Blind listening tests
We believe the best way to evaluate sound cards is to listen to them in blind tests. Subjective impressions provide a sense of how each solution sounds, and blind tests largely remove bias from the equation. Our
victims subjects listened to 30-second clips of various songs back-to-back on different audio solutions, and they had no idea what solution was being played when.
Since they're the limiting factor, let's introduce our listeners. They've all rated themselves on an impromptu audiophile scale between 0, which considers Apple's stock iPhone earbuds to be awesome, and 10, which classifies as garbage anything that isn't piped through multi-thousand-dollar speakers via gold-plated Monster cables from an original vinyl recording.
Brent, a friend of mine from university, was first to endure our barrage of 30-second song clips. He doesn't have any fancy audio gear and rates himself a 6 on our scale. Next, I convinced my girlfriend Mo to sit through a session. She thinks even basic Logitech speakers sound great, so we'll give her a 4 on the audiophile scale. TR staffer Cyril is definitely pickier when it comes to sound quality; he rates himself an 8. After one round of listening tests, Cyril took the controls and ran a second round with me in the hot seat. I'm not sure my ears are quite as sensitive as Cyril's, but I do appreciate good sound, and I'd give myself the same audiophile rating.
Die Antwoord — Hey Sexy
There isn't a whole lot of hip-hop in my music library, but I do have a weakness for South African rap group Die Antwoord. Hey Sexy layers vocals over looping guitars and a thumping bass line.
To my ears, Hey Sexy was "more gangsta" on the DSX than on the DGX. The former's bass line was almost over-emphasized, and the card was difficult to distinguish from the more expensive Xonar DX. The DGX sounded less boomy in comparison, with more prominent mid-range content and in-your-face vocals.
Cyril found the DGX more balanced than the other Xonars, perhaps because its bass was less prominent. He thought the DSX sounded a little metallic at the high end, as well. Cyril actually liked the sound of the Realtek codec versus the DGX and DSX.
Our other listeners agreed with the general sentiment that the Xonar DSX and DX offered the most bass of the bunch. Mo had a hard time telling the difference between those two, but Brent thought the DSX thumped a little harder. He also heard more pronounced vocals on the Xonar DGX, something that Mo didn't notice. Neither was particularly keen on the Realtek codec, which was singled out for lacking richness, being too tinny and rattly, and having too much treble.
LCD Soundsystem — Home
In the realm of increasingly difficult-to-classify modern music, LCD Soundsystem is perhaps best described as low-fi indie electronica. Home is from the band's final album. The track is an upbeat number filled with different instruments, some of which sound more familiar than others.
The Realtek codec failed to impress my ears with this song, and Cyril thought it sounded a little neutered. For me, the Xonar DSX offered better separation between the various instruments than the integrated motherboard audio. It also sounded more balanced than the DGX, which emphasized the mid- and high-range notes. That emphasis wasn't as apparent versus the Xonar DX.
Cyril and I both found the Xonar DSX and DX to be comparable. He seemed to prefer the Xonar DGX over the DX due to the DGX's greater separation between the song's percussive elements. When the DGX was played back-to-back with the DSX, though, that separation sounded a little over-sharpened to his ears.
With this track, our other listeners struggled to tell the different configs apart. Mo preferred the DSX to the DGX in their head-to-head matchup, but she couldn't explain why. Brent favored the DGX, which he said had more distinct vocals and treble.
Adele — Rolling in the Deep
It's rare to hear music I like playing in malls and department stores, but Adele's 21 is one of those albums. The first track, Rolling in the Deep, combines Adele's soulful voice with background vocals, bass, and a building piano track.
Once again, the Xonar DSX was a close match for the DX. All of the listeners agreed on that point, although there were differing opinions about the subtle differences between the two. To me, the DSX felt farther from the stage. Brent said it had deeper lows, but Cyril thought the opposite, adding that the DSX's vocals sounded a little metallic and less fleshed out than on the Xonar DX.
There was less consensus regarding the DGX, which some said offered clearer backing vocals and percussion. However, Cyril thought the DGX's vocals sounded like a "robot singing through a tin can." He also found the card's mid-range frequencies to be a little compressed, at least versus the other Xonars. Cyril didn't think the Realtek codec sounded as good as the DGX, but it sounded better to me due to a more even distribution of frequencies. To my ears, the DGX's vocals were over-emphasized, with a hint of distortion. Mo heard some distortion in the DGX's percussion, too.
Mo was reasonably happy with the Realtek codec overall, although she said it sounded a little tinny. Brent was less impressed, saying the Xonars were "just better" in all their matchups. With few exceptions, Cyril and I agreed. To my ears, the onboard audio didn't have as much body as the Xonar DSX. Cyril repeatedly called the Realtek solution neutral, but he did feel that its lows were a little muted compared to the discrete cards.
The Tea Party — Sister Awake
Apparently, few people outside of Canada and Australia have heard of The Tea Party, which sounds a little like Jim Morrison singing for Led Zeppelin somewhere in eastern Asia. The snippet of Sister Awake we used is purely instrumental and filled with multiple string and percussion instruments.
In our marquee matchup, the Xonar DSX versus the DGX, the results were mixed. Mo couldn't tell the difference between the two, but the rest of us could. Cyril thought the DGX was a little bit louder, and that the DSX's drums blended in with the mid-range a little. Brent said the DGX offered too much treble at the high end of the spectrum, to the point that it sounded tinny. I agreed with those sentiments and thought the percussion hit harder on the DSX.
To my ears, the Realtek codec sounded squished together compared to the DSX. I did think the onboard audio was pretty close to the DGX, though. Cyril thought the Realtek solution's drums were a bit subdued, with more of a focus on the mid-to-upper range of the spectrum.
Personal preferences inevitably taint subjective impressions, especially when subjects are simply asked what they hear. That's probably why Brent preferred the motherboard audio's more prominent mid-range tones to the stronger bass on the Xonar DX and DSX. Mo was firmly in the Xonar camp across the board and liked the DSX more than the others.
Tom Waits — Tell Me
Tom Waits' unmistakable voice sounds like what might happen if one gargled gravel every day for 20 years. In Tell Me, Waits' rough vocals are complemented by subtle percussion, sparing guitars, and what I believe is a xylophone. With Waits, one never knows.
Half of our listeners couldn't distinguish between the Xonar DGX and DSX this time around. Cyril and I had little trouble, however. He found the DGX's mids too sharp and the DSX more natural, while I thought the DSX had more low-end grunt and sounded a little subtler than the DGX.
I preferred the DSX to everything else it went up against. Cyril liked the DSX over the Realtek solution, which he said had no warmth, but he said the vocals were crisper on the DX.
Neither of us really liked the motherboard's built-in sound, calling it compressed and "slightly wrong." Brent was similarly critical of the Realtek audio, saying its rivals sounded fuller in comparison. He called the Xonar DX and DSX similar, and said Waits' deep voice was accentuated by the DGX. Cyril made a similar comment, noting that Waits' voice sounded more natural on the DGX. That may be the only time such a raspy bellow has ever been described as natural.
Mo didn't mind the Realtek audio as much as the rest of us, perhaps because listener fatigue had set in. She thought the Xonar DGX was a little crisper than the other Xonars, though.
The inevitable summary
Our listeners all thought the different audio solutions were more closely matched than in any of the listening tests we've conducted before. Their assessments of each config were largely consistent, but some of the songs and matchups teased out contradictory opinions. More often than not, the Xonar DSX was identified as having deeper bass and a fuller sound than the DGX. The DGX's mid-range bias was definitely apparent, and its output was often deemed sharper and crisper than the DSX and its other rivals.
Although it fared better than any other integrated audio implementation we've tested, the Realtek codec was clearly inferior to the DGX and DSX overall. The onboard audio was definitely short on bass, and it lacked the sharper mid-range tones of the DGX.
We weren't surprised to see the motherboard audio fall to the bottom, but we didn't expect the Xonar DSX to so closely match the pricier DX. Those two sounded more alike than any other pair, and our listeners usually preferred the DSX.
A few words about gaming audio quality
During our music listening tests, subjects had the luxury of closing their eyes and concentrating on sound alone. Getting a sense of gaming audio quality is more difficult because the soundtrack tends to fade into the background when you're actually playing. We didn't run a full set of blind listening tests in games, but I did take a few notes while testing Battlefield 3 and DiRT Showdown.
The biggest takeaway was that I didn't hear obvious differences in audio quality between the various solutions. Perhaps I was too distracted by the visuals and trying to play through the games in a repeatable fashion, so that I'd hear the same mix of sounds. Maybe the differences faded when it took minutes rather than seconds to swap sound cards.
There were, however, very big differences between the audio configurations offered by each game. Battlefield 3's Enhanced-mode surround virtualization added a real sense of immersion in the environment, especially with gunfire coming from all directions. The Dolby Headphone mode on the Xonar DGX and DX had similar surround content, but it sounded a little more distant and muffled. Remember that Dolby Headphone attempts to simulate sound coming from speakers at a distance. Personally, I prefer the in-game Enhanced mode.
In DiRT Showdown, there was a pronounced difference between the built-in software and Rapture3D audio modes. The latter felt natural, with in-game sounds at more appropriate distances than the default audio, which kind of crammed everything right into my ear. Again, Dolby Headphone output sounded a little muffled and far away—and not as good as the Rapture3D surround mode.
|Rockchip SoC powers $149 Chromebooks, sub-$100 dongle||3|
|The TR Podcast 173: Torquing the Titan||4|
|A fresh look at storage performance with PCIe SSDs||31|
|Leaked specs detail Intel's 14-nm Braswell SoCs||33|
|Here are our musings on the new MacBook||151|
|Microsoft unveils Atom-powered Surface 3 tablet||78|
|Source code references hint at Tegra X1 Chromebooks||2|
|Samsung's 850 EVO M.2 solid-state drive reviewed||32|
|New Windows 10 build includes Project Spartan browser||67|
|THIS IS THE INTERNET. THERE IS NO PLACE FOR FUN DISCUSSION.||+36|