1. Another Round with DHCAT: Some Interesting Results

We’ve written about the Intel Digital Home Capabilities Assessment Tool, aka DHCAT, in several recent stories, most notably on June 11 in We Test Our BYO HTPC’s With Intel’s DHCAT. In conducting that review and after subsequent analysis, we were stunned to observe that what we thought were some powerful and capable Home Theater PCs (HTPCS) didn’t fare too well on that test. As we pondered what DHCAT really tests, we formulated some theories about what it takes to ace that test, so to speak. For this story, we built some PCs to embody those theories, and to report on the results.

In building and testing those machines, we also had an interesting opportunity to put a claim from one of the Intel engineers we got to know while working through the last round of tests as well - namely, that Windows Vista produces results that are as much as 20% better on the same hardware as Windows XP when running DHCAT. We’ll take you through our theories, describe the systems we built and report on the results for both Windows Vista and Windows XP (MCE 2005, actually) in this story.

Re-introducing DHCAT

Let’s start with a quick DHCAT refresher (those interested in even more detail should consult our previous story on this topic). Intel designed DHCAT to focus on a PC’s media-handling capabilities rather than simply observing how quickly the CPU could handle certain processing tasks, how quickly the disk drives could read and write data, how fast memory could be accessed and so forth. What we like about DHCAT is that it’s designed to model reasonably common (if sometimes ambitious) types of media recording, playback and encode/decode activities, so we do think it has some interesting news to convey in the results it delivers when it’s run on one PC or another.

The 10,000-foot Overview Of DHCAT, Revisited

Essentially, what DHCAT does is to enact a number of scenarios that put a PC to work handling various kinds of media activities. It then measures the performance that results as those tasks are completed (or not, as the case may sometimes be). Let’s begin with the list of elements that DHCAT looks for as it begins its assessment:

DivX Pro 6.4 codec or newer: We used the DivX Pro 6.5 codec in our testing, because it’s the latest, most readily-available version.

SDTV and HDTV tuners: We started out with AVerMedia M780 combo SDTV/HDTV tuner cards (aka AVerTV Combo PCIe) but quickly learned that DHCAT wouldn’t recognize the HDTV tuner on those cards. We quickly added an ATI TV Wonder 650 PCI HDTV Tuner card to the mix, because DHCAT did recognize that hardware.

HD Audio: We used RealTek ALC889A chips for both sets of tests, because that’s what comes on the Gigabyte GA-P35T-DQ6 and GA-P35-DQ6 motherboards we used for testing. This chipset meets DHCAT’s requirements for high-definition audio.

A UPnP media server that complies with criteria set by the Digital Living Network Alliance (DLNA), and also includes the ability to transcode Windows Media Video (WMV) and DivX to MPEG2 format. Because these boards don’t support Intel ViiV, we used the Nero Home media server for our tests.

As it goes through its paces, DHCAT tests a media PC’s abilities to handle the following tasks:

Encode, play, record and transcode standard TV and HDTV video.

Synchronize and edit digital photos.

Serve, stream and transcode media to one or two devices across a network, with or without compression - this is where the DLNA Media Server comes into play.

Rather than actually using real hardware-based SDTV and HDTV tuners, or interacting with a server across a real network, Intel chose to stub these activities out to virtual devices and interfaces so as to keep its testing entirely focused on the machine under assessment. We followed that reasoning to its logical conclusion by also installing UPnP media server software on those same machines as well (as they do themselves).

What makes Intel’s assessments worth watching - and its results somewhat relevant to those who might wish compare one system against another - are the combinations of individual activities that Intel includes in its various test scenarios. The company also invested in some useful automated Mean Opinion Score (MOS) based video quality assessments. These allow a pristine source video to be compared to a degraded copy of the same video, to observe how well a system can play back noisy or otherwise imperfect video source material.

In the next section we step through all six scenarios in the DHCAT suite, and explain the individual assessments, from three to five per scenario, that go into each one.

Join our discussion on this topic