If you've ever read our System Guides, you know we like to harp on about sound cards. Like, a lot.
Our point is this: on most motherboards, the onboard audio just doesn't do digital-to-analog conversion all that well. The analog output may sound decent enough for cheap headphones and $50 computer speakers, but plug in anything nicer than that, and you'll notice flat highs, muddy mids, and boomy bass. All kinds of subtleties will be blurred over in music and movies.
Most good sound cards have high-quality DACs (digital-to-analog converters) with superior analog output. They typically aren't subject to electrical interference from the rest of the motherboard, either. That means even entry-level sound cards can sound noticeably better than onboard audio, provided you're listening through a pair of decent analog headphones or speakers. We've confirmed this in blind listening tests time and time again.
So, that's why we recommend sound cards.
Sound cards don't fit in every kind of PC, though. Building a Mini-ITX gaming rig? Your only expansion slot will probably be taken up by a graphics card. Using a NUC or some other ultra-small system? Forget about expansion cards altogether. The same goes for laptops, whose integrated audio is usually no better than that of desktop motherboards.
Enter Diamond's Xtreme Sound XS71HDU. This little box measures 3.3" x 3.1" x 0.9" (85 x 79 x 23 mm), plugs into your computer via USB, and essentially acts as an external sound card, complete with a Cmedia codec, 7.1-channel output, 24-bit/192kHz playback and recording, and a smorgasbord of inputs and outputs.
There are six 3.5-mm jacks: stereo headphone, rear, sub, side, mic, and line in. Also included are a set of red-and-white RCA ports, an optical S/PDIF input, a matching S/PDIF output, and a Micro USB port, which is needed for the USB connection to the host PC. In short, the XS71HDU can drive just about anything from studio monitors to headphones to 7.1 computer speakers, and it'll happily record either digital or analog audio, too. The connectivity options here are plentiful enough to put many traditional sound cards to shame.
Heck, there's even a volume knob up top. The knob has a nice, slightly clicky action, and because it controls Windows sound volume directly, it works with both the analog and digital audio outputs.
Too good to be true?
On paper, then, the Xtreme Sound looks like a fine sound card substitute. But how does its analog sound quality measure up?
To answer that question, I grabbed my Sennheiser HD 595 headphones and tested the Xtreme Sound side by side with Asus' Xonar Phoebus, a high-end PCIe sound card with a dedicated headphone amp. The Realtek integrated audio from my aging Gigabyte P55 motherboard was also in the mix, simply to provide a baseline reference. (The HD 595s don't have the kind of impedance that requires extra amplification, so they had plenty of volume even on the Realtek.)
To gauge analog sound quality, I listened to some 320Kbps and V0 MP3s from my personal collection. Some of you will cry murder at the very thought of testing sound quality with a lossy compression format, but I think there's a case to be made for that approach. High-quality compressed audio is what most people listen to these days, and in my experience, 320Kbps and V0 MP3s are virtually indistinguishable from FLAC and other uncompressed formats. This isn't just a subjective opinion, by the way: my own double-blind testing with Foobar2000's ABX comparator bears that out, at least for the kind of music in my playlist.
In any case, one doesn't need uncompressed audio to tell the Xtreme Sound apart from the Xonar (or the Realtek onboard sound, for that matter). Right after plugging my headphones into the Diamond box, I noticed a very faint background hiss in the left channel. No amount of fiddling with the connector or software configuration alleviated it. Then, in music playback, the Xtreme Sound turned out to be noticeably shrill and tinny. Mids were flat, bass was confined to drum hits, and vocals were weirdly metallic, even in quiet songs like Agnes Obel's Fuel to Fire. Busier tracks, like M83's Kim & Jessie, sounded close enough to white noise at times to be downright unpleasant. Opeth's brooding guitar riffs were a little more tolerable, but still harsh, shrill, and not nearly punchy enough.
By contrast, the Xonar Phoebus offered full and balanced sound with crisp highs and good separation between instruments, even with the headphone amplification setting turned all the way down. The Realtek audio was a lot muddier, and it obscured a fair amount of detail, but it was at least neutral enough to make prolonged listening sessions comfortable. Not so with the Xtreme Sound, whose shrillness quickly tired me out.
Happily, the Xtreme Sound offers a way to bypass the lackluster digital-to-analog conversion. Digital speaker setups can be plugged right into the S/PDIF output, which pipes out raw, uncompressed sound. That's good news for laptops and uber-small-form-factor builds, which may not have digital audio outputs of their own.
I tested the S/PDIF out with my Antec Soundscience speakers, and predictably, the sound quality was indistinguishable from that of my Realtek onboard audio's S/PDIF output. Digital speakers shoulder the burden of digital to analog conversion themselves, so provided they're fed an error-free digital signal, there should be no difference in output quality across different sound cards or audio adapters. Still, it's good to know the theory checks out in practice with the Xtreme Sound.
|G.Skill KM560 MX keyboard drops the numpad||10|
|Rumor: Acer Triton 700 may use an unreleased Pascal GPU||23|
|Silverstone Vital VT02 could hold a Core i7 in under two liters||10|
|Galax and KFA2 induct the GTX 1080 Ti into the Hall of Fame||22|
|Acer's Aspire GX-281 lineup brings Ryzen to the masses||18|
|Deals of the week: discounts on CPUs, mobos, and more||10|
|Asetek gets $600,000 from Cooler Master in AIO cooler patent spat||20|
|Acer Predator Triton and Helios laptops are ready for serious play||16|
|Intel enjoys healthy revenue and profits for Q1 2017||30|