Creative’s Sound Blaster X-Fi audio processor

Manufacturer Creative
Model X-Fi XtremeMusic
Price (Street)
Availability Now

FOR A WHILE THERE, the PC audio market was pretty boring. Creative had bought out Aureal and Sensaura, and seemed content to issue only incremental improvements to its Audigy line of sound cards. VIA challenged Creative’s dominance of the market with various respins of the Envy24, and although they were generally well-received, those chips’ lack of hardware acceleration for 3D audio ultimately limited their appeal. Then, in May of this year, Creative surprised many by announcing an all-new Xtreme Fidelity audio processor loaded with 10,000 MIPS of processing power—24 times that of the Audigy chip it would replace. A radical departure from architectures of old, the X-Fi arranged its on-chip components around a pipelined audio ring with a whopping 4096 internal audio channels. The X-Fi also upgraded EAX to support up to 128 simultaneous hardware-accelerated 3D voices, and promised to enhance compressed audio playback to sound better than the original CD.

Modesty has never been one of Creative’s strong suits, but could there be something to all the X-Fi hype? To find out, we subjected the most affordable X-Fi offering, the $110 Sound Blaster X-Fi XtremeMusic, to a punishing battery of performance, quality, and listening tests against competitors ranging from M-Audio’s Revolution 7.1 to integrated “Azalia” High Definition Audio. Read on to see how the XtremeMusic fared and why the X-Fi is such a bold departure from audio chips of old.

The X-Fi architecture
Before exploring where the XtremeMusic fits into the Sound Blaster Xtreme Fidelity family, let’s take a moment to explore the 51-million transistor X-Fi audio chip that’s at the heart of Creative’s new Sound Blasters. Manufactured using 0.13-micron process technology, the chip has roughly half the number of transistors of an Athlon 64 and more than 11 times that of the Audigy, so it’s quite a leap from previous generations.

The X-Fi’s processing power is divided between five internal units: the sample rate converter, digital signal processor, and mixer, filter, and tank engines. Much of the X-Fi’s muscle ripples through a sample rate converter (SRC) that Creative claims pushes over 7000 MIPS. The SRC is actually made up of 256 individual sample rate converters, all of which tackle sampling rate conversions in the same manner. First, the sampling rate of an incoming audio stream is doubled. Next, a poly phase Finite Impulse Response (FIR) filter is used to produce a sampling rate four times greater than the desired output sampling rate. Finally, the sampling rate is reduced by a factor of four for output. According to Creative, this process is nearly transparent, and any loss in quality during sample rate conversions is miniscule compared to the noise generated by even the best DACs available on the market. If you’re not convinced, the SRC can be bypassed when it’s not needed.

Although the X-Fi’s sample rate converter has significantly more processing power than the rest of the chip, it’s still only one of five main chip components. The next X-Fi component of interest is the Quartet DSP. Quartet, in this case, refers to the fact that the X-Fi’s digital signal processor is made up of four SIMD (Single Instruction, Multiple Data) hardware threads. Each of those hardware threads has two data paths, leading Creative to describe it as TIMD, or Thread Interleaved Multiple Data. Giving each DSP thread dual data paths is a clever way to deal with audio data, which generally arrives in multiples of two. With four “stereo” hardware threads, the DSP can tackle eight data streams at once—perfect for an eight-channel sound card. As one might expect, the Quartet DSP’s instruction set is audio-centric. It can handle both fixed and floating-point data types, and Creative claims that the interface is programmer friendly.

The Xi-Fi’s SRC and Quartet DSP are undoubtedly the stars of the show, but the chip has several other essential components worth mentioning. More than 1200 MIPS of processing power are dedicated to the X-Fi’s mixer, which handles the scaling, combining, and, of course, mixing of audio streams. The “Tank” engine handles all of the X-Fi’s delay-based effects, including reverb, chorus, reflections, and inter-aural time delays, while a filtering engine dedicates a couple hundred MIPS to environment modeling, equalizers, and positional 3D audio. The X-Fi also has a transport engine that interfaces with onboard memory and an I/O bus, such as PCI. As you might expect, the chip also has an audio I/O component.


The X-Fi audio ring (Source: Creative)

A typical audio card would organize the X-Fi’s various components in a line, each leading to the next with little flexibility in the signal path. Creative does things differently with the X-Fi, organizing the various chip components around what the company calls an audio ring. The audio ring is actually a pipelined bus with 4096 audio channels. Those audio channels can be used for not only the front, rear, center, and other external channels, but also as internal pathways for effects generation. Each chip component is free to access and route audio streams around the ring, giving the X-Fi a great deal of flexibility. That flexibility, combined with the all those internal audio channels, is perhaps the most exciting upgrade over Creative’s previous audio cards. The Audigy, for example, was limited to 64 internal audio channels, all of which were limited to the effects engine.

Speaking of the Audigy2’s effects engine, what would a Creative audio chip be without some measure of EAX support? The X-Fi is equipped to handle Creative’s latest, EAX Advanced HD 5.0, with hardware acceleration for a total of 128 voices—double that of the Audigy2. Battlefield 2 can already take advantage of the X-Fi’s support for extra voices, and most games with 3D audio will take advantage of the X-Fi’s hardware EAX acceleration. Oddly, though, Creative is pretty much still the only game in town if you want hardware-accelerated 3D audio. Others seem content with the performance hit associated with having the CPU do all the heavy lifting for positional 3D audio.

While the X-Fi architecture is impressive and its ability to accelerate 3D audio unmatched, one of the chip’s more interesting features is that it can be paired with onboard memory that Creative calls X-RAM Onboard memory is nothing new in the sound card world (the AWE32 was available with 2, 4, and 8MB of RAM), but X-RAM is different because it’s there specifically to be used by applications. Creative hopes that game developers will utilize X-RAM to cache decompressed audio, potentially saving precious CPU cycles in the process. Developers are also encouraged to use X-RAM to store higher quality audio assets that would otherwise be infeasible to decompress on the fly or stream from the hard drive. At least one developer is already taking advantage of X-RAM, with DICE’s Battlefield 2 using it to cache in-game audio.

Unfortunately, X-RAM is only available at the top end of the X-Fi lineup on Fatal1ty FPS and Elite Pro cards that cost $300 and $400, respectively. With X-RAM limited to high-end cards, many developers may not see enough of an installed base to justify dedicating development resources to taking advantage of X-RAM. Only time will tell.

Despite its fancy new architecture, onboard memory, support for Dolby Digital ES and DTS-EX output, and THX certification, the X-Fi won’t encode Dolby Digital Live. According to Creative, Dolby Digital Live support would make the X-Fi more expensive due to both licensing costs and the need for additional tank RAM on the chip. Creative also asserts that Dolby Digital Live encoding introduces a small amount of latency, and points out that it’s limited to 5.1-channel output. Perhaps more notably, Creative says that Dolby Digital Live encoding can’t handle DRM-protected content since it passes unprotected data through a digital output. That’s an interesting limitation, although PC enthusiasts have never really been enthusiastic about DRM-encrusted content.

 

The Sound Blaster X-Fi XtremeMusic
Creative is offering four Sound Blaster cards based on its new X-Fi technology: the Elite Pro, Fatal1ty FPS, Platinum, and XtremeMusic. Today we’ll be focusing our attention on the XtremeMusic, whose $110 street price puts it at the low end of the X-Fi family. The low end of the X-Fi line isn’t necessarily a bad place to be, though. Thanks to the wonders of trickle down, each member of the X-Fi family uses the same audio chip. In fact, with the exception of X-RAM, all X-Fi cards have the same core features and functionality.

X-Fi cards also share a common aesthetic, which is dominated by black and gold. It could be worse, I suppose, but I’ve never been a bit fan of bling on black—not that it matters much when the card is tucked away inside a case.

Fortunately, an audio card doesn’t have to look good to sound good, and when it comes to sounding good, the XtremeMusic’s resume is impressive. According to Creative, the card has a 109dB signal-to-noise ratio for all DAC channels and supports audio formats up to 24 bits and 192kHz. 24-bit/192kHz audio isn’t supported across the board, though. Creative is quite upfront about the fact that the XtremeMusic only handles 192kHz sampling rates for two-channel stereo playback. Recording and multi-channel playback are limited to 24 bits and 96kHz.

While a 96kHz sampling rate may be enough for most users, we were still curious about the nature of the XtremeMusic’s limitations. On the recording side, one could blame the Wolfson WM8775 ADC, which only supports audio up to 24 bits and 96kHz. Creative says that the X-Fi can technically handle 192kHz recording, but suggests that anyone that serious about recording will probably want to consider its E-mu line of professional products over an X-Fi. When it comes to playback, the card’s eight-channel Cirrus Logic CS4832 DAC claims to support 24-bit audio at up to 192kHz, so it shouldn’t limit multi-channel output. According to Creative, the X-Fi doesn’t support multi-channel 24-bit/192kHz output because no commercial content exists in that format. Most DVD-Audio content is multi-channel 24-bit/96kHz, and Creative only supports stereo output at 192kHz because some DVD-Audio discs offer that option.

Apart from the Cirrus Logic and Wolfson chips, users will notice that the XtremeMusic also has a 2MB Samsung memory chip onboard. However, the memory isn’t available to developers, so it’s not the same as X-RAM. For that, you’ll have to pony up $300 for an X-Fi Fatal1ty FPS or $400 for an X-Fi Elite Pro, both of which come with 64MB of X-RAM.

While the XtremeMusic is one of two X-Fi cards without X-RAM, it’s the only card in the X-Fi lineup that lacks a breakout box with extra input and output ports. That leaves the XtremeMusic with a rather sparse port cluster that relies far too heavily on port sharing. The card’s digital output, for example, shares the same physical port as the line input and microphone. Digital output also requires a 3.5mm TOS-Link adapter, which Creative apparently doesn’t include with the card.

As if not being able to run a microphone and digital speakers at the same time weren’t inconvenient enough, the XtremeMusic only has three analog output ports. That’s fine for six-channel output, but with eight-channel audio, things get a little messy. To compensate for its lack of a dedicated surround output, the XtremeMusic tacks the left and right surround channels onto the rear and center/sub outputs. This arrangement requires special cables that aren’t bundled with the card. Obviously, we’d prefer a dedicated surround output. While we’re at it, standard TOS-Link or coaxial S/PDIF ports would be nice, as well. To be fair, Creative does offer a more extensive array of outputs via a breakout box on its X-Fi Platinum, Fatal1ty FPS, and Elite Pro, but that’ll cost you extra.

Speaking of extras, or a lack thereof, the XtremeMusic lacks the Firewire port present on the Audigy and Audigy2. Firewire isn’t a necessity for a sound card, especially with virtually every enthusiast-oriented motherboard sporting at least one 1394 port, but it was a nice perk.

By now, you’ve no doubt noticed that the XtremeMusic is a PCI card. The entire X-Fi lineup is currently PCI-only, but Creative says that the X-Fi chip can work over USB, Firewire, and most notably, PCI Express. Can we get a PCI Express version ASAP, please?

 

Upgrading for Xtreme Fidelity
Although the X-Fi supports 24-bit/96kHz multi-channel output, most audio is only 16-bit/44.1kHz in stereo. The X-Fi has no problem faithfully reproducing 16-bit/44.1kHz stereo audio, and it can even do so without any sample rate conversions by bypassing the SRC, but that leaves much of the X-Fi’s potential untapped. Instead of being content with simple reproduction, Creative has also endowed the X-Fi with the ability to upgrade a 16-bit/44.1kHz stereo audio stream to multi-channel, 24-bit/96kHz sound. The philosophy behind this upgrade to “Xtreme Fidelity” is that a sound card should strive to create the best listening experience possible rather than simply defaulting to bit-accurate playback. This philosophy works in reverse, too, optimizing playback of multi-channel content through headphones or two-speaker setups.

Creative’s Xtreme Fidelity initiative has two components: 24-bit Crystalizer and CMSS-3D. The first tackles upsampling, while second expands and contracts audio streams over multiple channels to optimize playback for a listener’s speaker configuration.

24-bit Crystalizer
With claims that it “enhances MP3s and movies to sound better than they do on their original CD or DVD,” the 24-bit Crystalizer is easily the X-Fi’s most-hyped feature. Creative certainly has a history of hyperbole, but the audio world in general seems rife with wild claims. 128kbps MP3s that offer CD-quality audio are one of my personal favorites.

If you don’t buy into the hype, the 24-bit Crystalizer is easy to enable and disable within the X-Fi driver control panel. When enabled, audio streams are run through the X-Fi’s sample rate converter and upsampled to 24-bit and 96kHz. From there, the Crystalizer attempts to simulate how a sound engineer would go about remastering the audio stream. It’s common for studio engineers to compress the dynamic range of instruments to fit into a 16-bit/44.1kHz recording, and the Crystalizer tries to restore some of that lost dynamic range.

Creative says that the Crystalizer is able to extract crisper high frequencies, punchier mid-range percussion, and stronger kick bass hits from lower bitrate recordings, a claim we’ll examine in our listening tests. First, though, we thought it would be interesting to examine the Crystalizer’s impact on RightMark Audio Analyzer. We fired up RMAA and ran the X-Fi at 16-bit/44.1kHz with the Crystalizer disabled, at its 50% default, and at 75% and 100%. Only a couple of RMAA’s tests were affected by the Crystalizer, with the most dramatic difference observed in the frequency response test.

Notice how the Crystalizer pumps up high and low frequency sounds, but has poorer frequency response in the middle of the spectrum. In general, the Crystalizer’s impact is consistent across all three percentages, but the same can’t be said for its influence on intermodulation distortion.

Intermodulation distortion occurs when a sound card can’t accurately reproduce two sounds at the same time. With Crystalizer percentages above 50%, the X-Fi definitely struggles. Perhaps that’s why 50% is the default.

Although it’s hard to imagine that the 24-bit Crystalizer can divine enough information from an MP3 to make it sound better than the original CD, the technology may have merit, particularly with low bitrate recordings. Our listening tests will shed some light on just how useful the Crystalizer is in the real world.

CMSS-3D
While the 24-bit Crystalizer upsamples low bitrate recordings in an attempt to recover detail lost in the compression process, CMSS-3D expands and contracts stereo and multi-channel recordings to best match a user’s speaker configuration. CMSS-3D aims to make playback more immersive while still preserving the intent and quality of the original recording. This technology comes in 3DHeadphone, 3DVirtual, and 3DSurround flavors, each of which is optimized for a different speaker configuration.

As its name implies, CMSS-3DHeadphone optimizes playback for headphone listening. Headphones generally give listeners the impression that sound is arcing through their heads, but 3DHeadphone creates the illusion that playback is occurring through a pair of virtual stereo speakers. Creative says this effect can lessen the listening fatigue that some experience with headphones, although some listeners may prefer that the voices stay inside their head.

In addition to creating the illusion of stereo speakers, CMSS-3DHeadphone can also virtualize multi-channel speaker configurations for 3D audio playback. Various techniques are used to trick the ear into thinking that sounds are coming from imaginary speakers that surround the listener. It’s a neat effect, and although I’m in no rush to trade in my six-channel speaker setup for a set of headphones, CMSS-3DHeadphone worked better than I expected.

CMSS-3DVirtual is actually quite similar to 3DHeadphone. Both can virtualize multi-channel audio for stereo playback devices, but instead of concentrating on headphone playback, 3DVirtual is designed for stereo speaker configurations. The same battery of tricks is used to fool the listener’s ears, and since users have more freedom to place stereo speakers than they do headphones, 3DVirtual can be calibrated to take into account the distance between a user and his speakers.

CMSS-3DHeadphone and 3DVirtual both translate multi-channel recordings for stereo output, but CMSS-3DSurround works in the opposite direction, upmixing stereo content for multi-channel playback. 3DSurround doesn’t arbitrarily assign the voices and instruments from a stereo recording to surround speakers, though. Instead, it taps the surround channels to deepen listener immersion and uses the center channel to expand the sweet spot between the front left and right speakers. To achieve immersion without disrupting a stereo recording’s original intent, CMSS-3DSurround extracts ambient sounds from the original recording and routes them to surround channels. Then, it expands the front channel sweet spot by directing center-panned sounds to the center channel. This method keeps the primary image in front of the listener while surrounding him with the ambient noise of the recording. As far as preserving the intent of the original recording goes, that’s a pretty good compromise.

 

Modes
The X-Fi can be configured to run in gaming, entertainment, and audio creation modes. These modes cover the three most common usage scenarios for a sound card, and switching between them is as easy as clicking the driver control panel. Switching modes resets the X-Fi and retargets its computational resources at a short list of features most commonly associated with the mode in question, all without the need for a reboot. Here’s a quick rundown of the features available in each mode:

  Gaming Entertainment Audio creation
Video game frame rate Yes No No
Hardware 3D audio processing Yes No Optional
Environmental effects Yes Optional Optional
High resolution audio playback No Yes Yes
Audio enhancement processing No Yes Optional
Two-channel to multi-channel upmix Optional Yes No
Multi-channel audio recording No No Yes
Hardware MIDI playback No Optional Yes
Hardware effects Yes Yes Optional
Sample-synchronized record and playback No No Yes
Low audio streaming latency No No Yes
Bit-accurate audio capable No Optional Yes

Creative refers to the X-Fi’s mode-switching capacity as an Active Modal Architecture (AMA). The feature choices for each mode make sense, too, especially since switching between modes takes only seconds.

The X-Fi’s gaming mode is optimized for—you guessed it—gaming. In this mode, the X-Fi is capable of starting hardware voices ten times faster than the Audigy2. Users get hardware acceleration for up to 128 DirectSound and EAX Advanced HD 5.0 voices, and OpenAL is also supported.

Entertainment mode is geared towards movie and music playback, and it’s all about upsampling and upmixing content using the 24-bit Crystalizer and CMSS-3D. This mode also supports DVD-Audio playback, and gives users compatibility with Dolby Digital ES and DTS-EX.

Designed for would-be producers and musicians, the X-Fi’s audio creation mode may be the least popular of the three. In this mode, the X-Fi supports ASIO at 24 bits and 44.1, 48, 88.2, and 96kHz with no CPU hit. According to Creative, ASIO 2.0 latencies are as low as one millisecond.

Audio creation mode also supports 3D spatialization for MIDI, allowing for multi-channel MIDI creation. The mode allows up to eight hardware-accelerated effects to be applied to any audio stream moving through the card, and 24-bit sound fonts are also supported. Users are even given some control over the X-Fi’s SRC; the X-Fi’s internal sampling rate can be switched between 44.1, 48, 88.2, and 96kHz, and users can enable a bit-accurate playback mode that bypasses the SRC, the 24-bit Crystalizer, the equalizers, and the card’s smart volume management features.

Software extras
Creative’s audio drivers are notorious for feature bloat, but the X-Fi’s drivers are rather well behaved in that respect. Creative offers a performance install option that only gives you the essentials, including the base drivers and the audio control panel. Most users probably won’t need anything beyond this performance install, but there are a few extras you might want to consider adding to the mix.

The first extra you might want to consider installing is Creative’s MediaSource player, not for its playback capabilities, but for Super Rip. Super Rip leverages the 24-bit Crystalizer and CMSS-3D to rip CDs to 24-bit/96kHz multi-channel WMA Pro and WMA Lossless formats, and is a handy feature if you’re a fan of the X-Fi’s upsampling and upmixing capabilities. MediaSource can also convert existing recordings, including MP3s, to high-bitrate, multi-channel WMA formats.

Curiously, the X-Fi’s DVD-Audio playback capabilities aren’t tied to the MediaSource player. That’s not necessarily a bad thing, though, since you can install the DVD-A playback app without having to install MediaSource. The X-Fi’s updated DVD-Audio player has a slicker interface than the player Creative includes with the Audigy2, but the new version doesn’t appear to offer any significant new features over its predecessor.

Another of the X-Fi’s enticing software extras is the THX control panel, which, among other things, lets users calibrate playback to take into account the distance, level, and angle of each attached speaker.

We’ve only touched on a handful of extras here, but there are plenty of other extras on the X-Fi install CD. At the very least, every X-Fi user should install the DVD-Audio player and experience the aural glory of a 24-bit/96kHz multi-channel DVD-A recording.

 

Our testing methods
We’ll be comparing the X-Fi XtremeMusic’s performance to that of M-Audio’s Revolution 7.1, our Sapphire motherboard’s integrated Realtek ALC880 “Azalia” audio, and Creative’s own Audigy2 ZS. As the X-Fi’s predecessor, the Audigy2 ZS’s inclusion is easy to justify. We also chose to include M-Audio’s Revolution 7.1 because, in our opinion, it sounds better than any other consumer-level Envy24 implementation on the market. Our Sapphire motherboard’s integrated Realtek HD audio isn’t a direct competitor to the X-Fi, but we’ve included it to show what one can expect from onboard audio.

All tests were run three times, and their results were averaged, using the following test systems.

Processor AMD Athlon 64 FX-53 2.4GHz
System bus HyperTransport 16-bit/1GHz
Motherboard Sapphire PI-A9RX480
BIOS revision Beta 07/28/2005
North bridge ATI RX480
South bridge ATI SB450
Chipset drivers Catalyst 5.7
Memory size 1GB (2 DIMMs)
Memory type OCZ PC3200 EL Platinum Rev 2 DDR SDRAM at 400MHz
CAS latency (CL) 2
RAS to CAS delay (tRCD) 2
RAS precharge (tRP) 2
Cycle time (tRAS) 5
Hard drives Western Digital Raptor WD360GD 37GB SATA
Audio SB450/ALC880 M-Audio Revolution 7.1 Creative Audigy2 ZS Creative X-Fi XtremeMusic
Audio driver Realtek HD 1.24 M-Audio 1024 Creative 1.84.55 Creative 5.12.5.1141
Graphics NVIDIA GeForce 6800 GT with ForceWare 77.77 drivers
OS Microsoft Windows XP Professional
OS updates Service Pack 2, DirectX 9.0c

Our test system was powered by OCZ PowerStream power supply units. The PowerStream was one of our Editor’s Choice winners in our latest PSU round-up.

We used the following versions of our test applications:

The test systems’ Windows desktop was set at 1280×1024 in 32-bit color at an 85Hz screen refresh rate. Vertical refresh sync (vsync) was disabled for all tests. Most of the 3D gaming tests used the Medium detail image quality settings, with the exception that the resolution was set to 640×480 in 32-bit color.

All the tests and methods we employed are publicly available and reproducible. If you have questions about our methods, hit our forums to talk with us about them.

 

Gaming performance
To test the X-Fi’s impact on in-game frame rates, we rounded up a handful of games and 3DMark03’s handy audio benchmark.

3DMark03

In 3DMark03’s 3D audio benchmark, the X-Fi has a modest lead over the Audigy2, but one that gets more pronounced as the number of sounds increases. The ALC880 and Revolution 7.1 don’t give the X-Fi much of a run for its money, either; the former doesn’t support 3DMark03’s 60-sound setting, while the latter is almost ten frames per second off the pace.

DOOM 3
DOOM 3 didn’t initially support hardware acceleration for 3D audio; however, the latest 1.3 patch adds EAX 4.0 HD support for the Audigy2 and X-Fi. We tested all the cards with the game’s standard surround sound setting, and we also tested the Creative cards in EAX 4.0 mode.

Not much to see here. The X-Fi leads the pack, but it’s only faster than the other cards by a hair.

Half-Life 2

Although barely faster than the Audigy2 in Half-Life 2, the X-Fi has a clear lead over the M-Audio Revolution 7.1, and to a lesser extent, the ALC880.

Unreal Tournament 2004
We ran through our Unreal Tournament 2004 timedemo in each of the game’s three 3D audio modes.

The X-Fi doesn’t have much opportunity to stretch its legs in the first 3D mode, which doesn’t take advantage of hardware acceleration. However, in the game’s two hardware-accelerated modes, the Creative cards run away from the competition. Note that, again, there’s not much difference in performance between the Audigy2 and X-Fi.

Battlefield 2
We used FRAPS to record in-game frame rates for a two-minute run through Battlefield 2. This method of benchmarking isn’t quite as reliable as in-game timedemos, but our results were reasonably consistent over five test runs.

We tested each of the cards with Battlefield 2’s high quality 3D audio setting. Since the game also has an ultra quality mode for the X-Fi, we tested that, as well.

It doesn’t quite catch the Audigy2, but the X-Fi performs extremely well in Battlefield 2, in particularly because there’s no performance hit associated with the game’s “ultra” quality mode. The additional sounds provided in the ultra mode were easily noticeable, at least to my ears.

 

CPU utilization
RightMark 3D Sound measures CPU utilization with a variable number of hardware voices for 2D, 3D, and EAX audio. The ALC880 doesn’t support more than 32 voices, so you won’t find scores for it at 64 and 127. There are scores for the Audigy2 ZS and Revolution 7.1 at 64 voices, but we should note that the Audigy2 actually maxes out at 63 voices.

The advantages of hardware acceleration are readily apparent in RightMark 3D Sound, with the Creative cards easily consuming fewer CPU resources than the competition. Note that the X-Fi uses less CPU power with 127 voices than the M-Audio Revolution 7.1 does with just 64.

While running RightMark 3D Sound, we wondered if the X-Fi’s 24-bit Crystalizer and CMSS-3D had much of an impact on CPU utilization, so we put the XtremeMusic through a few extra tests with various Crystalizer and CMSS-3D settings.

With the exception of CMSS-3DHeadphone, neither the 24-bit Crystalizer nor CMSS-3D appears to have much of an impact on CPU utilization. CMSS-3DHeadphone’s CPU utilization doesn’t really ramp up until we hit 127 voices, either.

 

Listening tests – CD-quality audio
For our listening tests, I conscripted a couple of friends, who have asked to be referred to here as Frylock and Master Shake, and subjected them to several hours of music playback spanning all four cards. Tests were conducted with a pair of 5.1-channel Logitech Z680 speakers connected to each sound card using the same analog cables. Volume levels were normalized to within one decibel, and all environmental effects and equalizers were disabled, including the X-Fi’s CMSS-3D. Our listeners were also unaware of which sound card was being used for playback at any given time, and the order of playback was randomized throughout.

This first set of listening tests examines CD-quality audio playback. In order to preserve quality for playback, tracks were ripped directly to WAV with no compression. The X-Fi’s 24-bit Crystalizer was disabled for this round of listening tests; we’ll get to it in a moment.

Below you’ll find a summary of our thoughts on how the X-Fi’s playback compared with the Revolution 7.1, Audigy2 ZS, and ALC880.

Santa Esmerelda – Don’t Let Me Be Misunderstood
Ripped from the Kill Bill Volume 1 soundtrack, this song has a decidedly latin feel, with a clapping back track and delicately plucked acoustic guitar.

The X-Fi and Revolution 7.1 were virtually indistinguishable with this track, even after several back-to-back tests. Against the Audigy2 ZS, the X-Fi did a better job reproducing the subtle pluck of each string note, although the two were otherwise very similar. The X-Fi also sounded much better than ALC880, whose bass was more muffled and strings less twangy. Integrated audio didn’t even come close with this track.

Antonio Vivaldi – The Four Seasons Winter 1
A classical, er, classic, The Four Seasons Winter 1 is dominated by soaring strings that are almost piercing in their ferocity.

The X-Fi outshines the Audigy2 on this track, providing sharper violins and richer background. Our listeners were split on whether the X-Fi sounded better or worse than the Revolution, though; the X-Fi did a better job with the foreground violins, but the Revolution had a little extra background depth. Consensus was easy to find with the ALC880, which all parties agreed sounded much more distant than the X-Fi. The ALC880 also lacked the subtle violin squeak heard with the other sound cards.

Chemical Brothers – The Test
Richard Ashcroft provides vocals for this addictive electronic track, which features a rolling bass line, plenty of otherworldly sounds, and what can only be described as a sonic liftoff.

Unfortunately, the Audigy2’s liftoff wasn’t quite as vibrant as that of the XtremeMusic. The two Creative cards were pretty close, though. M-Audio’s Revolution 7.1 wasn’t so close, and came off sounding a little dull next to the X-Fi, with less separation at the low end of the spectrum. The Revolution didn’t fare nearly as poorly as the ALC880, which also sounded dull—across the board, in fact, when compared with the X-Fi.

Tori Amos – Caught a Lite Sneeze
This track laces Tori Amos’ sultry vocals over piano, percussion, and harpsichord.

Our listeners were split on whether the X-Fi or Revolution 7.1 sounded better on this track; the X-Fi beat the Revo when it came to vocal depth, but the song’s piano was a little stronger on the M-Audio card. The X-Fi also outshined the Audigy2’s vocal depth, stretching Amos’ ragged voice further than its predecessor. By comparison, the ALC880’s vocal reproduction was weak, sapping the emotion and urgency from the track. The ALC880 also couldn’t match the punch to the chest that the listeners felt when listening to the X-Fi.

The Killers – Mr. Brightside
Most of today’s popular “The” bands profess to be punk, or at least what passes for punk these days, but The Killers’ punchy, distortion-laced Mr. Brightside flirts with what kids these days call screamo.

Mr. Brightside played back similarly on the XtremeMusic and Revolution 7.1, but the track’s voice echoes and cymbals were a little clearer on the X-Fi. The Audigy2 was close as well, although it tended to favor the vocals too much at the expense of bass and background instrumentation. Unfortunately, the ALC880 seemed to get bogged down in the track’s distortion, producing much more muted vocals than the X-Fi’s vivid performance.

Rob Dougan – Furious Angels
The title track of a double album that anyone who loved the music in The Matrix should own, Furious Angels cuts a soulful female wail through Rob Dougan’s raspy vocals.

The Audigy2 has always done well with vocals, but on this track, the X-Fi did better, letting listeners hear Dougan’s lips move before each word. The Revolution’s vocal performance wasn’t as detailed as the X-Fi’s, either, with the M-Audio card’s bass slightly muffling the track’s piercing wail. Surprisingly, the ALC880 didn’t fare as badly against the X-Fi here as it did with other tracks. Sure, it didn’t sound as good as the Creative card, but the Realtek managed a pretty balanced performance with our last CD audio track.

 

Listening tests – DVD-Audio
To test the XtremeMusic’s 24-bit/96kHz multi-channel playback, I selected a track from each of my three DVD-Audio discs. These tracks were played for our listeners in the same manner as our CD audio listening tests, but we were only able to include the X-Fi and Audigy2. Unfortunately, the Revolution 7.1, like all Envy24-based audio solutions, doesn’t support playback of DVD-Audio content. We couldn’t find an application that would play DVD-Audio tracks on the ALC880, either.

Nine Inch Nails – Only
Nine Inch Nails’ latest studio album is available in a dual-disc format, with a CD on one side and 24-bit/48kHz DVD-Audio on the other. Only is one of the most distinct songs on the album, as Trent Reznor channels disco, of all things.

Our listeners couldn’t distinguish any difference between the X-Fi’s output and the Audigy2’s with this track. It sounded great on both, of course, but neither Creative card sounded better than the other. Perhaps the track’s 48kHz sampling rate prevented the X-Fi from stretching its legs.

Blue Man Group – White Rabbit
Blue Man Group’s The Complex serves up DVD-Audio in all of its 24-bit/96kHz glory, and the group’s haunting cover of Jefferson Airplane’s White Rabbit comes from a very different world than those cheesy Intel commercials.

After several back-to-back tests, our listeners through the X-Fi sounded marginally better than the Audigy2 ZS. The two were so close that our listeners weren’t entire sure which sounded better, and I was hard pressed to tell the difference between the cards, too.

Brahms – Piano Quintet in F Minor
Featured on Creative’s DVD-Audio sampler disc, Brahms’ Piano Quintet in F Minor plays back at 24-bit/96kHz and features strings over, you guessed it, piano.

Our listeners found a minor difference between the two Creative cards with this track. The X-Fi had a slightly richer feel that conveyed more body than the Audigy2, but both listeners concluded that the two cards were very close.

Listening tests – 24-bit Crystalizer
For this set of X-Fi-only tests, we played several tracks in WAV, LAME “extreme” MP3, and 128kbps constant bitrate MP3 formats. Our listeners heard each track, in each format, with the X-Fi’s 24-bit Crystalizer disabled and enabled at its default 50% setting. For the first two tracks, we didn’t change the volume levels between normal and Crystalizer playback.

LCD Soundsystem – Disco Infiltrator
Dubbed indie electronica, LCD Soundsystem’s Disco Infiltrator delivers plenty of drums and just a hint of funk.

The Crystalizer definitely delivered on Disco Infiltrator, with harder snare hits and crisper vocals than normal playback. Listeners didn’t hear any new depth so much as they heard more of what was already there. This impact was more pronounced with our 128kbps MP3, but considerably more muted with our LAME “extreme” MP3 and WAV files.

Incubus – Aqueous Transmission
A fluttering flute and eastern strings play over a subtle bass line in Incubus’ Aqueous Transmission, which wouldn’t sound out of place as the opening track for an epic Kung Fu movie.

Aqueous Transmission‘s flute flutter and string twang were both enhanced by the 24-bit Crystalizer, but again, the effect faded as we moved from low-bitrate MP3s to uncompressed WAV files. Interestingly, our listeners didn’t think that Crystalized 128kbps MP3s sounded better than normal WAV playback.

At this point, I began to suspect that our listeners were favoring the 24-bit Crystalizer’s playback because it sounded louder than normal playback. Our trusty sound level meter confirmed that the playback was peaking three to four decibels higher with the Crystalizer than normal playback, despite the fact that the control panel and speaker volume levels hadn’t been touched. To even the playing field, I nudged up the volume of normal playback to match the Crystalizer’s louder output for the next two tracks. This didn’t match volume levels perfectly, since the Crystalizer only seems to turn up the volume on the foreground instruments that it touches, but it’s as close as we could get.

Johnny Cash – The Man Comes Around
Although the man has come around for Johnny Cash, his timeless voice lives on here, backed up as always by an acoustic guitar.

Turning up the volume on non-Crystalized playback made enough of a difference that our listeners began favoring normal playback over the 24-bit Crystalizer. Instead of noting the Crystalizer’s sharper foreground sounds, our listeners noticed the clearer background separation of normal playback. The Crystalizer still had great highs and lows, but less in the middle. This disparity was particularly clear with uncompressed WAV playback.

U2 – Two Shots of Happy, One Shot of Sad
Written by Bono for the late Frank Sinatra, Two Shots of Happy, One Shot of Sad conjures an orchestra to back Bono’s unmistakable vocals.

Again, our listeners found background instrumentation stronger without the Crystalizer. Background instruments were certainly louder without the Crystalizer, although the volume and quality of foreground vocals was closely matched.

 

RightMark Audio Analyzer – Loopback – 16-bit/44.1kHz
Moving from subjective listening tests to something more objective, we used RightMark Audio Analyzer to evaluate the X-Fi’s audio quality versus the competition. Our first set of RMAA results are from “loopback” tests that route a sound card’s output through its line input. We’ll kick things off with 16-bit/44.1kHz CD-quality audio.

To keep things simple, I’ve translated RightMark’s word-based quality scale to numbers. Higher scores reflect better audio quality, and the scale tops out at 6, which corresponds to an “Excellent” rating.

The XtremeMusic almost runs the table with 16-bit/44.1kHz CD-quality audio, but the Audigy2 ZS’s noise levels are slightly lower. However, the X-Fi’s frequency response does prove superior to its predecessor.

 

RightMark Audio Analyzer – Loopback – 24-bit/48kHz
My Nine Inch Nails With Teeth dual disc’s DVD-Audio tracks are 24-bit/48kHz, so we ran RMAA at that bitrate and resolution.

The Creative cards come out ahead in each test, but the X-Fi is unable to distance itself enough from the Audigy2 to affect RMAA’s rating scale. Note the large disparity in intermodulation distortion, noise levels, and dynamic range between the Creative offerings and the rest of the pack.

 

RightMark Audio Analyzer – Loopback – 24-bit/96kHz
RMAA won’t run loopback tests on the Revolution 7.1 at 24-bit/96kHz, citing a lack of full duplex support for 24-bit/96kHz playback and recording.

Creative rules again, but this time, the XtremeMusic manages to edge out the Audigy2 in the frequency response test.

 

RightMark Audio Analyzer – DMX 6fire – 16-bit/44.1kHz
To remove each sound card’s line input as a variable, we switched from RMAA loopback to straight playback tests using a Terratec DMX 6fire 24/96 installed on a separate system for recording.

The XtremeMusic leads the field in the frequency response test, but falls a little short when we look total harmonic distortion (unwanted noise introduced to a recording as it’s played back through the sound card) and intermodulation distortion.

 

RightMark Audio Analyzer – DMX 6fire – 24-bit/48kHz

The X-Fi leads four of six tests with 24-bit/48kHz playback, losing to the Audigy2 and Revolution 7.1 in the stereo crosstalk and total harmonic distortion tests, respectively.

 

RightMark Audio Analyzer – DMX 6fire – 24-bit/96kHz

Our 24-bit/96kHz results are similar to those at 24 bits and 48kHz. However, at 96kHz, the Audigy2 scores a little better in the intermodulation distortion test.

 

Detailed RMAA results – Loopback – 16-bit/44.1kHz
If you want to geek out over a bunch of detailed RMAA graphs, we’ve provided the raw results for all of our RMAA tests over the following six pages. The results are included as a bonus; feel free to skip to the conclusion of the review if you wish.

Frequency response

Noise levels

Dynamic range

Total harmonic distortion

Intermodulation distortion

Stereo crosstalk

 

Detailed RMAA results – Loopback – 24-bit/48kHz

Frequency response

Noise levels

Dynamic range

Total harmonic distortion

Intermodulation distortion

Stereo crosstalk

 

Detailed RMAA results – Loopback – 24-bit/96kHz

Frequency response

Noise levels

Dynamic range

Total harmonic distortion

Intermodulation distortion

Stereo crosstalk

 

Detailed RMAA results – DMX 6fire – 16-bit/44.1kHz

Frequency response

Noise levels

Dynamic range

Total harmonic distortion

Intermodulation distortion

Stereo crosstalk

 

Detailed RMAA results – DMX 6fire – 24-bit/48kHz

Frequency response

Noise levels

Dynamic range

Total harmonic distortion

Intermodulation distortion

Stereo crosstalk

 

Detailed RMAA results – DMX 6fire – 24-bit/96kHz

Frequency response

Noise levels

Dynamic range

Total harmonic distortion

Intermodulation distortion

Stereo crosstalk

 

Conclusions
Some have criticized the X-Fi for being too expensive, in part because the Fatal1ty FPS and Elite Pro flavors cost $300 and $400, respectively. However, the X-Fi XtremeMusic can be had for as little as $110 online. At that price, it’s certainly affordable, although still more expensive than a $65 Audigy2 ZS or $80 Revolution 7.1. We think the premium is justified in both cases. The Revolution 7.1, for example, lacks hardware acceleration for 3D audio and DVD-Audio support. It doesn’t sound as good as the XtremeMusic, either. Considering that Revolution’s generally excellent playback quality has made it one of our favorites for quite some time, the fact that the XtremeMusic surpasses it says a lot.

Picking the XtremeMusic over the Revo is an easy call, but handicapping it against Creative’s own Audigy2 ZS is more difficult. The Audigy2 costs 40% less, and you still get hardware acceleration for 3D audio and DVD-Audio playback support. However, the Audigy2’s EAX’s only scales to 64 voices—half that of the X-Fi. The XtremeMusic also offers much more balanced playback than the Audigy2, avoiding its predecessor’s tendency to favor foreground vocals at the expense of background instrumentals. And that’s without the 24-bit Crystalizer and CMSS-3D effects.

Of course, the 24-bit Crystalizer and CMSS-3D aren’t for everyone. Despite Creative’s claims, we didn’t find that the Crystalizer made MP3s sound better than the original CD. Some listeners may favor the sharpness that the Crystalizer lends to the foreground elements of certain recordings, but we prefer higher quality recordings over Crystalized low-bitrate MP3s. We also prefer a real multi-channel speaker setup to virtualized speakers via CMSS-3D, but for those occasions when headphones are a must, CMSS-3D does come in handy. Users should also enjoy CMSS-3D’s ability to expand stereo recordings for multi-channel playback, and we’re quite pleased with Creative’s philosophy of not messing around with the original intent of the recording.


Creative Sound Blaster X-Fi XtremeMusic
October, 2005

As much as we like the XtremeMusic, we’re not crazy about a couple of the card’s features, or more accurately, lack thereof. First, we’d rather Creative made X-RAM standard across the X-Fi line, if only to widen the technology’s install base and encourage developer support. We also wish that the XtremeMusic had individual analog surround output, line input, and digital output ports. Such apparent luxuries may only available on more expensive, breakout box-equipped X-Fi variants.

Despite its faults, the X-Fi XtremeMusic is highly recommended. In fact, the card’s strong performance in our gaming and listening tests, its low CPU utilization, and the X-Fi architecture’s considerable potential warrants an Editor’s Choice award. If I were buying a sound card for gaming and entertainment today, I’d spring for an XtremeMusic in a heartbeat. 

Comments closed

Pin It on Pinterest

Share This

Share this post with your friends!