Sound Device Considerations for Gamers

[Balthazor]

This post is primarily for gamers; if you watch movies or listen to music much more than game on your PC, then you can skip it. Gamers, read on:

If you game, and you have games that use EAX or plan to play games with EAX, then EAX will be your driving decision maker. EAX is usable on Vista over OpenAL or over the older DirectSound3D API using a .dll that changes DirectSound3D calls to OpenAL calls. EAX is still promoted by Creative by developers and is still being implemented on current games.

You can only access EAX through internal sound cards. True EAX support is available via X-Fi cards, including the Auzentech Prelude (and to a limited extent the older Audigy cards) and emulated EAX support is available through the Asus Xonar series of cards.
Without cards with EAX support you won't hear certain effects in games. You will still get multi-channel sound, but just won't hear all of the effects that developers can easily add using EAX. EAX allows developers to add reverb effects, making audio sound different depending on where the audio takes place in the virtual game world, or muffling sounds if an intervening virtual object is between 'you' and the sound in question.

I personally find playing most games with EAX and without EAX to be a night and day difference; but test for yourself.

Not all games use EAX. If you happen to only play a few games, and these don't use EAX, then I would definitely recommend that you consider one of the non-EAX sound options, whether an internal sound card, your onboard sound's digital out & a nice receiver, or external USB sound device.

Hardware acceleration used to be a big deal, but multi-core CPUs are handling audio processing much better than their single-core predecessors. Although hardware acceleration of audio might save you as much as 5% of CPU utilization, the actual FPS impact of software processing of audio is very minimal when you consider that the CPU is only one factor of your resulting gaming FPS. In games that are largely CPU-limited you may notice a larger difference.

If you are using an older computer with a single core, then I would recommend you upgrade your CPU; you'll get much more bang for your buck that way. If you are dead set on getting a sound card with hardware acceleration then you're only option would be the X-Fi or Audigy equipped cards, from Creative or Auzentech.

Hardware acceleration of sound in games is only available if the game supports OpenAL or the older DirectSound3D API. For every game that uses these there is one or two that use a software-only audio processing engine, and no sound card will be able to hardware process a game using this type of audio.

For several months I've been reading posts on this forum recommending external USB DACs or external sound cards / sound devices as an alternative to internal sound cards for gaming.

As I understand it the primary advantage to doing this is do the digital to analog conversion in a less electronically noisy environment; outside the computer case is clearly better than inside the computer case.

But USB has its own problems. Everything you put over USB is going to eat up CPU processing cycles. Even changing the report rate of USB mice can have a dramatic effect on your FPS in certain games.

I do not own a USB sound device so can't test myself, but I did do a brief bit of research that suggests that there may be a small but negative impact on the FPS of games when using a USB sound device vs. an internal sound card. These comparisons are both using audio processed on the CPU, not 'hardware accelerated', so they seem to be a good apples to apples comparison of the two.

Ascully.com compared the USB Philips Aurillium to a PCI Turtle Beach Santa Cruz (in 2003):

USB Phillips / PCI Turtle Beach:
in Quake 3: 153 / 170
in Comanche 4: 38 / 46

This is an older review, and probably using a single-core machine. What about something more recent? There aren't a lot of reviews showing the FPS impact of an external sound device, but I did find this one from January 2008 from PC Perspective. Please note I do not know what product they are referring to as the 'X-Fi Premium' so am just leaving that out.

USB Asus U1 / PCI Xonar D2
in Crysis: 18 / 19

Unfortunately the review doesn't seem to note what sort of CPU is running this game; the difference is small, so is that due to a dual or quad core processor at work?

Guru3D has this review comparing Creative's PCI-E eXtremeAudio to a USB Sonar headset (May 2008):

USB Sonar / PCI-E eXtremeAudio
in Call of Duty 4: 53 / 56

This above test was using a dual core computer. I ommitted the Bioshock results because the reviewer didn't know whether or not EAX was enabled with the eXtremeAudio, which wouldn't have made for a good comparison if it were. EAX on the eXtremeAudio is processed in software and would have altered the overall results.

And that is all I could find without becoming obsessive. I was actually expecting a bigger difference in overall FPS between USB sound devices and internal sound devices, but the evidence appears to suggest that as computers increase the number of cores the effect of putting your sound data over USB has less of a real impact on game performance.

There other factors to consider when purchasing a sound card or sound device for gaming; inputs and outputs, quality of DACs (if you are using analog out), multi-channel encoding using DDL or DTS (if you are using digital out), and quality of the microphone input (if you are one who likes to use voice chat while gaming.) You can find a variety of USB sound devices and internal sound cards that offer you a wide field of these options.

Happy gaming!

 

[ashmedai]

Thanks for the great post! You're right, there does seem to be a lot of confusion over the "sound card" issue - in no small part due to the fast one Microsoft pulled with Vista. However, it is a trend that will hopefully eliminate the sort of sound technology compatibility issues we had even before Vista, and it's one that modern computing power can hopefully make the most of.

I would emphasize that EAX only matters if the game developers decide to make some effects only available over EAX. The entire point of interface layers like DirectX is that it allows developers to create a game that runs on as many hardware combinations as possible with as many of its "bells and whistles" as possible. If you don't see the shiny audio effects that they spent many large piles of money and many hard hours working on...well, that is absolutely the opposite of what game devs want! Since processors are so powerful compared to the earlier days of computer games, it's very easy to offload many complicated audio tasks to the CPU that would have been impossible before, and more games are taking this route because it makes it easier to get the complete game to as many gamers as possible.

With Vista & ALchemy, you're already running the sound over a virtualization layer...so there's no real penalty to using the implementation of this virtualization for the Xonar cards instead. The Xonars use very respectable DAC chips. As far as sound cards go, the Xonar should sound better than most. If you're not already stuck with the Vista problem, well, given a decent multicore CPU, it's not really a huge tradeoff - a couple % of CPU usage in exchange for a good DAC.

And the few % CPU cycles to use USB - unless you have a ridiculously crappy USB host implementation - weren't you going to be using those (and more) for ALchemy already? Sound costs CPU cycles no matter which way you slice it. And anyway...it's not like it uses your GPU to do it.

 

[Balthazor]

I suspect that developers are given incentives by Creative to add EAX effects to their games. It would make sense.

Of course in a perfect world EAX would be open to be implemented in software by third parties so you could process the effects on another sound card or on your CPU, but that isn't the case. And since EAX effects do sound great on the games that use them, whether or not you purchase a card capable of supporting EAX (or at least emulating it, as the Xonar cards do) is going to be a significant factor for gamers for the time being.

But you should always research your games and see just how many of them actually use EAX before making EAX a factor.

And keep in mind that older implementations of EAX (2.0 and earlier) are 'open' and supported by most major sound cards. It is only the 3.0 and later versions (current iteration is 5.0) that are limited to X-Fi hardware. Audigy 2 can do 3.0 and 4.0; Audigy is limited to 3.0.

And from what I understand EAX is not backwards compatabile, so an EAX 5.0 game will not run on Audigy hardware. Some games with EAX have an option to change to a lower version of EAX.

This list is the official list of EAX games from Creative (click on the EAX Games List), but it doesn't note what versions of EAX are used. Generally speaking games from 2002 or earlier have a good chance of being EAX 2.0; after that it gets a little fuzzy.

 

[topdog5050]

Hey,I think the Asus Xonor are the Creative Xi-Fi would be great to have with a set of Logitech great combo.