And here's Ghost Recon: Advanced Warfighter 2. This is with the "Extreme" audio quality preset, which is exclusive to X-Fi cards, as shown here.

In this screenshot, taken during the very first mission, it's using 127-2=125 simultaneous hardware voices! This is the most I've seen in any game that I've tested so far. And all that with just 7ms of latency. Incredible!

Just to confirm, you are using the Daniel K driver pack for Windows 7?

DrLucienSanchez wrote on 2026-03-08, 14:13:

Just to confirm, you are using the Daniel K driver pack for Windows 7?

Yes.

I'm using DanielK's latest X-Fi support pack under both Win7 and Win10. The only place where I don't use it is under WinXP.

I also wanted to check if games which utilize EAX via DirectSound3D could be tested when using Creative's ALchemy under Win7, and it looks like they can! Here's a screenshot from Halo: Combat Evolved, a game which is officially supported by ALchemy. For the record, this is the retail CD version + latest official patch. No mods or fan made fixes of any kind have been applied.

It's using 127-65=62 simultaneous hardware voices, so pretty much as expected of an EAX 3.0 game. Nice that Falcosoft's OpenAL Test tool can monitor this as well.

Joseph_Joestar wrote on 2026-03-08, 18:02:

I also wanted to check if games which utilize EAX via DirectSound3D could be tested when using Creative's ALchemy under Win7, and it looks like they can! Here's a screenshot from Halo: Combat Evolved, a game which is officially supported by ALchemy. For the record, this is the retail CD version + latest official patch. No mods or fan made fixes of any kind have been applied.

The attachment HaloCE_Test01.jpg is no longer available

It's using 127-65=62 simultaneous hardware voices, so pretty much as expected of an EAX 3.0 game. Nice that Falcosoft's OpenAL Test tool can monitor this as well.

Hi,
1. I have noticed on your screenshots that on your X-Fi the busy loop takes more time than on my Audigy 1 and Aidigy 2 ZS. So I have made some optimizations.
Namely instead of trying to allocate buffers/sources one by one version 1.5 uses the 'binary search' pattern so first it tries to allocate 128 buffers then 64, 32, 16, 8, 4, 2, 1. This way instead of 127 calls (as the worst case on X-Fi) it uses only maximum 8 calls. On Audigy 1 and Aidigy 2 ZS this new algorithm is about 1 ms faster but on X-Fi maybe it matters more.
The best way to test this is to start monitoring with the previous version first without running other OpenAL programs and notice what the 'Query Time' field shows.
Then close the previous version, start the new one and make the same monitoring test.

2.Also added the ability to manually select the EAX reverb preset during the live audio playback test.

Falcosoft wrote on 2026-03-08, 19:56:

The best way to test this is to start monitoring with the previous version first without running other OpenAL programs and notice what the 'Query Time' field shows.
Then close the previous version, start the new one and make the same monitoring test.

Cheers! I just tested it, and the difference it pretty big on my X-Fi Titanium. It went from 27ms in version 1.4 to only 3ms in version 1.5. Nice work!

I might have missed this—have you tried to check how many hardware voices are used in UT3 and F.E.A.R.? Would be nice to know for Mass Effect and Gears of War as well, if you have them handy 😀

AndreaColombo86 wrote on Yesterday, 22:46:

I might have missed this—have you tried to check how many hardware voices are used in UT3 and F.E.A.R.? Would be nice to know for Mass Effect and Gears of War as well, if you have them handy 😀

UT3 has already been tested here.

I do have F.E.A.R. and it is officially supported by ALchemy, so I will test that when time permits. But I don't have Mass Effect and Gears of War on PC, only on my Xbox 360. If someone else wants to test those, feel free to post the results here.

Ok, here's the F.E.A.R. test:

In this screenshot, taken during a firefight with multiple opponents, it's using 127-55=72 simultaneous hardware voices.

P.S.

Creative's ALchemy doesn't detect the GOG version of F.E.A.R. I've looked into this, and it seems to happen because GOG doesn't set the relevant registry keys that ALchemy is looking for. The retail DVD version of the game does of course set those keys, and is detected just fine. I've seen this behavior with some other GOG releases such as Splinter Cell.

Joseph_Joestar wrote on Today, 10:07:

Ok, here's the F.E.A.R. test: […]
Show full quote

Ok, here's the F.E.A.R. test:

The attachment FEAR_Test01.jpg is no longer available

In this screenshot, taken during a firefight with multiple opponents, it's using 127-55=72 simultaneous hardware voices.

P.S.

Creative's ALchemy doesn't detect the GOG version of F.E.A.R. I've looked into this, and it seems to happen because GOG doesn't set the relevant registry keys that ALchemy is looking for. The retail DVD version of the game does of course set those keys, and is detected fine. I've seen this behavior with some other GOG releases such as Splinter Cell.

It's interesting that in case of some games the X-RAM on the X-Fi is not used at all despite the documenattion says that X-RAM usage is automatic even without explicit support...
UT3 was another game that did not utilize X-RAM at all if I remember correctly.

It's interesting because X-RAM usage can be disabled but it requires explicit X-RAM related OpenAL function calls. So a game must know the X-RAM extension to be able to disable it.
So implicitly such games should be compatible with EAX 5 level. I do not know if it's true for UT3 and FEAR.

Falcosoft wrote on Today, 10:36:

It's interesting that in case of some games the X-RAM on the X-Fi is not used at all despite the documenattion says that X-RAM usage is automatic even without explicit support...
UT3 was another game that did not utilize X-RAM at all if I remember correctly.

Well, F.E.A.R. isn't a native OpenAL game. I'm not sure if X-RAM can be accessed via DirectSound3D.

BTW, if you look into ALchemy.ini you can see how it detects all the supported games. Specifically, for F.E.A.R. it's looking for the following registry entries:

1[F.E.A.R]
2RegPath=HKEY_LOCAL_MACHINE\SOFTWARE\Monolith Productions\FEAR\1.00.0000\InstallDir
3Preset=1
4
5[F.E.A.R. Extraction Point]
6RegPath=HKEY_LOCAL_MACHINE\SOFTWARE\TimeGate Studios\FEARXP\1.00.0000\InstallDir
7Preset=1
8
9[F.E.A.R. Perseus Mandate]
10RegPath=HKEY_LOCAL_MACHINE\SOFTWARE\TimeGate Studios\FEARXP2\1.00.0000\InstallDir
11Preset=1

If those aren't present, it won't find the game. And the GOG release doesn't set any of them. Which is why I had to use the retail DVD version for the test above.

Joseph_Joestar wrote on Today, 10:43:

Well, F.E.A.R. isn't a native OpenAL game. I'm not sure if X-RAM can be accessed via DirectSound3D. […]
Show full quote

Falcosoft wrote on Today, 10:36:

It's interesting that in case of some games the X-RAM on the X-Fi is not used at all despite the documenattion says that X-RAM usage is automatic even without explicit support...
UT3 was another game that did not utilize X-RAM at all if I remember correctly.

Well, F.E.A.R. isn't a native OpenAL game. I'm not sure if X-RAM can be accessed via DirectSound3D.

BTW, if you look into ALchemy.ini you can see how it detects all the supported games. Specifically, for F.E.A.R. it's looking for the following registry entries:

Copy code to clipboard

1[F.E.A.R]
2RegPath=HKEY_LOCAL_MACHINE\SOFTWARE\Monolith Productions\FEAR\1.00.0000\InstallDir
3Preset=1
4
5[F.E.A.R. Extraction Point]
6RegPath=HKEY_LOCAL_MACHINE\SOFTWARE\TimeGate Studios\FEARXP\1.00.0000\InstallDir
7Preset=1
8
9[F.E.A.R. Perseus Mandate]
10RegPath=HKEY_LOCAL_MACHINE\SOFTWARE\TimeGate Studios\FEARXP2\1.00.0000\InstallDir
11Preset=1

If those aren't present, it won't find the game. And the GOG release doesn't set any of them. Which is why I had to use the retail DVD version for the test above.

1Well, F.E.A.R. isn't a native OpenAL game. I'm not sure if X-RAM can be accessed via DirectSound3D.

1. The point of Alchemy is that the game does not have to access the X-RAM through DirectSound3D. The wrapper translates calls to native OpenAL calls so OpenAL should be able to use its automatic X-RAM allocation algorithm.
But of course it's possible that the wrapper itself calls functions to disable X-RAM. Although this still cannot explain the case of UT3 that is a native OpenAL game.

PS: we have proved here that X-RAM usage does not require explicit OpenAL function calls, it works automatically by default:
Re: EAX appreciation thread
According to this result X-RAM is an opt-out feature that is enabled by default and you have to enable manual management mode and disable the usage of X-RAM for samples explicitly if you do not want to use it.
Of course according to the documentation there are special cases when the driver decides to put samples into system memory instead of X-RAM according to some defined rules.

2. BTW, for me Alchemy works even with unsupported apps/games if Alchemy's dsound.dll and dsound.ini are copied into the app/game's folder. Actually my version of Alchemy does not do anything else even with supported games than copying the dsound.dll/dsound.ini files into the games folder.

Falcosoft wrote on Today, 10:52:

1. The point of Alchemy is that the game does not have to access the X-RAM through DirectSound3D. The wrapper translates calls to native OpenAL calls so OpenAL should be able to use its automatic X-RAM allocation algorithm.
But of course it's possible that the wrapper itself calls functions to disable X-RAM. Although this still cannot explain the case of UT3 that is a native OpenAL game.

Interesting. From the games that I've tested so far, I haven't seen X-RAM get used through ALchemy. But maybe there are some edge cases out there that can use it that way. Not sure.

Falcosoft wrote on Today, 10:52:

2. BTW, for me Alchemy works even with unsupported apps/games if Alchemy's dsound.dll and dsound.ini are copied into the app/game's folder.

Yup, there's even a button to add unsupported games right in the ALchemy window. But for the purposes of determining the number of used hardware voices, I deliberately chose to only test officially supported games. Just to ensure that everything is working exactly as Creative intended.