Just to remark on the X-Fi's frequency boosting in the lower and higher frequencies. Myself and a freind have just purchased the extreme music version, we both used a digital amp and although use different speakers we also both use the Behringer Ultracurve pro deq2496.
Now the ultracurve provides a 32 band equaliser in the digital domain and also allows in room frequency response measurements of systems and the offers correction.

I have to say that although I suspected that the x-fi was acting as a glorified loudness control (haven't measured it yet with the ultracurve) I still insticntively feel that it doing more than just that. When I get round too measuring the x-fi and those low end and high end boosts are removed I still feel that the other attributes that it offers will clearly manifest themselves.

In fact anyone with a 32 band equaliser can mimic the settings shown in the test graph and see if that sounds the same with the crystaliser off!

I've tried those boosts settings before I had the x-fi by using either the ultracurve or the descisoft winamp plugin 32 band equaliser and I never got the resolving power that the x-fi offers

LTJ

No gameport, eh? Suckage. Yeah I still use my MS Sidewinder pad for emulators.

Not that I'd dare buy another Creative card again, ever. The driver problem I've had with the Audigy Line never got fixed, and I won't take that chance again.

But excellent review. TR rox.

I'm also disenchanted by the connections setup of the basic X-Fi. For the relatively high $120, all they needed to do was have a second passive backplate with ports that connected to the appropriate spots on the card. It would have been a cheap and easy solution to implement without changing card production, and might have been justification in the eyes of people comparing the X-Fi and the Audigy2 ZS.

The differences are subtle enough that I couldn't justify the cost, especially as Circuit City is currently selling the Audigy 2 ZS Platinum with the LiveDrive for $89 ($150 - $60MIR). Of course, I can't really justify upgrading my OEM Audigy 2 card for a 2 ZS even, for what I do with it.

I wish we could get some benchmarks of BF2 with the XRam and without to see if you get better performance.

The review was great other than the choice of equipment used for the listening tests. Wouldn't using a speaker set more accurate than the Z-680 be useful? IMHO the logitech seem to have a bloated low frequency response. Something that sounds flatter and more accurate would have been more appropriate to judge the output. (Altec Lansing MX5021 comes to mind)

RIP, Hardware accelerated audio. The CPU hogging M-Audio is still able to pull a very playable framerate. You might as well call the "X-Fi" the Audigy 5. You can only see a performance delta for X-Fi and Audigy 2ZS in a CPU-whoring game like UT2K4.

The XRAM on the Platinum and "Wallet-Fatali1y" edition are practically useless marketing gimmicks. It would make more sense to place it on a true professional-level model, but Creative is lacking in the software department.

Creative never places any of the standard Optical and Codxial outputs on their cards. It's to justify the existence of "Creative Drive" and other external boxs found on their higher-end, more $$$$$$ sound cards. Just another reason to loathe Creative's business tactics.

Thanks for ther review.

Notice how the Crystalizer pumps up high and low frequency sounds, but has poorer frequency response in the middle of the spectrum.

The frequency response isn't "poorer" in the middle of the spectrum. It has increased gain in the lower bass and treble, and reduced gain in the midrange around 1kHz, but that isn't "poorer", it's just reduced gain. If that's all it is doing, then it is just a glorified-but-fixed tone control, and you could do virtually the same thing by lowering the volume a bit and then turning up the bass and treble tone controls on your average simple stereo. What a wank.

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.

Whilst the intermodulation distortion rises a little over -80dB at 100% Wank setting (-80dB = 0.01% distortion), if anyone can here that on the shithouse speakers most people use by necessity on their PCs, I'd be very surprised.

Interestingly, our listeners didn't think that Crystalized 128kbps MP3s sounded better than normal WAV playback.
Bloody hell, I'd hope (expect) not. Once you've thown away data by MP3 compression, you can't retrieve it. The Crystallizer might in principle be able to make an MP3 sound nicer/sharper/whatever than without it turned on, but there is no way it could ever make an MP3 sound better than the original WAV file!

Anyway, as you noted, without exact matching of volumes, these tests are invalid. I'm interested in your comment that the Crystalizer only seems to turn up the volume on the foreground instruments that it touches - how would it "know" which are "foreground" instruments?! It is working on a file format that does not differentiate between "foreground" and "background" instruments.

Re the charts from page 16 onwards; thanks for going to the trouble for including these. However, to put some of those figures in perspective (e.g. for posters like #11 Dirge, who are unsure of what the figures mean), it seems that in absolute terms the cards are going to give better raw figures than most crappy PC speakers can do justice to.

- these cards all show similarly flat frequency response, and it's unlikely you'd hear any difference on most PC speakers. (Also, the frequency response charts are so flat, they'd be better if rescaled to show much finer detail, e.g. 1dB vertical scale rather than 20dB)

- depending on the format, the noise levels and THD shown are all at -80dB (0.01%) to -100dB (0.001%) or less, over the audible hearing range. Again, you just won't hear that on most speakers, which will themselves introduce far more distortion than that! It seems you can safely remove that as a consideration when choosing between these cards

- likewise, stereo crosstalk is at -75 to -80dB or so (almost 0.01%) from 200Hz or so upwards, i.e. the range at which PC speakers are likely to be able to reproduce these sounds. Anyway, crosstalk isn't so important at lower frequencies, where most info is coming from a single subwoofer anyway

Thanks for the review and the loads of info! The almost-blind listening tests (apart from differing volume levels) are a nice inclusion.

So how does the Terratec DMX 6Fire score itself in a RMAA loopback test battery? Might as well find out how the "reference" recorder scores.

I wonder if pci-e versions would be about to use a hypermemery/turbocache type arcitecture using system memory to suppliment the onboard memory for sound buffers. A month or two Creative had made a statement elluding to waiting for 4x pci-e slots because of all things latency. Maybe just not enough bandwidth for sharing system memory? X-FI X-RAM MegaMemory Ultra Gamer ZS Gold Fatal1ty has a certain marketing ring to it.

Nice review. Few comments:

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.

I'm not sure what this means. Compression IS a common technique, but a competent studio engineer deliberately uses the technique to produce a desired effect. (An incompetent engineer uses it to make the whole album sound 'loud,' and s/he should be drug out into the street and shot. At that point the album is basically ruined regardless.) Altering it is fine if you happen to like the altered sound or need to compensate for weaknesses in your stereo (hence, the "tone" controls), but regaining lost dynamic range isn't within the range of capabilities here.

As I believe someone else pointed out, the Crystalizer is merely doing tone control at the expense of accuracy. I've been doing that with WinAmp equalizer for years.

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.

This sounds suspiciously like a "stereo width controller." In conventional analog hardware, a width controller merely introduces deliberate crosstalk between the channels to give the listener the effect of an open room, where both ears obviously hear portions of both stereo channels. A nice feature for those who want it (especially since very few receivers have width controls anymore), but not exactly innovative. Likewise for virtualized 3D -- that's been with us for a while, too. A comparison to how the X-Fi performs against previous implementations would have been useful.

What I don't get is why they make their cards rely so heavily on the breakout box. I like having all of my outputs in the back of my computer to hide the cables. I won't be getting this because first of all, I would need to fork up $200 just so I can use digial output and a microphone at the same time. PITA. Second, I have a door on my case that I would have to leave open with that card. lame.

All kidding aside about drivers, are the remaining 2 questions:
- any issues with the drivers?
- the thing still hogging PCI bus or not?

Is there a "software" graphic equalizer? If not, then I really don't care about the card as I listen to a lot of audio through headphones / headset.

<troll>Just a thought: "Creative" used to be "Creative Labs". Do you think they dropped the "Labs" part so they'd sound less like a research company and people would lower their expectations, innovation-wise?</troll>

I don't know if anyone's brought this up already but AWE32 can actually use up to 28MB of RAM if you stick 2 16MB 30pin SIMMs on it. 4MB is unusuable though.

But 28MB of RAM on a card designed in 1993 or so is pretty significant. Windows9x actually uses some of the RAM for DirectSound buffers as well. Otherwise it's all for Soundfonts.

Minor nitpick, but I wouldn't consider The Killers anywhere near punk or screamo. Theres no screaming/yelling first off. I'd just consider it rock, or pop. Things like Hawthorne Heights, Silverstein, and Thursday are more along the lines of screamo, with things such as Pennywise and NOFX being closer to punk. But thats just me.

Ahhhh audiophiles: As I think back to my studies of the nerd pyramid, you are amongst the top of the heap.

Good ol' on-board audio with digital out is working just fine for me.

I don't see this card offering much other than 15% more CPU that I probably wouldn't really notice, and less money in my pocket.

Going forward, I think TR should consider adding "64bit driver" as a feature to evaluate in any reviewed hardware -- is there one, is it out of beta, does it work? I don't (yet) expect tests/benches on Windows x64 but I think hw vendors should be held accountable for driver availability at this point.

Still waiting for the obligatory carping about Creative's drivers.....

Minor quibble....it's DTS-ES and DD-EX, not DTS-EX and DD-ES.

Hey thanks Dissonance for another sound review, its great to see your impression of the X-FI.

I was wanting to 'geek out' as you put it over the RMAA graphs you have included. But I, as I am sure are allot of TR readers, was unable to make sense of the graphs.

I know they are an additional extra, but I would have appreciated an explanation of their significance. Sorry I am a bit of an audio novice.

Compliments on the review. You have just the right mix of computerized and human impressions of each card. Other sound card reviews don't get that quite as right. Well done.

Due to the terrible port cluster on this thing, I think you should have been a bit harder in the conclusion. The convenience of discrete ports is very important when you consider having to reach behind the case to swap cables, if that's possible at all. In my mind, that makes this comparison good for audio quality alone. A more realistic comparison woudl be between the Audigy, Revo, and the X-Fi on the next highest rung of the X-Fi ladder.

Judging from the tables and geek graphs in the review, the X-Fi doesn't look like it's /that/ much better than the Aufigy 2 or Revo7.1. It's a good successor to the Audigy 2, but once you plow through the Crystallizer nonsense, it's another incremental upgrade.

Anyone think of folding on this audio proc? :P

Looks like the kind of card that is good to replace the A2ZS line, but not really enough to make we want to upgrade from my Audigy 2 ZS platinum.
As was stated, this card looks to be the 'lite' version with most of the people on here needing to step up to the next best mode.

Good review though. Thanks 8)

I knew there was a reason why I never sleep ha time to read