OUT OF its Singapore base, Creative Technology has firmly ruled the PC sound landscape since the very first Sound Blaster card over 20 years ago. From more polyphonic sounds and realistic effects, to better frequency response, dynamic range, lower noise, and finally on-board DSP and effects accelerators to offload the CPU from processing high-bit rate multichannel surround audio, Creative led all the way - up to the current X-Fi product line.
Its early local Singapore sound competitors, like Aztech and AVS, gave up by the middle of last decade, as well as well known US players as Turtle Beach, with only specialised companies like Razer offering now some entries better than the generic, CPU-loading motherboard audio.
The biggie of the mainboard business, Asus, came into the fray this month, with its Xonar line of cards. Aimed at the very same high end of audio as Creative's X-Fi, Xonar seems to be the first really serious competitor to Creative's bread & butter, from a very large resourceful brand at that, not some niche player. So, the comparison was inevitable.
Firstly, when you compare the specs of two cards - which we won't repeat here since the brochures are on the Web, you'll see that, actually, they don't exactly seem to shoot for the very same market segment. For instance, Asus specs claim slightly better sound quality on every parameter from sampling rate to frequency, dynamic and noise, while Creative seems to have a definite edge on the on-board processing strength around its DSP & effects core.
Then, Creative pushes its own EAX - now at 5.0, supported in more PC games now - and OpenAL standards further, relying on its market presence to expand their coverage, while Asus opted for Dolby and DTS standards, more prevalent in movie and console gaming environments. Looking at this alone, one would conclude that Xonar will be more of a solution for the ultimate movie and music listening buffs, especially the new breed of BD and HD-DVD ones, while the gaming performance "enthusiasts" would opt for Creative due to, among others, higher FPS and overall performance.
Even the looks hint at that: with its shiny black noise shield, Xonar looks more like a part of a hi-fi combo, while X-Fi looks and feels like a typical high-end PCI card of some kind.
Do things really work out that way? Let's see.
I set up a common platform, an Intel QX6850-based Asus Striker Extreme Nforce 680i high-end board running at a its defauly 3 GHz CPU / FSB1333 speed water-cooled with Corsair Nautilus 500, with 8 GB of Geil Black Dragon DDR2-800 RAM, running here at DDR2-667 4-4-4-8 setting. Nothing special speed-wise as all three components can do way better, but still a good, stable platform. The GPU was an Asus water cooled GeForce EN8800GTX Aquatank using Thermaltake solution. As for the audio output, Creative 5+1 speaker system was on hand when needed, although Creative Aurvana high-end headphones were sufficient in most cases.
Both cards use PCI 33 bus, and both vendors are expected to come out with PCI-E x1 updated versions within a month or so. Asus has a notch up on user friendliness when it comes to messing up with cables, as its ports are lit up in colours corresponding to the input type, so even in the dark, you'll insert it right. Both are also about the same size, with Asus being some three millimetres longer. You can see them next to each other on the photo here.
Since Vista isn't still exactly smooth when it comes to hardware accelerated audion, I stuck with Windows XP SP2 32-bit for this round of tests. Of course, I do look forward to Vista tests when the stuff has settled (that is, after SP1), and then, well by that time I expect both of these cards to have decent Linux support as well.
For some of the audio quality parameters, I ran the input / output audio loopback on both cards with Rightmark Audio Analyzer 5.6 tool - running it on both 16-bit 44 kHz (standard CD audio) and 24-bit 96 kHz SACD / DVD Audio samplig rates. The typical parameters like frequency response, harmonic distortion, dynamic range and signal to noise ratio were compared.
For CPU utilisation, RightMark 3D sound utility with its CPU utilisation metre was used - again, I ran both 44 and 96 kHz samplers under DirectSound, OpenAL and ASIO options.
WMV HD playback with surround expansion - I played a 1080p HD Alexander scene in Micro$oft Windows Media Player 10 with surround sound, and checked the CPU usage graph on all four cores. Same for high bit-rate 384 kbps MP3 song playback with stereo-to-surround expansion and added effects, including audio real time clean-up and 'crystallisation'.
Finally, I ran the Painkiller game with FRAPS measurement of game frame rate running in background, and Dolby (for Asus) or EAX (for Creative) surround effects. Again, besides any FPS difference, the observable difference in overall CPU utilisation was checked.
So, here is the results table:
X-Fi Elite Pro
|RightMark 3D Sound CPU Mean|
|DirectSound 2D 44 kHz||4.44||3.79|
|DirectSound 3D EAX 44 kHz||4.66||4.92|
|DirectSound 3D EAX 96 kHz||3.47||3.71|
|OpenAL 2D 96 kHz||1.39||-|
|OpenAL 3D EAX 96 kHz||2.41||-|
|RightMark 3D Sound CPU peak|
|DirectSound 2D 44 kHz||5.21||7.81|
|DirectSound 3D EAX 44 kHz||5.13||8.06|
|DirectSound 3D EAX 96 kHz||6.37||8.29|
|OpenAL 2D 96 kHz||4.52||-|
|OpenAL 3D EAX 96 kHz||4.52||-|
|MP10 WMVHD play, peak||15.00%||18.00%|
|FRAPS Painkiller XGA fps idle||281||263|
As you can see, the expectations were confirmed - but each party achieves just a hairline win in its own department - by a thin margin. Asus' better amp circuitry and funky noise shield did help in audio quality, while Creative's much more powerful combined DSP & effects engine made a difference in CPU utilisation. However, that difference has only real effect in gaming, as every FPS increase matters there. If you are just playing a HD-DVD or listening to hi-fi surround music, another percent or two of CPU usage won't matter, as, most probably, you're only doing that on that machine at the time.
In my mind, the RM 3D sound benchmarks were the most interesting, where you could see the effects of using Creative's on-board 64 MB X-RAM in the OpenAL mode. The CPU utilisation drastically dropped, especially in the top DirectSound3D + EAX mode. Keeping the traffic off the PCI bus does help.
In summary, the two head-on competitors seem more like two complementing products - for now. Xonar does a (slightly) better "hi-fi" audio job, while X-Fi is, again by a slight margin, the ultimate gamer's audio card. After this, Creative has no alternative but to up the ante on the audio quality with better circuitry on its upcoming PCI-Express X-Fi cards using EMU20K2 chip - and maybe add a noise shield too, while Asus will have to go for more DSP horsepower on the next major revision of Xonars.
At that time, we'll talk really level, head-on competition - and the outcome of that race might not be decided by the card sales, but by the decision of the mainboard makers which solution to put as standard on their high end mobo offerings. After all, just like the time is overdue for more "intelligent" I/O like RAID, Ethernet and wireless with protocol and processing offloaded from CPU to the chipset, same applies to the audio: Accelerated, hardware audio with combined quality and performance is the goal. ?
Fantastic design, top-notch hi-fi quality, lit-up connectors
Needs more DSP power
Nothing really, except the price
More computing power than yesteryear's laptops, "the" PC audio standard
Could use better amp componentry for a high end card
Same, price could always be lower. Some consolation: The X-Fi Platinum Fatal1ty will give you the same features for Xonar-type price without the console
Sign up for INQbot – a weekly roundup of the best from the INQ