Auzen X-Fi Bravura 7.1 Sound Card Review

Auzentech made sure that even those music lovers who do not have a huge budget could enjoy their best technologies in an affordable form – for a little over $100. But is it really possible to make an audiophile grade product at this price point and at what expense will that be? We will try to answer all these questions today.

by Sergey Romanov
12/28/2010 | 12:53 PM

Auzentech made sure that even those music lovers who do not have a huge budget could enjoy their best technologies in an affordable form – for a little over $100. But is it really possible to make an audiophile grade product at this price point and at what expense will that be? We will try to answer all these questions today.


The Auzen brand didn’t even exist some four years ago but now the company’s product range includes a number of sound cards and we covered the X-Fi Forte 7.1 and X-Fi HomeTheater 7.1 models in our previous reviews. It’s time to take a look at yet another of Auzentech’s products which is called X-Fi Bravura 7.1. This sound card may seem to be just a different combination of the constituent parts we’ve already seen in its predecessors, but the reality is far more sophisticated.


I want to begin this review by telling you that the letters “X-Fi” in the name of the sound card I am talking about have no relation to the CA20K audio processor the X-Fi abbreviation was originally used to denote. The Creative X-Fi Native PCI Express audio processor mentioned in the product specs and press releases is in fact a PCI Express audio controller which was first employed in the affordable SoundBlaster PCI Express X-Fi Xtreme Audio card that Creative positions separately from its gaming products as a solution for music and movies. Auzentech’s marketing people have a different opinion. They say that nearly all of the Auzen sound cards are good for music. Then, the HomeTheater is best for movies, the Forte being somewhat deficient in this respect. And when it comes to gaming, the Bravura has a rating of three out of five stars whereas the Auzen cards with full-featured X-Fi processors have four stars (that’s rather odd as there seem to be no better sound cards for gaming as yet).

In our previous review and on the back of the product box you can see a somewhat different classification from the same source. The marketing folks from Auzentech now think that the Bravura has become better for movies and gaming. This is just one more incentive for me to test the card carefully because manufacturers’ claims often fall short of reality. For example, the Prelude, Forte and HomeTheater are far from identical in terms of music playback despite the same star ratings and the fact that the latter two cards seem to have almost identical analog sections. The Bravura is similar to them as well, judging by the employed components, yet I wouldn’t venture to suppose how it may sound.

But let’s get back to the electronic chips. The Bravura specs tell us that the card is based on a Creative X-Fi Native PCI Express audio processor. The same processor is specified for the X-Fi Forte and the HomeTheater but the Bravura is actually based on a different chip. The text on the product page says the processor is on the back of the PCB but that’s not so, as you can see in the photos. But maybe these chips are so similar that there is no reason for me to worry? Well, the difference is large, unfortunately.

Just like the CA20K2 is an enhanced CA20K1, the CA0110 is a revised CA0111 chip with such additions as a PCI Express-PCI bridge, an UAA-compatible (universal audio architecture) HDA controller, playback support for 44.1 kHz audio, and a fifth independently clocked stereo output. However, there was no audio processor in the CA0111 and none is present in the CA0110. Therefore the Auzen X-Fi Bravura lacks many capabilities we have got used to after the previous products from Auzentech and Creative. All sound processing is carried out by the driver which is almost the same as Creative offers to mainboard makers under the X-Fi Motherboard program.

Thus, the main difference between a mainboard with X-Fi driver and a discrete CA0110-based sound card is that the latter has a higher-quality analog section and features a headphone output which is independent from the main (multichannel) section. Let’s now discuss these features in more detail.

Packaging and Accessories

As opposed to the Auzen X-Fi HomeTheater HD which comes in two separate boxes, the Bravura is shipped in standard packaging.


Besides the sound card, the box contains:

The accessories are standard for Auzentech products and should be sufficient for a majority of users.

PCB Design and Technical Specifications

The specs and features of the Auzen X-Fi Bravura 7.1 are detailed at the manufacturer’s website and I don’t want to simply copy and paste that information. I will just dwell on a few of the most important things.

The Bravura is extremely close to Auzentech’s X-Fi Forte and HomeTheater products in its functions, components and exclusive technologies. It too has a symmetric power circuit, two microphone amplifiers with independent converters, and an individual headphone output. The similarity with the Forte is emphasized by the AK4396 DAC in the headphone amplifier, but the rest of the circuitry is different. In the Forte, the headphone output duplicated the line output of the front channel, making the individual converter unnecessary, so it was removed in the HomeTheater. Now it returns with an important mission. It allows the headphone amplifier to work independently of the line output.

The amplifier itself has changed, too. It used to be driven by the front output’s low-pass filter but now the headphone amplifier is itself part of the filter. ASUS implemented this solution in its Xonar Essence STX, by the way. Otherwise, it is the same dual-channel amplifier circuit with a buffer of eight transistors. As if in response to my thoughts about the weak power supply of this circuit on the Auzen X-Fi HomeTheater, the engineers installed four polymer capacitors instead of two aluminum ones into the power converter and added three thick printed conductors into the amplifier itself, obviously to better cool the transistors. We’ll see the effect of this shortly.

As opposed to the previous Auzentech products and to the card’s own line outputs, the Bravura’s headphone amplifier filter employs higher-quality capacitors. It is the metallized propylene FKP2 manufactured by WIMA. The opamp in the headphone output is swappable without soldering, which is handy. By the way, there are four empty places for transistors in the headphone amplifier and a place for one more opamp on the reverse side of the PCB. It seems that the engineers had wanted to separate the headphone amplifier from the DAC filter at first.

The headphone output is designed for users who prefer 6.3mm rather than 3.5mm connectors. Besides professional audio equipment, 6.3mm headphones could only be directly connected to the ASUS Xonar Essence. Owners of other sound cards had to use an awkward adapter or an external amplifier.

Judging by the opamp compatibility table, the design of the line output filters has changed somewhat since Auzentech’s earlier products. Although the Bravura is not declared to deliver optimum sound quality with the decompensated opamps of the OPA637 series, I didn’t have any problems with them in my tests. So, you can use such opamps with your Bravura if their price suits you.

The inputs have remained the same as in the previous Auzentech products except for the different way of switching the operation mode of the microphone amplifier. Instead of a switch, there is now an electronic selector. You can refer to the How To section of the product page for details. This unit is also responsible for the operation mode of the headphone output DAC and allows to enable a smooth frequency response decay or lower the peak output signal level. The headphone and line outputs are controlled in sync but by different rules.

It is unclear why the Auzentech engineers preferred to implement this array of hard-to-access micro-switches and barely visible LEDs. Software-based control over these features would be far more practical.

Testbed Configuration and Testing Participants

The sound cards were tested on a testbed with the following configuration:

I will compare the Auzen X-Fi Bravura with the Auzen X-Fi Prelude 7.1 and the ASUS Xonar D2 as well as with the Creative PCI Express X-Fi Xtreme Audio which is based on the same CA0110 audio controller as the Bravura itself.

The PCI Express Xtreme Audio makes use of Asahi Kasei DACs instead of Cirrus Logic ones you can see on other Creative cards. An eight-channel AK4359 (106 dB DR, -94 dB THD) is responsible for the card’s line outputs and a stereophonic AK4388 (104 dB DR, -90 THD) for the headphone output. The converters of the other sound cards covered in this review have better specs. For example, the Bravura’s line outputs are serviced by an eight-channel CS4382A (110 dB DR, -100 dB THD), and a stereophonic AK4396 (117 dB DR, -99 dB THD) is responsible for its headphone output.

The analog-to-digital converters of the PCI Express Xtreme Audio were chosen regardless of the tradition, too. These are two stereophonic AK5358 ADCs with rather mediocre parameters (about -100 dB DR and -93 dB THD depending on the sampling rate). For example, a WM8782 chip with the same specs is only responsible for the front-panel microphone input in the Auzen Bravura and many other Creative products. On the other hand, the WM8775 chip in the line inputs of these cards is not much better in its specs and is far inferior to the parameters of their analog outputs.

Winding up this electronics-related part of the review, I want to take a look at the operation amplifiers employed in the playback sections of the sound cards.

The Auzen Bravura has three different opamp models and you can replace them in each of the five output channels. Thus, users of this sound card are offered a unique opportunity to evaluate the effect that opamps may have on sound quality without purchasing additional chips. The default configuration is like that: an LME49720NA in the headphone amplifier, an NJM4580 for the center and subwoofer and an NJM5532 for the rest of the output channels.

Creative are more conservative in this respect and do not offer swappable opamps in their products. The PCI Express Xtreme Audio has NJM4556A opamps in the front channel of its line output as well as in the headphone output (for the front panel of the system case). It features very low distortions even at a 200-Ohm load. The rear channel is based on an NJM4558 which is not really meant for audio.

The table below summarizes the key specs of the abovementioned opamps together with those of a related model which was employed in the Creative X-Fi Elite Pro as well as in some E-MU products.

The specified distortion data for the LME49720 were gathered under special conditions, precluding us from directly comparing them to the New Japan Radio data. However, I can tell you that the LME49720 has a distortion level of less than 0.0001% in most applications. The NJM4556A is second best, ensuring the lowest level of distortions under harsher conditions than the rest of these chips. The downside of this opamp is its rather slow performance, resulting in a sudden growth of distortions at high frequencies. The NJM4580 has low distortions at low frequencies and is better high frequencies than the NJM4556, therefore I have no objections to Auzentech's installing it into the center and subwoofer channels. The use of the NJM5532 for the rest of the line outputs is questionable, though. We shall see shortly if this opamp should be replaced with something else.

Interestingly, there are four chips from New Japan Radio that are designed in an identical way but differ in their operation modes. The current balance is biased in the NJM4580 towards the input cascade. The NJM4558 can be viewed as a slowed-down NJM4556 with a weaker output cascade. The low operation current of this chip provokes such distortions that the manufacturer is ashamed of voicing the numbers. You can get a notion of them by the example of the NJM2068 which features the same circuit design and differs from the NJM4558 in slew rate and gain band product: the NJM2068 is declared to have a distortion level of 0.001% whereas the slower NJM4580 has a distortion level of 0.0005%.

Thus, we can see that the rear channels of the Creative PCI Express X-Fi Xtreme audio are inferior in quality to any components of the Auzen Bravura as well as of the rest of the sound cards I'm going to use in my tests. As for the front line output, I can’t make any predictions as yet.

Installation and Software

Unlike in my previous sound card reviews, I tested my Auzen X-Fi Bravura 7.1 in two OSes: Windows XP 32-bit and Windows 7 64-bit.

I installed the Windows XP driver from the included disc without problems but the sound quality was far from promised and vastly inferior to the Creative card with the same audio controller. I then installed the Windows 7 driver but it would not work. The beta driver downloaded from the Auzentech website did not provide the expected quality of sound, either.

Fortunately, PCI Express sound cards with Creative chips are compatible with Microsoft Unified Audio Architecture and with the generic Windows driver which, as my tests showed, can be used for listening to music and any other applications except for games.

You can refer to the next section for details about using the Bravura in games. Right now, let's see what the Auzen/Creative software can do.

The application suite isn’t very extensive, but the list above doesn’t show a few components that are installed anyway and appear as individual items in the Add/Remove Programs menu. These are Auzentech Control Panel, Creative Sound Blaster Properties, and Host OpenAL. The Volume Panel component is not installed by default although it is an indispensible means of setting the sound card up for games.

Originally expected in February and postponed to the second quarter, the Auzen Bravura driver with support for DTS Connect was only released in late July and only for Windows Vista and 7. Later on, there appeared two drivers for the Bravura, one of which supports DTS and another is WHQL-certified. I used the DTS driver for Windows 7 and the WHQL driver for Windows XP. DTS Connect became available to Windows XP users in October when the driver was updated to version 1.40.

The latest versions of the driver have improved the sound of the Bravura to the level comparable to its competitors but the driver's functionality is still limited in Windows XP. The headphone volume cannot be controlled independently and there is no OpenAL. Sampling rates of 192 and 44.1 kHz are not available although the latter is most important for a sound card that claims to reproduce music with high quality.

The mentioned features are all available under Windows 7. ASIO is the only missing one. The sound card is identified by the OS as three independent playback devices (Speakers, Headphones and S/PDIF), each with its own set of enhancements and sampling rates.



If you use the Microsoft High Definition Audio driver, the sound card is represented with the same three devices and supports the full range of sampling rates. The sound processing options are even superior to those offered by the Creative driver because the sound volume can be equalized according to the acoustic properties of the environment. The latter feature was only available with sophisticated and expensive studio equipment until recently but now it is offered to every Windows user!

If the native Creative X-Fi Xtreme Audio driver is installed, the Xtreme Audio card loses the Headphone device whereas the Bravura is still separated into multiple devices, which is somewhat problematic as I will explain below.

The sound card being separated into two devices, you can set up their sound volume independently. Auzentech also promised an additional control in the audio console which had been used to adjust the volume of the external unit of the Creative X-Fi Elite Pro.

Having installed the Auzen Bravura software, I could not see anything like that.

Thus, you can only adjust the volume of the Bravura’s headphone output in the Windows control panel or via the system tray icon if the Headphone device is selected as the computer’s default sound playback device.

As opposed to the sound cards with the full-featured X-Fi audio processor, the volume control of both CA0110-based cards does not change the volume level in a logarithmic manner. Thus, you have to select the volume in the bottom part of the range with more accuracy whereas moving the control in the top part of the range doesn't change the volume much. The top level of the Bravura’s output signal is 3 volts against 2 volts with the other manufacturers' cards, so getting the desired volume from it proves to be as inconvenient as with sound cards based on C-Media audio controllers whose volume control works in a linear manner (50% volume corresponds to a mere 6dB decrease in the output signal power). This is especially annoying when you use high-sensitivity headphones with the Bravura.

The software of both CA0110-based cards has the user interface of the audio console of the full-featured X-Fi in the Entertainment mode but its functionality is limited. Speaker system parameters are specified using Windows settings. The headphone and digital output settings have become unnecessary because those devices are separate from the sound card proper. You can enable EAX reverberation, set up Crystalizer and CMSS-3D parameters and choose the operation mode of the input connectors.

Despite the familiar names of the exclusive technologies, the 3D sound, reverberation and Crystalizer effects are calculated in the driver and differ in capabilities and quality from those of sound cards with a full-featured X-Fi audio processor. For example, CMSS-3D technology lacks the MacroFX and Elevation Filter options here.

I had noted how slow Creative’s software is in nearly each of my reviews of Auzen sound cards, but this time around the programmers overdid themselves. The test configuration was fast enough to avoid any discomfort in Windows XP, but everything was downright slow in Windows 7 64-bit. I had to wait for many seconds for a reaction to my changing some settings, one of the CPU cores being 100% loaded at that. The audio console and the Audio Control Panel would hang up for a few seconds even when I just plugged my headphones into the sound card!

That’s all I have to tell you about the card’s software. Let’s move on to the tests.

Listening to Music

I reproduced music using foobar2000 beta 2 with the latest sound output plugins. I used the same plugins for all the sound cards for the comparison to be correct since the Bravura doesn't support ASIO in any OS. The plugins were Kernel Streaming for Windows XP and WASAPI for Windows 7.

I didn’t use foobar2000’s standard output interface for a number of reasons. It is susceptible to software processing in the Windows audio subsystem under Windows 7. When used in Windows XP, DirectSound is much quieter than Kernel Streaming with the tested sound cards. As I found out later, the Wave control has to be set at its maximum in the mixer to achieve a normal volume level. I guess the reduced volume is necessary to minimize distortions when special effects are applied to audio. By the way, the special effects are not turned off in Windows XP even when Kernel Streaming is selected for audio output whereas the WASAPI interface provides additional advantages in Windows Vista and 7: you can keep CMSS-3D enabled when listening to music and do not have to worry about whether the sampling rate of the audio material agrees with the Windows settings. In other words, WASAPI saves you the trouble of adjusting the sound card’s settings and interacting with the sluggish software. This is better than, for example, the behavior of ASUS Xonar series cards which apply special effects (particularly, Dolby 3D sound algorithms) irrespective of the audio playback method under any version of Windows.

As I mentioned above, only two sampling rates, 48 and 96 kHz, are supported in Windows XP. It means that CD-DA music is always resampled, which is not good news for users who have accumulated their music collections by grabbing audio CDs. Many modern games use 44kHz audio, too. Fortunately, the resampling algorithms have improved since the SoundBlaster Live! and Audigy 2, so listening to music in Windows XP was quite a pleasure with every sound card included into this test session (considering that the latest driver version was installed for the Bravura).

The bulk of this test session was performed in Windows 7 64-bit where the sampling rate of 44.1 kHz is available both for the headphone and for the line outputs. Availability doesn’t mean full support, though, as was made clear by the integrated SoundMax audio in RightMark Audio Analyzer. Both CA0110-based sound cards are far from perfect in this respect, too. In my reviews of all ASUS Xonar series products I mentioned the fact that the sound card's SNR deteriorates at a sampling rate of 44.1 kHz. Well, this deterioration is far smaller than what you can hear with the Auzen X-Fi Bravura. The latter’s sound quality degenerates so much that it becomes a real problem when listening to music at a volume of 40% or lower. And while you can set the volume of the line outputs higher and then reduce it on your amplifier, this high amount of noise is a real problem for the integrated headphone amplifier. No other sound card I have ever tested behaves so defiantly. You can only get rid of the incessant hiss that rises up whenever you play music or launch a game by switching the card to a sampling rate which is a multiple of 48 kHz, i.e. by using software resampling. As a result, in both operating systems your music will be converted into the higher sampling rate which is hardly an optimal way of reproducing music. The Creative X-Fi Xtreme Audio also behaves like that, yet somewhat less irritatingly for some reason.

Besides a few select audiophile recordings from different test discs, I listened to Nina Simone’s A Single Woman, Roger Waters’ Amused to Death, Judas Priest’s Demolition and many others. Every sound card was switched to a sampling rate of 44.1 kHz in the Windows control panel and in the card’s own driver (the ASUS Xonar). The Auzen Prelude worked in the Audio Creation mode with Bit-Matched Playback enabled. For the rest of the cards I just turned off all kinds of sound processing. You can refer to my previous review for a description of terms that I use to characterize sound quality.

At first I compared the line outputs of the four sound cards by connecting them one by one to my headphone amplifier C.E.C. HD53R-80. As the ASUS Xonar D2 and the Auzen Prelude have LM4562 opamps in their front channels, I also installed the same opamp into the Auzen Bravura, but to a most unexpected effect.

First of all, the Bravura was very noisy. I could avoid the audible noise in its line output by raising the sound volume above 50% on the sound card and reducing it on the amplifier.

Second, its sound quality was far from impressive. While delivering good high frequencies and dynamics, it produced muffled bass and lacked details (echoes and aftersounds).

The Prelude was softer, more natural and comfortable. It sounded freer, so to say. The reverberations could be heard longer, the contrabass did not buzz annoyingly, the violins had a soul and the vocals had more softness.

The D2 had a peculiar way of building the sound scene. The bass was kind of more fluid, the high and medium frequencies were muddier, and the vocals were less distinct. The violins were rather irritating but the string instruments and the piano were reproduced excellently. This sound card did not lose resolution when reproducing polyphony.

The XA was inferior in resolution only: the vocals were less distinct and the high frequencies were blurrier and louder, which may be an indication of a higher level of distortions. The hissing wasn’t as oppressive as with the Bravura and the piano had a more impressive sound. The contrabass was sharper and more powerful and the echoes would fade out in a more natural way, even though rather too fast.

Although the inexpensive XA was inferior to the Bravura in the resolution and depth of the sound scene, in the detailed reproduction of trebles and in the distinctiveness of reverberations, it was more comfortable and exciting to listen to with most of my music compositions.

Sorted in the order of decreasing comfort and sound quality, the sound cards would have the following standings, depending on the particular music material:

Prelude — Bravura — XA — D2
Prelude — XA — D2 — Bravura
D2 — Prelude — XA — Bravura
Prelude — XA — Bravura — D2
Prelude — Bravura — XA — D2

Such unexpected results made me do some measurements to compare operating amplifiers. Renewing my tests of the sound cards a few weeks later, I found out that the Bravura's line output had improved greatly in terms of sound quality. I guess that it is the blocking capacitors in the card's line output that are the reason for the changes in the Bravura's sound quality over time. These are the tiny green Nichicon Muse that Auzen first installed on the X-Fi Forte and later on the Home Theater. All of these three sound cards have the same symptoms: the level of detail and resolution would increase and the timbres and bass would improve after a while. These non-polarized electrolytic capacitors seem to take a lot of time until they cease to spoil the sound. For example, the Bravura’s sound became clearer throughout the entire frequency range, producing a deeper sound scene. The instruments and vocals got more distinct, clear and realistic. The changes were the most conspicuous in the lower frequency range: the bass got sharper and saturated while the contrabass was not muffled anymore.

My measurements helped me find the reason for the surprisingly poor performance of the Xonar D2. This sound card occasionally enabled signal processing like what was noted in my and some other reviews.

This error is not related to ASIO and disappears when the data format for the playback device is changed in the Windows control panel. I will discuss the problem of setting up ASUS sound cards in Windows Vista and 7 in my upcoming review. Now let’s get back to the Auzen Bravura.

The final comparison of the sound cards dethroned the Auzen Prelude.

Judas Priest

The D2 features a clear sound with quick bass and inconspicuous trebles. The only unnatural thing about its sound is the dimensions of the sound scene.

The XA puts an emphasis on bass. Its sound is kind of slower, duller but not tiresome. Its resolution is very good with an accurately built sound scene, high transparency, and an exquisite separation of individual voices and instruments. Its bass is somewhat too soft.

The Bravura was not comfortable to listen to due to an emphasis on trebles. It had less of bass, prominent top middle frequencies, and hissing trebles. The scene lacked depth, worsening the sound resolution.

The Prelude seemed to have everything in good measure, yet its sound was rather dull, pale and unexciting. Its resolution was inferior to that of every other tested card. Once again I have to confess that this card just cannot reproduce a bass guitar properly.

Nina Simone

The DS was almost perfect with this material.

The Prelude produced a less spacious and less detailed scene. It was not as dynamic as the leader and had somewhat lisping trebles.

The XA is closer to the D2 is dynamics but has muddy trebles and a lower resolution. Its reverberations interfere with the vocals, making the latter less clear, and fade out unnaturally quickly.

The Bravura showed a good resolution at medium frequencies but lacked the absorbing detailedness of the first two cards. The vocals, piano and cymbals were rather unpleasant with this sound card.

Roger Waters

The Bravura delivers a voluminous, incredibly clear and lush sound, especially with string instruments. It is detailed but lacks depth. The low frequencies are not stiff enough whereas the high frequencies are, on the contrary, too stiff.

The Prelude features a high level of detail, a very good stereo panorama, a clear and voluminous but rather too soft sound.

The D2 offers the most agreeable bass and an expressly voluminous sound scene. Occasionally it sounds too muddy and Roger begins to lisp.

The XA has poor depth and a low level of detail. Roger lisps occasionally.

Thus, a sound card may be the best in reproducing one musical material but fail with another. The four tested cards are roughly at the same level. There are no obvious losers among them and the quality of sound depends largely on the specifics of a particular recording. Here is another example with a couple of audiophile tracks.


The Bravura offers the richest timbre of the instrument and a voluminous sound scene.

The D2 has a detailed bottom range and a broad sound scene.

The Prelude allows to localize extra sounds and has a deep sound scene.

The XA is just good, although with a somewhat lower level of details compared to the other cards.

Piano + Male vocal

The D2 has a voluminous and impressive sound image and detailed high frequencies.

The Prelude renders soul-touching vocals but with hissing sibilants. Its level of detail is lower and the instrument has a muffled sound.

The XA has less airy and expressive, but is good otherwise.

The Bravura has a certain hiss which spoils the atmosphere. Both the vocal and the instrument lack expressiveness. The sound seems somewhat muddy overall.

I want to remind you that I reproduced the music through the WASAPI interface because the Bravura and XtremeAudio do not support ASIO. I tested the different interfaces briefly and made sure that the Auzen Prelude sounded clearer and purer via ASIO and thus would have been a leader across a lot of music materials.

I also have to note that the ASUS Xonar D2 had been modified by replacing the NJM2114 opamps with AD8066 ones. This helped me improve the card’s dynamics and get rid of an unpleasant coloring of its sound. Thus, the Xonar did somewhat better that it would have done by default, but the automatic driver-based resampling makes the Xonar the worst card in this review in many situations.

Finding the Optimal OPAMP

With sound cards of this level and sound quality, even small trifles are important.

The Auzen X-Fi Bravura 7.1 didn’t do well in terms of its line output quality. Even after it had warmed up, it would be often inferior to the Creative card based on the same controller and costing only half its price. The high-quality components help the Bravura outperform the more expensive cards in some situations, yet there’s something wrong with it anyway. The Auzen HomeTheater HD with the same analog section components was free from the downsides I observed with the Bravura. Besides, its LM4562 opamp didn’t suit it well. While drawing a good sound scene and delivering very detailed medium frequencies, it was not just vivid but aggressive.

I combined my search for the optimal opamp with tests of the Bravura’s integrated headphone amplifier by comparing the latter’s sound quality with that of the external amplifier C.E.C. HD53R-80 connected to the card’s line output. For the comparison to be fair, I installed identical opamps into both of the card’s channels by using all opamps I could find in two samples. I mostly used Evanescence’s Origin as the music material. The comparison took a long time and coincided with the card’s warming up and improving the quality of its line output but I didn’t re-listen to all the opamps, limiting myself to but a few models.


The headphone output provides a darker sound with somewhat more clarity and dynamics. These small improvements combine to make the listening to music via the card’s own amplifier far more exciting but Judas Priest was more fierce and detailed when listened to via the line output with external amplifier even though the integrated amplifier had more powerful and detailed low frequencies, good bass being so important for rock music.

To find out if it was the problem with the integrated amplifier I had noticed during my tests of the Auzen X-Fi HomeTheater HD, I connected the external amplifier to the headphone output but didn’t find any more details in the sound. It must be due to the output filter because the AK4396 played rock most vigorously on the Auzen X-Fi Forte.

Thus, the line output has a lower resolution with calm music, yet it doesn’t lower more with heavy music. The line output was also tiresomely clamorous with this opamp.

The chips surprised me with their large variations in sound quality. I had no complaints about only one of the three I tested whereas the other two delivered muddy trebles and a flat scene, even though to a different degree. This explains my negative impressions from the LME49720 in the reviews of the Auzen Forte and Auzen HomeTheater. The good sample proved to produce a clear and consistent sound.


This one delivers a viscous bass and smoothed-out dynamics on both outputs. Evanescence was reproduced downright sloppy, with a lack of energy and an unstable sound scene. The line output seems to produce more details but this may be due to a higher level of harmonics because the headphone output plays the music in a more natural way with better bass. The piano was listless when listened to through the line output whereas the drums were clearer through the headphone output and the subsequent low-frequency hum was more saturated. The two outputs of the sound card were equal in terms of vocals. High frequencies differed but I couldn’t determine the better output for reproducing them: the line output was perhaps more natural in some recordings.

Overall, the headphone output produced a more voluminous and freer sound with a saturated bottom range.


The line output seemed to be somewhat more comfortable to listen to at first, but also simpler and far less exciting than the headphone output. But after the card’s warming up for a few months my preferences shifted towards the line output even though its sound was still far from ideal.

This opamp wasn’t very expressive in the headphone output, lacking a sharp bass even though its macrodynamics was okay. It was merrier in the line output: although there might be less bass overall, the low frequencies were very detailed.

One sample of this opamp distorted high frequencies, made the sound scene flat, cut off low frequencies and reduced the level of detail. It produced a distorted sound of the piano, an annoying sax and uncomely high vocals. The sound of rain would be something like hail with it. Another sample was more voluminous, brighter and prettier to listen to except that its high frequencies were not very detailed.


This opamp is most energetic but somewhat flat. There is no punch, the high frequencies are smoothed out, the level of detail is far from impressive, and the vocals do not excite, either. On the other hand, the sound is not irritating, contrary to the Creative X-Fi Elite Pro that has this opamp in its front channel. So, the NJM2114 turns to be able to sound comfortably under some conditions, yet it is still inferior even to the NJM5532.

The two samples of this opamp I had differed in nearly everything: dynamics, the depth of low frequencies and the quality of high ones, the spaciousness and detailedness of the sound scene. One sample was obviously better than another and produced a rather good sound in the headphone amplifier.


This opamp produced a condensed and saturated sound in the headphone output but high frequencies were fuzzy (without a clearly defined attack) and lusterless. The sound was not very detailed and the bass was just barely noticeable.

When installed into the card’s line output, this opamp delivered a lively and voluminous sound with beautiful vocals, but the piano didn't sound pretty. The bass was soft but there was enough of it; the high frequencies were more sonorous but not very detailed.

It’s hard to prefer one output here: the headphone output was clearer and more detailed but the line output had more spaciousness. The sound quality wasn’t very high in both cases, though. It’s similar to the NJM5532 but somewhat more comfortable to listen to.


This opamp isn’t very good in terms of bass but delivered detailed high frequencies and a characteristic sound scene. Its sound doesn’t resemble the other opamps. It may be unpleasant in some places but fascinating in others. The more I was listening to it, the more I liked it. I wonder if opamps take some time to warm up, too.

The two samples of this opamp were different, too. One was somewhat inferior in terms of dynamics and low frequencies, producing a flat and irritating sound. The other produced a full, comfortable and vigorous sound; the drums and thunder had the necessary sharpness.

When used with the OPA2228P, the headphone output produced a smoothed-out sound, like with the NJM2114, yet with clearly defined details.

The line output, on the contrary, was voluminous and dynamic, and the two samples of this opamp differed less dramatically in it. This seems to be the first opamp worthy of the line output.


When in the line output, it produces a sound scene which is deep but not broad. It also gives you a detailed bass, good medium frequencies and rather odd trebles which are clear and detailed at low sound volume (like the sound of rain, for example) but loud sounds (sibilants or cymbals) become more like a hiss.

When in the headphone amplifier, this opamp, to my surprise, was not so good. While delivering somewhat more details than I could hear via the external amplifier connected to the card’s line output, its sound was not so rich in timbres, the sound scene lacked depth and the bass lacked energy.

The LT1364 adds transparency and “darkness” to the line output. When playing various genres of rock, the line output with the LT1364 was often superior to the headphone output with the LME49720.

Unfortunately, the two samples of this opamp that I tested differed greatly in terms of sound quality. The worse one emphasized medium frequencies too much, also producing hissing trebles and lowering the level of detail. While I couldn’t decide which of my two LM6172IN opamps was better, this choice was simple with the LT1364 pair: listening to music was a pleasure with one sample and an annoying experience with the other. However, the annoying one had better dynamics and reproduced the sound of rain in a more realistic way whereas the other one muffled trebles somewhat.


The two samples of this opamp didn’t differ much. One was clearer, deeper and had more bass whereas the other was more aggressive and had muddy high frequencies.

At first I liked them more in the headphone amplifier as they played many compositions (e.g. Roger Waters) with more spaciousness, neatness and purity compared to the line output where their sound was less transparent. The situation changed dramatically after the card had warmed up: they then delivered a very energetic and comfortable sound via the line output, with a responsive bass, detailed trebles and attractive vocals. When used in the headphone amplifier, they seemed to lack dynamics and accuracy.

Overall, I liked the OPA2228P and the LT1361CN8 the most on the line output whereas the LME series opamps seemed to be the best choice for the sound card’s headphone amplifier. The series includes the LME49710 you can order from Auzentech. The LME49710UA produces a spacious and accurate sound scene with a good bass and nice trebles whereas the LME49710HA is less energetic but had the best results according my measurements.

I also liked the AD826 in the headphone amplifier. It produced a very dynamic and spacious sound but it was the metallic OPA637SM that, like with the Auzen X-Fi HomeTheater, had the biggest effect on me. I could not hear so many details, so much air and liveliness with any other opamp, including the similar OPA637AU which was somewhat more clamorous. Alas, the pleasure of using this opamp is going to cost twice the price of the sound card itself!

As a matter of fact, I wonder if purchasing replacement opamps is really worth the trouble because the difference in subjective sound quality between different opamp models is not much larger than between different samples of the same model. This may be the reason why expensive opamps that have passed a stricter quality assurance check are so valued among audiophiles. Anyway, if you are not satisfied with the sound of your Auzen Bravura half a year after purchasing it, try swapping round its default opamps. If you do notice some positive changes then, you may want to purchase other opamps to improve the sound even more.

As for the quality of the integrated headphone amplifier, I could not find any fault with it. Even at a load of 32 Ohms it was as good as the expensive external amplifier C.E.C. HD53R-80 in purity, dynamics and power which was more than enough for my Grado SR 325i headphones. The only problem was the incessant hiss when the card was working at a sampling rate of 44.1 kHz. For the background noise to get inaudible, less sensitive headphones are necessary.

Winding up my subjective tests of sound quality, I checked out the different operation modes of the sound card’s outputs. Pressing the onboard switch changes the sound volume of both outputs. A smooth decline of high frequencies is enabled for the headphone amplifier in three out of the four modes. I personally think that the integrated amplifier is more comfortable and natural in modes 2, 3 and 4. For more details about that see the Measurements section.

Sound in Games

It’s no secret that after Windows Vista put an end to hardware DirectSound3D acceleration more and more new games do no rely on the sound card’s capabilities but process audio with their own means. For example, Left 4 Dead and other games running on the updated Valve Source Orange Box engine output mixed audio depending on the specified speaker configuration (headphones, 2.0, 4.0, 5.1 or 7.1). This audio doesn't differ from what you get when playing a DVD, for example. As a result, such games have the same sound irrespective of whether you use an integrated sound core or an expensive SoundBlaster, except for the difference in the quality of the analog sections which is discernable in music. On the other hand, most games with Unreal Engine 3 employ OpenAL and rely on software emulation only if the sound card doesn't support OpenAL. In this case, like when the audio is processed by the game itself, the highest quality of sound, special effects and spatial positioning is achieved when the sound card's audio processing features are disabled.

Every card included into this test session supports OpenAL in Windows 7 to some extent. The Auzen Prelude is the only card to process OpenAL on the hardware whereas the others do that on the level of the OS driver. There is a special Host OpenAL driver for the Auzen Bravura and Creative Xtreme Audio whereas the ASUS Xonar relies on the Generic Hardware path of the OpenAL driver.

The OpenAL core, which has been developed under Creative’s aegis since 1999, offers three types of playback devices to applications. The Native type directly accesses the card’s hardware and provides the highest quality of effects because it applies them in 3D space and only then mixes the sounds up for the required number of output audio channels. The Generic Hardware type translates OpenAL calls into the DirectSound3D space model to enable the capabilities of any sound cards that offer hardware DirectSound3D acceleration. Unfortunately, the translator limits the number of available voices (sound sources) and the quality of effects is inferior to the Native type. Besides, the hardware DirectSound3D was cut out of the Windows audio subsystem, making Creative and C-Media develop Alchemy and Xear3D EX, respectively (ASUS renamed the latter into Gaming eXtensions or GX). The third OpenAL device type is a simple audio engine which offers some basic sound processing. Its audio is meant for stereo speakers and doesn’t sound natural in headphones. The distance to a sound source is only indicated by volume rather than timbre.

The CA0110 audio controller doesn’t have an integrated audio processor but the programmers wrote a native OpenAL driver for it which is free from the limitations of a Generic Hardware device. This driver is compatible with Windows Vista and 7 and works with only select sound cards because Creative is not so willing to grant its technologies for free. C-Media’s programmers began to write a similar driver back in 2007 but haven’t yet produced a functioning result. ASUS sound cards are still dependent on the Generic Hardware driver which limits their functionality and works only when the GX feature is turned on. While I had no problems with the version 8.17.77 driver for Windows 7, users at different forums report all manner of abnormalities the active GX feature can provoke. It must also be noted that turning GX off under Windows Vista or 7 deprives the sound card of all 3D sound processing in games except for the algorithms of converting the number of sound channels.

A sound card’s functionality can be useful even in a game that processes audio itself if you specify in the Windows control panel that your playback device has 8 audio channels and choose the real configuration of your speaker system in the sound card settings. In this case the game can output its sound with the highest quality possible whereas the sound card will be transforming it into the necessary format by utilizing its own exclusive technologies. This is especially important for headphones because the audio engines of most games do not do any headphone-optimized virtualization or do it poorly. The same goes for the sound effects library in Windows 7: every sound card I tested could prepare audio for headphones with much higher quality.

In the early versions of the Bravura driver for Windows Vista/7 the CMSS-3D technology could be enabled separately for Speakers and Headphones but the standard control options for the sound card would confuse these two outputs. The Enable X-Fi CMSS-3D checkbox in the Audio Control Panel and the Audio Console did not represent the actual status of that technology. The only reliable way to control it was to right-click the Volume Panel in the system tray but you had to install that application yourself.

To get the correct 3D sound with the early driver versions you had to choose the appropriate speaker configuration in the Audio Control Panel.

The version 1.40 driver lacks this setting and you can choose the speaker configuration when installing the driver. Perhaps this setting doesn’t affect anything now.

The separation of the Bravura into two devices in Windows Vista/7 has its pitfalls. For example, a game that has its own audio engine will not output multichannel audio to the Headphones device because you cannot change the number of channels for it. Besides, CMSS-3D will not work if you disable the Sound Blaster extensions for the audio device in the Windows control panel.

I tested the sounds cards’ gaming capabilities in a couple of games that have good ambient audio. The Auzen Prelude was switched into gaming mode with default settings. For the rest of the sound cards I enabled the sound processing features for headphones: CMSS-3D with the Auzen Bravura, Dolby Digital and GX with the ASUS Xonar D2. When you enable DH, the Xonar driver allows using 7.1 Virtual Speaker Shifter which affects game audio, so I additionally tested the Xonar with VSS enabled. By the way, there is a bug in the driver that leads to an incorrect distribution of audio among the virtual channels. It is cured by rolling VSS settings back to their defaults.


The game uses a rather high-quality audio engine, has no sound settings except for volume, and calculates the output for as many channels as are specified for the playback device in the Windows control panel.

In the 2.0 mode, when 3D sound technologies like Creative's CMSS-3D or ASUS’s Dolby Headphone are turned off, there is but a slight hint of front-to-back positioning but the distance towards the sound source is not reproduced at all. The tonal balance is shifted towards high frequencies: there is a lack of bass and sound pressure. The sound is flat and unexciting whereas the additional processing on the sound card is harmful in this mode, adding an artificial echo. For example, turning on CMSS on the Bravura puts every sound farther away but does not improve their positioning in space; the voice in the intro seems to be speaking into a glass jar then. Turning on CMSS on the Prelude in the Audio Creation mode also provides some distance towards the sound sources but the attack gets even worse than before. You can play the game at such settings but with little comfort because the sounds are hollow and you don't hear much from your back: the sound sources are all distributed in front of you or at the sides, producing a very odd effect. Turning on Dolby Headphone on the Xonar makes the sounds hollow, too. It becomes easier to localize sound sources in front of you but there is absolutely no positioning of sounds from behind. The sound gets too flaccid and unexciting overall. The Virtual Speaker Shifter option solves these problems but doesn’t improve the reproduction of distance; it produces too many low frequencies while high frequencies are muffled.

Switching the sound card to eight channels makes the game’s sound richer: the attack improves and the sound gets more detailed, high frequencies get more sonorous, and there are more low frequencies. The overall result is similar to the above-described DH+VSS but with better timbres and more details. For example, gunshots produce an echo in between buildings.

The Bravura with enabled CMSS did excellently with the intro clip (except that the high frequencies were somewhat screechy), but the in-game sound lacked depth. For example, the din of a crowd could only be heard from the sides. Besides, the background music was reproduced with a wheeze which was not so conspicuous with the other sound cards.

The Xonar with DH produces a better spatial sound even in the game menu and creates an atmosphere of horror from the very first sounds, even before the menu. The intro clip has more bass, but only due to the suppression of the rest of the frequency range. The bass itself isn't stiff. The in-game sound isn't impressive: the high frequencies are muffled, details are lacking, the different sounds are not as distinct as necessary, and there is an overall impression of a closed space. The surround effect is good, though. The sound sources are all accurately positioned in space and the distance to them is reproduced correctly. VSS brings back the lost details and saturation, adds more spaciousness by moving the closest sounds farther away, but the distant sounds are suppressed more than necessary. Still, this variant is the most exciting and scariest of all: there is a lot of low frequencies, the sounds are detailed, it is easy to orient oneself by ear.

The Prelude with CMSS boasts pure high frequencies which I couldn’t get from the two other cards even when the text was being printed during the game’s loading a level. The sounds are even richer and more detailed than with the Bravura both in the intro and in the game itself. The 3D positioning is as good as with the Xonar. The distinctiveness of individual sounds is amazing: they are perfectly identifiable even from a long distance and are positioned correctly. Numerous voices do not mingle into a mess.

Thus, it is the Prelude that gives this game the best-sounding voice. To be exact, it is the X-Fi audio-processor in gaming mode (its performance is definitely worse in the other modes). Its junior cousin in the Bravura card is only good with the intro clip but produced an unrealistic in-game sound with a lack of details. The Xonar in DH+VSS mode takes second place but switching its sampling rate to 48 kHz in the Windows control panel and to 192 kHz in the ASUS Audio Center brings a significant improvement in the clarity of high frequencies and makes the Xonar as good as the Prelude.

Mirror’s Edge

This is a rather old, yet engaging game that presumably uses OpenAL. Thus, it is the sound card that is responsible for delivering the required number of audio channels. Without turning CMSS-3D on, there is absolutely no 3D positioning of sound sources and you get a dizzy impression of being right in the middle of a high-traffic thoroughfare. The only sound setting available in the game menu is the choice between hardware support and software emulation.

The Bravura does not ensure realistic 3D positioning. Like in Prototype, there is a lack of depth. You hear sounds closer than necessary and their perceived positions are unstable when the player is moving and turning around. There are a lot of loud echoes in closed environments, yet you get the feeling of a closed environment even when you are outdoors. The buzzing tone in the sound suggests a high level of intermodulation distortions.

The Prelude is far more realistic. Its high frequencies differ dramatically from those of the Bravura in this game: the ringing of the metallic mesh, the slaps of hands against the pipes and walls, the falling drops of water are all rich and natural. Distances are reproduced more correctly but 3D positioning of sounds is unrealistically precise whereas the background music is clicking in your head. These problems can hardly be the sound card's fault. In a closed environment the street sounds can be heard without the required fading-out. Once I didn’t hear the noise of water after automatically loading a control point but the wind was wailing at twice its normal volume.

The Xonar D2 has almost the same 3D positioning as the Prelude but the changes in the perceived position of sound sources when turning around are more sudden and somewhat jerky whereas the sound itself is more hollow and close to the Bravura’s in quality. The wind is louder. There are no reverberations in closed environments. Some sounds stop abruptly. There are odd sounds occurring occasionally when your character is running or hitting something. The Xonar would be better than the Bravura if it were not for the excessive amount of such artifacts in its sound.

To my surprise, choosing Generic Software in the game’s sound options not only corrected all the downsides of the Xonar but even pushed the realism of its sound to a much higher level. The sound got far more detailed, spacious, and clear, especially when I switched the Xonar to a sampling rate of 48 kHz in the Windows control panel and in the card’s own Audio Center. Don’t think I’m crazy or anything. Of course, I checked out the Generic Software option with each of the three sound cards and could only hear the sound deteriorate. However, the Xonar behaves differently from the Prelude and Bravura that can change the sound output mode right from the game. The Xonar reacted to the setting only after quitting the game. If the Hardware mode is selected when launching the game, the change of mode doesn't provoke any changes in sound. But if the Generic Software setting is selected...

The wind is whistling in your ears, you hear your breath, the sound of your steps, and the rustle of your clothes clearly, yet not obtrusively. Every sound around you is pure and ideally positioned according to the visuals. The echo is less hollow and, to my taste, more natural than with the Prelude. The Xonar was even better in reproducing the game atmosphere than the full-featured X-Fi of the Prelude, but the competition between the X-Fi and Xonar in games is a separate story. Let’s return to the Bravura.

Summing up the gaming section of this review, I can tell you that the Auzen Bravura turned to be rather weak in games because its software CMSS-3D implementation is far inferior in quality not only to the namesake technology of the sound cards with the full-featured X-Fi audio-processor but also to the 3D sound algorithms of the Xonar series products. Of course, EAX 4 and OpenAL are the Bravura’s advantages over simple integrated audio cores but mainboards with Creative X-Fi MB2 technology can already boast EAX 5 and the advanced sound processing algorithms called THX TrueStudio Pro.

Test Results

Like many other sound cards, the Auzen X-Fi Bravura doesn’t have an ADC of a high enough quality to measure the characteristics of its own outputs. Therefore for this review I employed a professional audio interface E-MU 1820 for that purpose. I would have gladly used it in my earlier tests if Auzen’s CA20K2-based X-Fi Forte and HomeTheater didn't conflict with it. Based on a different audio controller, the Bravura is perfectly compatible with the E-MU 1820 as well as with the Creative Xtreme Audio, Auzen Prelude and ASUS Xonar D2. To my surprise, these sound cards were all quite content working together under Windows 7 x64. The Prelude was the only one to occasionally conflict with the E-MU 1820 over ASIO, but this didn't prevent me from measuring its characteristics as well.

Another innovation in this review is a test of impulse response during which the tested sound card is made to reproduce a square wave with a sampling rate of 44.1 kHz. The result is recorded at maximum resolution the E-MU 1820 supports, namely 192 kHz/24 bit, which helps reveal any ringing at the wave fronts.

Finally, I measured the quality of cancellation of the byproducts of digital-to-analog conversion: ultrasound noise and aliasing noise. To perform this test I reproduced a sliding pure tone on the tested sound card, i.e. a sinusoid wave with a frequency of 20 Hz to 22 kHz.

Let’s first check out the results of traditional tests in Rightmark Audio Analyzer Pro 6.2.3. For the results to be as accurate as possible, I disabled all effects in the sound card's settings and selected a sampling rate in the Windows control panel to match the reproduced signal (44.1 or 48 kHz).

The results above suggest that the Auzen Bravura’s integrated ADC is greatly inferior to the professional audio interface in every parameter, yet the playback quality of 44.1kHz signal is even worse due to high noise. However, both measurements via the E-MU 1820 show excellent noise results at 44.1 kHz. Why?

The fact is, just when I began my tests with the E-MU 1820, the Auzen Bravura suddenly began to reproduce such signal with high quality when the device’s sampling rate was set at something other than 44.1 kHz in the control panel. The results were excellent when the sound was reproduced via DirectSound, but there were too much noise with WASAPI. My luck ended as suddenly as it had come. After some events, the most likely of which was the installation of the ASIO4ALL driver, the frequency automation that allowed reproducing audio of any quality without distortions just stopped to work. Below you can see two graphs that illustrate how great is the difference between the card's operation modes and outputs as well as between the Bravura's line input in comparison with the E-MU 1820.

Auzen Bravura Frequency

Auzen Bravura intermodulation distortions

As you can see, the Bravura’s headphone output delivers excellent results that surpass the capabilities of the E-MU 1820’s ADC, let alone the Bravura’s own line input. The line output distortions, on the contrary, are high and can be easily spotted even if measured via the card's own line input. The distortions provoked by the nonmatching sampling rates of the reproduced audio and the audio device setting in Windows 7 are just awful. Interestingly, when the sound card's headphone output is set at 96 and, especially, at 192 kHz, the distortions are not so conspicuous whereas the stereo panorama acquires more depth and the voices and instruments get a specific crystal-like clarity that some users may even like.

Anyway, sound cards with the Creative CA0110 chip generally deliver their best sound quality when the highest sampling rate is set in the Windows control panel. Users of such devices may want to learn the simple procedure of measuring with RMAA available at the official website in order to make sure that their sound card works properly. When measuring the parameters of an Auzen Bravura, you should keep it in mind that the maximum signal level is 3 volts on each of the card's outputs whereas the allowable input level is only 2 volts for some reason. Auzentech engineers didn’t follow the conventional level of 2 volts before, but they used to reset the sensitivity of the line input accordingly.

Taking advantage of the unexpected complaisance of the E-MU 1820 which agreed to work with all of the sound cards and was free from the ground loop problem which might worsen the measured characteristics, I could carry out some previously undoable measurements with my Auzen X-Fi Prelude 7.1. Its own line input didn’t provide the required quality and its results were not very high when measured via a Creative X-Fi Elite Pro or ASUS Xonar D2.

The Auzen Bravura isn’t bad in terms of headphone output quality, but the line output distortions are worse than with the other tested cards, including the Auzen Forte/HomeTheater that employ very similar components. I hope this is not a flaw in the circuit design but a manufacturing defect although neither gives the manufacturer any credit.

The Bravura’s headphone output surpasses the three other tested cards but in my ASUS Xonar D2 review I could see that sound card deliver higher results in terms of noise and distortions when tested via its own line input. Thus, the similar results of the three tested sound cards are due to their interaction with the E-MU 1820's inputs.

Frequency response of the four sound cards

Comparing the frequency response of the sound cards, we can note the odd hump in the Auzen Prelude’s graph and the waviness of the Creative PCIe XtremeAudio’s one. The latter can be explained by the documented feature of the digital filter of the AK4359 converter whereas the Prelude’s behavior is not so easy to comprehend. I will discuss it shortly. Here, I can note that the Auzen Prelude has an early decline in the low frequencies that the other sound cards do not have. I hadn’t noticed this before with this card, but it is the oldest among the samples I tested, so I suppose this must be the consequence of a loss of capacitance by the blocking electrolytic capacitors due to aging. Interestingly, the ASUS Xonar D2 and the Creative PCIe Xtreme Audio do not have capacitors in their line outputs and have a similar frequency response in the area of low frequencies. The Auzen Bravura's headphone output features the highest uniformity in high frequencies.

Intermodulation distortions of the four sound cards

The graphs showing correlation between distortions and frequency indicate a lack of anomalies that might be provoked by resampling when the sound card's settings do not match the reproduced audio material. Thus, the mystery of the accidentally discovered “magic” mode of the Auzen Bravura remains unclear. If it is software resampling, its quality is just amazing. The smooth rise in distortions at higher frequencies is due to the insufficiently high quality of the E-MU 1820’s inputs which limit the results of the Auzen Bravura. The Auzen Prelude is the closest to the Bravura’s performance despite being inferior to the other two cards in terms of noise which affects this test.

Now let’s take a look at the harmonics of all the four cards:

Bravura Headphones Out

Bravura Line Out


Xonar D2

PCIe Xtreme Audio

Despite the different circuit design and converters, the distortions on the Bravura’s two outputs are very similar but their level on the line output is many times that of the headphone output. Interestingly, such a high level of distortions is barely noticeable when listening to music. It is also interesting that the improvements in sound quality over time, which are quite perceptible to the ear, only resulted in a tiny improvement in the distortions level which might as well be due to a thermal drift of parameters or measurement inaccuracies.

The Creative PCIe Xtreme Audio, on the contrary, is a nice surprise. Being the cheapest product in this review, it proved to be the best in terms of the high harmonics, being only inferior to the Auzen Bravura in the second harmonic level and to the ASUS Xonar D2 in the third harmonic level. No wonder that it was quite competitive against the more expensive and renowned opponents in my subjective tests.

And now I'm going to show you the new tests. Below you can see square wave oscillograms illustrating the operation of the playback section’s digital and analog filters. The contribution of the E-MU’s line input filters is negligibly small because the recording was performed at a sampling rate of 192 kHz while the analog input filter is set at a rather high cutoff frequency.

Bravura Line Out

You can see the smooth increase in the signal amplitude on the Auzen Bravura’s line output which is not observed with the other sound cards. The Mute feature is usually implemented with field transistors or a relay but the extra components worsen the sound quality, therefore Auzentech's engineers selected some nontrivial method like switching the DAC into sleep mode. The operation of that mechanism is audible when the Bravura is set at a sampling rate of 44.1 kHz: when an application (a video game, for example) uses the sound card, there appears an audible noise. Otherwise, the Bravura's outputs are absolutely silent.

PCIe Xtreme Audio


The Auzen Prelude’s graph looks rather unusual, too. Judging by the article from Ayre, the maker of disc players and DACs, the developers of the Auzen Prelude implemented a minimum-phase filter in it. This is not the normal operation mode for the AK4396 converter because the other sound cards with this chip have the characteristic pre-ringing typical of a line-phase filter. The Prelude doesn’t have it at all but its post-ringing is far larger than usual in amplitude and duration. It is no wonder then that the Prelude sounds differently from other cards.

The Bravura’s line output filter is even more unusual: the pre-ringing is shortened to two cycles rather than eliminated altogether. That's what the phrase “we accurately designed the DAC filter” means! Interestingly, this optimization didn't concern the headphone output filter but you can enable a special filtering mode for it.

Bravura headphones out in Sharp filter and Smooth filter modes

Bravura headphones out in Sharp filter mode

Bravura headphones out in Smooth filter mode

In the Smooth filter mode the extra ringing is eliminated completely. Many people consider it to be an inherent downside of delta-sigma converters but you can see that the converter’s architecture has nothing to do with it. The “pretty” switching response is achieved by means of the digital filter with slow decay. However, this filter has one serious drawback. Take a look at the filtering quality.

The sonograms below show you the spectrum of the signal: the X-axis shows time and the Y-axis, frequency. The brightness of a dot indicates the intensity of the signal at that frequency.

Bravura headphones out in Sharp filter and Smooth filter modes

The lilac cloud at the top is ultrasound noise which is largely determined by the characteristics of the E-MU 1820’s ADC, but we can see quite clearly that it is brighter in the Smooth filter mode. That’s not the whole problem, though. The noise might be lowered but it is absolutely impossible to get rid of the aliasing which is supposed to be suppressed by the filter. The aliasing is harmful as it provokes intermodulation with the true signal in the subsequent analog section, making different sounds less distinct. The parasitic lines are perfectly visible in the sonogram of the line output and barely noticeable in the sonogram of the headphone output.

Bravura and PCIe Xtreme Audio Line Out

The purple lines deflecting from the yellow line in the sonogram of the Bravura’s line output represent harmonics of the main signal. Judging by the sonogram, the third and fifth harmonics are especially high, which agrees with the results I obtained with RMAA. We can also see a less distinct line of the seventh harmonic and a mesh of intermodulation products which is not visible with the Creative card but can be seen with the Auzen Prelude.

Prelude at 100% and 75% volume

The nonstandard implementation of the Auzen Prelude’s filter is good at suppressing aliasing but provokes unexpected artifacts that occur in the analog section of the sound card itself or in the E-MU 1820. When the level of the output signal is lowered, the problematic frequency band gets smaller and the artifacts from the audio range move up to at least 100 kHz. These data explain the oddities of the Auzen Prelude such as its genre predilections and its fastidiousness about connecting cables. It doesn't like recordings with constituents higher than 19 kHz whereas most of today's sound editors do not care to filter their recordings before digitizing them. When reproducing such materials, this sound card has wide-spectrum and rather intensive noise in its output.

The rest of the sound cards are not sensitive to the ultrasound constituents, including the Auzen Bravura which has the same AK4396 DAC in its headphone output and whose line output filter is based on a CS4382A converter that has a similar switching response. By the way, the Cirrus Logic converter is better at suppressing aliasing than either of the Asahi Kasei converters.

And finally, I can show you the test results for the opamps you can buy from Auzentech.

As you can see, the Bravura’s headphone amplifier reacts eagerly to the change of opamps, the difference being obvious in the results. However, I won’t comment on them because I did not check the results out for repeatability.


Well, that was such an auspicious beginning: a sound card with high-quality components, an excellent headphone amplifier and an X-Fi audio processor for an affordable $130! But as it turns out, there are too many negative factors to all this.

The technological innovations have no practical value whereas many of the manufacturer's claims are misleading, to put it mildly. The Bravura does not have an X-Fi audio processor that we have come to expect from Auzentech sound cards. The chip employed instead it does not ensure a better signal separation as the manufacturer claims but worsens the signal-to-noise ratio when playing ordinary CDDA music so much that you can use the integrated amplifier with low-Ohm headphones only. If you do not switch the card's outputs to 44.1 kHz, it will produce too much distortion (even though rather agreeable to the ear) whereas the noise at this frequency prevented me from evaluating the headphone amplifier, which is improved over the one in the Auzen X-Fi HomeTheater, and the unique capability of switching the operation mode of the DAC’s digital filter. Hopefully, this will be corrected by driver updates because, for a short while, I did make the card reproduce 44.1kHz signal with no noise or distortion.

Software used to be a scourge of the Auzen Bravura, but it is being steadily improved. It is still awfully slow in Windows 7 x64 but its driver does not get confused between the two individual playback devices as the OS identifies the sound card. This separation into two devices in Windows 7 and Vista allows changing the line output and headphone volume independently from each other and I don’t know why the developers didn’t implement the same in Windows XP. After all, semiprofessional sound cards have long been able to do this trick. On a second thought, it would be better if this particular sound card lacked this feature even in Windows 7. With the headphones and speakers being separate, you cannot switch between them without stopping playback or even re-launching the application altogether. You cannot force the number of sound channels for the Headphone device, which has a negative effect on most of modern games. Interestingly, the Creative PCI Express X-Fi Xtreme Audio with the same audio controller does not split up into headphones and speakers after you install its driver.

Another feature that I would gladly disable is Jack Sensing. There is an additional contact in each of the sound card’s connectors that identifies whether there is a cable plugged into that connector, so that Windows could turn off devices when cables are disconnected. Some applications that were playing something via the disconnected device would halt with an error whereas the card’s own software falls into a long stupor during that operation.

In the Windows XP environment the Auzen Bravura is hardly more functional than an integrated sound core because it lacks support for 44.1 kHz sampling rate, ASIO and OpenAL. It’s better in Windows 7 and Vista where the card only lacks ASIO. However, the sampling rate of 44.1 kHz can only be used when you set the volume level above 40% whereas 3D sound is processed on the software level, even in OpenAL-based games. This doesn’t provoke a performance hit on modern and fast computers but the resulting quality of sound effects is hardly better than what you get with sound cards based on C-Media Oxygen chips with Dolby algorithms. It is a shame that the Bravura software does not offer such capabilities that we have come to expect even from mainboard-integrated audio.

The quality of manufacture is another issue to discuss. Out of the four tested sound cards from Auzentech the X-Fi Home Theater was the only one to work blamelessly in accordance with the manufacturer’s specs except that its headphone amplifier was not perfect with a load of 32 Ohm. As for the other products, the X-Fi Prelude had problems with the recording quality of the line input while the X-Fi Forte had some playback quality issues of an unclear nature and was also very hot at work, which may lead to a quick drying out of the electrolytic capacitors.

The Bravura has learned from its predecessors and its hottest capacitors have been replaced with solid-state ones of higher capacitance. Its improved headphone amplifier is now as good as external ones but its line output has some distortion that I did not notice with the previous products. For all the technological advances of Auzentech, the Bravura’s line output is occasionally inferior to the much cheaper Creative PCI Express X-Fi Xtreme Audio. The long “warm-up” period the Bravura takes to reveal its potential isn’t a good thing, either. If this is due to the characteristics of the electrolytic capacitors, they should have been “trained” back at the factory.

So, who might need such a sound card? First of all, it is going to be okay for people who use low-impedance headphones which do not sound good with other sound cards. The Bravura can outperform the ASUS Xonar Essence when it comes to headphones, let alone any sound card with no headphone amplifier. I have no doubt it will cope with any pair of headphones but audiophiles will have to use software resampling in WinAMP or foobar2000 players until the problem with noise in 44.1 kHz mode is solved. Besides, the multichannel output of the Auzen Bravura is high enough quality to avoid being a bottleneck in a home theater system. Finally, someone may want its advanced microphone-related capabilities although the lack of ASIO support makes this sound card unattractive for musicians.



I hope these lists of highs and lows will allow you make an informed buying decision.