With the original Live! drivers there should be an option in surround mixer for the speaker layout called Live! Surround. It's quite basic and uses something closer to dolby surround instead of HRTFs like later cards with CMSS-3D Virtual. I found it too annoying.

The setting is missing if you use the later Audigy based drivers.

SBLive has no HRTF capabilities AFAIK. It was only good at 2D positional audio if you ran a 4.0-5.1 speaker config. Vortex 2 was the headphone and 2 speaker 3D virtualization solution back then and later the cards using Sensaura were solid too. I don't remember if the old kx drivers tried to do something here but I have a feeling the EMU10K1 DSP isn't flexible enough.

Some games have software based spatial/HRTF options. They might do something with their own custom code or through libraries such as Miles Sound System and Directsound3D's software processing. Some games expose all the Miles Sound System audio backends for you to try.

AndreaColombo86 wrote on 2026-02-04, 11:37:

Hello, everyone.

I was wondering whether it is possible to get virtualized surround sound on headphones with an SB Live! 5.1 sound card.

Say the game has an option for 5.1 surround sound. I could select it, then set Windows to 5.1 speakers and play. The problem is the SB Live! software does not have an explicit option to enable CMSS-3D like later SB cards. Do you know whether it does surround virtualization?

Yes, it is possible, just like on any other card, using software utils like EqualizerAPO + HeSuVi (requires Windows 7+) for 5.1/7.1 to stereo (headphone) down-mixing.

CMSS-3D Headphone (or Virtual) also runs on the CPU, even on the "powerful" X-Fi cards.

Also you can install modified Audigy 2 driver on SB Live! and check if CMSS works, but I'm not sure if this will be same CMSS-3D generation, I think it appeared later with X-Fi series.

ott wrote on 2026-02-05, 03:32:

CMSS-3D Headphone (or Virtual) also runs on the CPU, even on the "powerful" X-Fi cards.

got any source for this? MIDI playback is one thing, but if even CMSS is done in software there is hardly a reason for that much advertised DSP to be on the card, is there?

there is the better SRC, but ironically the vista and later resampling means that the earlier cards are better off than under XP/9x, in terms of playing back 44.1 khz...

auron wrote on Today, 01:32:

got any source for this? MIDI playback is one thing, but if even CMSS is done in software there is hardly a reason for that much advertised DSP to be on the card, is there?

Haha, this question is also mentioned in my source 😀
I found this in old iXBT review (in Russian): the author compared performance of Audigy2 Value and X-Fi and saw CPU spikes when CMSS-3D was enabled on X-Fi:

When CMSS is enabled, X-Fi activates Sensaura algorithms, and as you can see from the CPU load, it's more software-based than hardware-based. Again, on powerful PCs, this will be 1-2%, but that begs the question: what's the point of a mega-powerful DSP with fifty million transistors clocked at 400 MHz? Essentially, the DSP handles SRC, mixing, and reverb.
https://web.archive.org/web/20061025022555/ht … gy2-value.shtml

It's also known that CMSS on X-Fi USB is entirely software-based.
Of course, it uses a different chip, but this likely speaks to CMSS being simplified and scaled through the software layer.

>>Does this mean thet the X-Fi,Crystalizer features of the USB X-Fi are implemented purely in software? I think this might be th […]
Show full quote

>>Does this mean thet the X-Fi,Crystalizer features of the USB X-Fi are implemented purely in software?
I think this might be the case.
I started SniffUSB under WindowsXP and tried changing the settings using
the audio console software provided by Creative.

When changing the Speaker selection (2.1/4.1/5.1), data is sent over usb to the device.
However, when I try to change the settings for the X-Fi Crystalizer or EAX or CMSS-3D or Graphic Equalizer
nothing gets sent to the device but the audio output does change and I am able to hear the corresponding effects.

And as mentioned in my previous post, the lsusb output doesn't show any effect descriptors or feature units

If this is in fact true that all the xtreme fidelity(X-Fi) stuff is happening in software alone, my purchase has been
a total waste!! Does anybody on this list have a USB X-Fi 5.1 that they can use to verify my observations?
https://mailman.alsa-project.org/pipermail/al … ary/014533.html

ott wrote on Today, 05:36:

Haha, this question is also mentioned in my source 😀
I found this in old iXBT review (in Russian): the author compared performance of Audigy2 Value and X-Fi and saw CPU spikes when CMSS-3D was enabled on X-Fi:

Just a quick note here, X-Fi CMSS-3D is more advanced than the version which Audigy cards use. At the very least, the X-Fi iteration has Elevation Filter and MacroFX, while those features don't exist on Audigy cards. It's possible that there are other differences as well.

So I'm not sure if a direct comparison between two different cards is actually a good idea. To test if CPU utilization is higher, it might be better to simply run a few tests on an X-Fi card, with CMSS-3D turned on/off from the Creative Console Launcher, and then compare the results. Ideally, these would be 1:1 tests (e.g. a built-in game benchmark) where the only difference is the presence/absence of CMSS-3D.

Joseph_Joestar wrote on Today, 08:31:

So I'm not sure if a direct comparison between two different cards is actually a good idea. To test if CPU utilization is higher, it might be better to simply run a few tests on an X-Fi card, with CMSS-3D turned on/off from the Creative Console Launcher, and then compare the results. Ideally, these would be 1:1 tests (e.g. a built-in game benchmark) where the only difference is the presence/absence of CMSS-3D.

I'm not super familiar with how CMSS-3D works, but if it takes in a multi-channel mix and then virtualises that, you also need to compare the multi-channel mix and CMSS-3D rather than 2.0 and CMSS-3D since there will be some additional overhead in preparing the multi-channel mix

NeoG_ wrote on Today, 08:43:

I'm not super familiar with how CMSS-3D works, but if it takes in a multi-channel mix and then virtualises that, you also need to compare the multi-channel mix and CMSS-3D rather than 2.0 and CMSS-3D since there will be some additional overhead in preparing the multi-channel mix

Good idea. I think it would also be advisable to run the test on a single-core CPU, just to exclude the possibility of the workload being handled by one of the other cores. So ideally, you'd have things set up like this:

• Test 1: Game set to stereo (if available), Windows/Creative speakers set to stereo, CMSS off
• Test 2: Game set to stereo (if available), Windows/Creative speakers set to stereo, CMSS on
• Test 3: Game set to stereo (if available), Windows/Creative speakers set to Headphone, CMSS-3D Headphone on

Comparing tests 1 and 2 would give you an understanding whether basic CMSS uses extra CPU cycles. While comparing tests 1 and 3 would determine the same for the CMSS-3D Headphone implementation, with its more advanced features. Of course, some people like to set the game and Windows speakers to 7.1 while keeping the Creative Console Launcher in Headphone mode. That would require additional testing, since the game would likely use more CPU power simply by processing 7.1 surround sound, regardless of whether CMSS is turned on/off.