in conjunction with the super-resolution option:
Enabling super-resolution increased the time to encode the video by more than two times as long, which was even slightly longer than with the pure-CPU encoding method. This isn't so bad if the end result is stunning, but if it doesn't provide a notable increase in quality, then this is a hard pill to swallow.
Next, we're going to have a look at the image-quality aspects of software versus hardware encoding.