VOGONS


EAX appreciation thread

Topic actions

Reply 900 of 941, by Joseph_Joestar

User metadata
Rank l33t++
Rank
l33t++
shevalier wrote on 2026-01-15, 07:00:

yep...

Like I said, on my system AIDA64 always shows 127. Even if I'm not running any games and just sitting on the desktop.

The attachment AIDA64_Trial.jpg is no longer available

I even tried using an older version like the one in your screenshot, and Creative's official drivers instead of DanielK's, but it made no difference. Unless that capability isn't available in the Trial version of AIDA64 which I'm using.

PC#1: Pentium MMX 166 / Soyo SY-5BT / S3 Trio64V+ / Voodoo1 / YMF719 / AWE64 Gold / SC-155
PC#2: AthlonXP 2100+ / ECS K7VTA3 / Voodoo3 / Audigy2 / Vortex2
PC#3: Core 2 Duo E8600 / Foxconn P35AX-S / X800 / Audigy2 ZS
PC#4: i5-3570K / MSI Z77A-G43 / GTX 980Ti / X-Fi Titanium

Reply 901 of 941, by shevalier

User metadata
Rank Oldbie
Rank
Oldbie

1. The game is not in DirectX/OpenGL output mode.
2. Aida64 displays data at the moment of its launch. To update the data, click the ‘Update Data’ icon. (circle with an arrow).

Aopen MX3S, PIII-S Tualatin 1133, Radeon 9800Pro@XT BIOS, Audigy 4 SB0610
JetWay K8T8AS, Athlon DH-E6 3000+, Radeon HD2600Pro AGP, Audigy 2 Value SB0400
Gigabyte Ga-k8n51gmf, Turion64 ML-30@2.2GHz , Radeon X800GTO PL16, Diamond monster sound MX300

Reply 902 of 941, by Joseph_Joestar

User metadata
Rank l33t++
Rank
l33t++
shevalier wrote on 2026-01-15, 10:04:

2. Aida64 displays data at the moment of its launch. To update the data, click the ‘Update Data’ icon. (circle with an arrow).

That solved it. What I did was always start AIDA64 before Prey. But this time, I did the reverse, and then it does work. Anyway, here's Prey using 121 hardware voices per AIDA64 during the final boss battle at the end of the game:

The attachment Prey_121.jpg is no longer available

I've attached the relevant savegame if someone else wants to test this.

The attachment Prey_Finale.zip is no longer available

PC#1: Pentium MMX 166 / Soyo SY-5BT / S3 Trio64V+ / Voodoo1 / YMF719 / AWE64 Gold / SC-155
PC#2: AthlonXP 2100+ / ECS K7VTA3 / Voodoo3 / Audigy2 / Vortex2
PC#3: Core 2 Duo E8600 / Foxconn P35AX-S / X800 / Audigy2 ZS
PC#4: i5-3570K / MSI Z77A-G43 / GTX 980Ti / X-Fi Titanium

Reply 903 of 941, by Falcosoft

User metadata
Rank l33t
Rank
l33t
Joseph_Joestar wrote on 2026-01-15, 11:14:
That solved it. What I did was always start AIDA64 before Prey. But this time, I did the reverse, and then it does work. Anyway, […]
Show full quote
shevalier wrote on 2026-01-15, 10:04:

2. Aida64 displays data at the moment of its launch. To update the data, click the ‘Update Data’ icon. (circle with an arrow).

That solved it. What I did was always start AIDA64 before Prey. But this time, I did the reverse, and then it does work. Anyway, here's Prey using 121 hardware voices per AIDA64 during the final boss battle at the end of the game:

The attachment Prey_121.jpg is no longer available

I can provide a savegame if someone else wants to test that.

Correct me if I'm wrong but for me AIDA64 -> OpenAL -> Hardware Sound Buffers always shows the currently available/remaining and not the actually used buffers. So if I'm starting AIDA64 and no OpenAL software is running it shows 63 buffers (I'm using Audigy). When an OpenAL software is running it shows less than 63. According to the same logic on your screenshot we see that 121 free buffers are available so 6 buffers are used (127 - 121).

Last edited by Falcosoft on 2026-01-15, 11:32. Edited 1 time in total.

Website, Youtube
Falcosoft Soundfont Midi Player + Munt VSTi + BassMidi VSTi
VST Midi Driver Midi Mapper
x86 microarchitecture benchmark (MandelX)

Reply 904 of 941, by Joseph_Joestar

User metadata
Rank l33t++
Rank
l33t++
Falcosoft wrote on 2026-01-15, 11:27:

Correct me if I'm wrong but for me AIDA64 -> OpenAL -> Hardware Sound Buffers always shows the remaining and not the actually used buffers. So if I'm starting AIDA64 and no OpenAL software is running it shows 63 buffers (I'm using Audigy). When an OpenAL software is running it shows less than 63. According to the same logic on your screenshot we see that 121 free buffers are available so 6 buffers are used (127 - 121).

No idea, this is the first time I'm using that tool.

Maybe @shevalier can explain how it works.

PC#1: Pentium MMX 166 / Soyo SY-5BT / S3 Trio64V+ / Voodoo1 / YMF719 / AWE64 Gold / SC-155
PC#2: AthlonXP 2100+ / ECS K7VTA3 / Voodoo3 / Audigy2 / Vortex2
PC#3: Core 2 Duo E8600 / Foxconn P35AX-S / X800 / Audigy2 ZS
PC#4: i5-3570K / MSI Z77A-G43 / GTX 980Ti / X-Fi Titanium

Reply 905 of 941, by Falcosoft

User metadata
Rank l33t
Rank
l33t
Joseph_Joestar wrote on 2026-01-15, 11:30:
Falcosoft wrote on 2026-01-15, 11:27:

Correct me if I'm wrong but for me AIDA64 -> OpenAL -> Hardware Sound Buffers always shows the remaining and not the actually used buffers. So if I'm starting AIDA64 and no OpenAL software is running it shows 63 buffers (I'm using Audigy). When an OpenAL software is running it shows less than 63. According to the same logic on your screenshot we see that 121 free buffers are available so 6 buffers are used (127 - 121).

No idea, this is the first time I'm using that tool.

Maybe @shevalier can explain how it works.

The 'Correct me if I'm wrong' part was only rhetoric 😀 . I'm actually pretty sure it is working this way. If I start 2 instances of my test software that use 1 test sound and buffer per instance both AIDA64 and Rightmark3DSound show 61 available buffers that means 2 buffers are used (63 - 61)

The attachment openalbuffers.png is no longer available

When no OpenAL software is running both software show the maximum available hardware buffers on Audigy that is 63.

The attachment openalbuffers_nouse.png is no longer available

BTW, I think the overall available hardware buffers for OpenAL is one less than provided by the hardware because 1 primary hardware buffer is always reserved for the Windows audio stack and OpenAL has no exclusive mode (Other audio software can work even if an OpenAL application is running) .

Last edited by Falcosoft on 2026-01-15, 11:55. Edited 1 time in total.

Website, Youtube
Falcosoft Soundfont Midi Player + Munt VSTi + BassMidi VSTi
VST Midi Driver Midi Mapper
x86 microarchitecture benchmark (MandelX)

Reply 906 of 941, by Joseph_Joestar

User metadata
Rank l33t++
Rank
l33t++
Falcosoft wrote on 2026-01-15, 11:39:

The 'Correct me if I'm wrong' part was only rhetoric 😀 . I'm actually pretty sure it is working this way. If I start 2 instances of my test software that use 1 test sound and buffer per instance both AIDA64 and Rightmark3DSound show 61 available buffers that means 2 buffers are used (63 - 61)

I haven't been able to find any documentation on that aside from this page which doesn't mention the Hardware Sound Buffers field at all. But what you say does make sense to me.

My problem with AIDA64 is that it doesn't seem to update in real time, even if I have the game and the tool open simultaneously. Instead, I have to click its refresh button to get a new reading. And yeah, I got different readings while loading different saves, possibly depending on how the map was designed and how many enemies were on the screen.

For a proper test, we would need something akin to MSI Afterburner which can display an in-game overlay. But instead of showing FPS, it would display free/used hardware buffers and X-RAM in real time.

PC#1: Pentium MMX 166 / Soyo SY-5BT / S3 Trio64V+ / Voodoo1 / YMF719 / AWE64 Gold / SC-155
PC#2: AthlonXP 2100+ / ECS K7VTA3 / Voodoo3 / Audigy2 / Vortex2
PC#3: Core 2 Duo E8600 / Foxconn P35AX-S / X800 / Audigy2 ZS
PC#4: i5-3570K / MSI Z77A-G43 / GTX 980Ti / X-Fi Titanium

Reply 907 of 941, by shevalier

User metadata
Rank Oldbie
Rank
Oldbie
Falcosoft wrote on 2026-01-15, 11:27:
Joseph_Joestar wrote on 2026-01-15, 11:14:
That solved it. What I did was always start AIDA64 before Prey. But this time, I did the reverse, and then it does work. Anyway, […]
Show full quote
shevalier wrote on 2026-01-15, 10:04:

2. Aida64 displays data at the moment of its launch. To update the data, click the ‘Update Data’ icon. (circle with an arrow).

That solved it. What I did was always start AIDA64 before Prey. But this time, I did the reverse, and then it does work. Anyway, here's Prey using 121 hardware voices per AIDA64 during the final boss battle at the end of the game:

The attachment Prey_121.jpg is no longer available

I can provide a savegame if someone else wants to test that.

Correct me if I'm wrong but for me AIDA64 -> OpenAL -> Hardware Sound Buffers always shows the currently available/remaining and not the actually used buffers. So if I'm starting AIDA64 and no OpenAL software is running it shows 63 buffers (I'm using Audigy). When an OpenAL software is running it shows less than 63. According to the same logic on your screenshot we see that 121 free buffers are available so 6 buffers are used (127 - 121).

Yes, that's right. These are the display features of AIDA64 — this programme simply reads the current values of available buffers without analysing how much the hardware can provide in total.
But this is much better than having no monitoring utilities at all.

Joseph_Joestar wrote on 2026-01-15, 11:14:

Anyway, here's Prey using 121 127-121=6 hardware voices per AIDA64 during the final boss battle at the end of the game:

This will provide a more complete description of your screenshot.

Aopen MX3S, PIII-S Tualatin 1133, Radeon 9800Pro@XT BIOS, Audigy 4 SB0610
JetWay K8T8AS, Athlon DH-E6 3000+, Radeon HD2600Pro AGP, Audigy 2 Value SB0400
Gigabyte Ga-k8n51gmf, Turion64 ML-30@2.2GHz , Radeon X800GTO PL16, Diamond monster sound MX300

Reply 908 of 941, by Joseph_Joestar

User metadata
Rank l33t++
Rank
l33t++
shevalier wrote on 2026-01-15, 12:14:

This will provide a more complete description of your screenshot.

So in your screenshot it's using 127-69=58 hardware buffers? Like I said, it may vary based on which map is loaded, how many enemies are on the screen, and how the soundstage of the area was designed.

Without a real time monitoring tool, we can't exclude the possibility that the game uses up to the advertised 124 hardware voices simultaneously under certain circumstances. I'm guessing that might happen during heavy firefights, or possibly in multiplayer matches. But again, this is just me speculating. Without real time monitoring, there's no way to be sure.

PC#1: Pentium MMX 166 / Soyo SY-5BT / S3 Trio64V+ / Voodoo1 / YMF719 / AWE64 Gold / SC-155
PC#2: AthlonXP 2100+ / ECS K7VTA3 / Voodoo3 / Audigy2 / Vortex2
PC#3: Core 2 Duo E8600 / Foxconn P35AX-S / X800 / Audigy2 ZS
PC#4: i5-3570K / MSI Z77A-G43 / GTX 980Ti / X-Fi Titanium

Reply 909 of 941, by shevalier

User metadata
Rank Oldbie
Rank
Oldbie
Joseph_Joestar wrote on 2026-01-15, 12:19:
shevalier wrote on 2026-01-15, 12:14:

This will provide a more complete description of your screenshot.

So in your screenshot it's using 127-69=58 hardware buffers? Like I said, it may vary based on which map is loaded, how many enemies are on the screen, and how the soundstage of the area was designed.

Without a real time monitoring tool, we can't exclude the possibility that the game uses up to the advertised 124 hardware voices simultaneously under certain circumstances. I'm guessing that might happen during heavy firefights, or possibly in multiplayer matches. But again, this is just me speculating. Without real time monitoring, there's no way to be sure.

I don't remember where I read that OpenAL does not have a built-in garbage collector for unused buffers.
That is, they must first be manually initialised, used, and then closed.
The number obtained from AIDA is the number of pre-initialised buffers linked to sources.
Whether data is flowing into them or they have simply not been closed is something only the game engine developers can say.

https://www.openal.org/documentation/OpenAL_P … mmers_Guide.pdf

Aopen MX3S, PIII-S Tualatin 1133, Radeon 9800Pro@XT BIOS, Audigy 4 SB0610
JetWay K8T8AS, Athlon DH-E6 3000+, Radeon HD2600Pro AGP, Audigy 2 Value SB0400
Gigabyte Ga-k8n51gmf, Turion64 ML-30@2.2GHz , Radeon X800GTO PL16, Diamond monster sound MX300

Reply 910 of 941, by MattRocks

User metadata
Rank Member
Rank
Member
shevalier wrote on 2026-01-15, 12:32:
I don't remember where I read that OpenAL does not have a built-in garbage collector for unused buffers. That is, they must firs […]
Show full quote
Joseph_Joestar wrote on 2026-01-15, 12:19:
shevalier wrote on 2026-01-15, 12:14:

This will provide a more complete description of your screenshot.

So in your screenshot it's using 127-69=58 hardware buffers? Like I said, it may vary based on which map is loaded, how many enemies are on the screen, and how the soundstage of the area was designed.

Without a real time monitoring tool, we can't exclude the possibility that the game uses up to the advertised 124 hardware voices simultaneously under certain circumstances. I'm guessing that might happen during heavy firefights, or possibly in multiplayer matches. But again, this is just me speculating. Without real time monitoring, there's no way to be sure.

I don't remember where I read that OpenAL does not have a built-in garbage collector for unused buffers.
That is, they must first be manually initialised, used, and then closed.
The number obtained from AIDA is the number of pre-initialised buffers linked to sources.
Whether data is flowing into them or they have simply not been closed is something only the game engine developers can say.

https://www.openal.org/documentation/OpenAL_P … mmers_Guide.pdf

There is no GC. If the buffer state goes bad, it stays bad - not a real world problem because audio programmers are very specific.

The OpenAL approach is the same on all systems, but the systems behave differently:

  • XP initialises the endpoints once so the audio paths persist - it was possible to programme a DSP, download microcode, set states, etc.
  • Vista clobbers and reinitialises the endpoints periodically - this is what makes hardware offloading on Vista impossibly hard, and why Creative stopped selling hardware DSPs for games.

This is why ct_oal.dll and soft_oal.dll only tell you how a pathway started (same on both systems), not how it ended (different on both systems).

If Creative solved the Vista puzzle then why did Creative abandon their solution, and why has nobody else solved it since? And, why did so many people complain that Creative had not solved it?

What all these threads show is people blaming Creative. Only a few on Vogons thought to blame Microsoft!

Creative's big mistake was to try to work with the impossible. They should have instead released a Gold Final Super Pack for WinXP as a send-off, and then continued on Linux. They would have become smaller, but never as small as Vista+ made them.

Sadly, due to having only Vista+ drivers, my X-Fi Titanium HD THX is a beautiful lemon - an anomaly that existed only for the small number of users who didn't see Microsoft's trap.

Reply 911 of 941, by shevalier

User metadata
Rank Oldbie
Rank
Oldbie

Creative Labs dug its own grave when it kept EAX 3.0 for itself.
When Microsoft grew tired of solving DirectSound problems with a third-party API, Creative Labs was left alone.

PS If you're talking about Titanium PCI-e HD, even now it's a decent sound card for Windows 11.
I like CMSS-3D in films and games.
The 5.1 conversion to headphones with HRTF works great.
Although the sound of the unmodified card is, alas, rather dull.

PPS
I really did get a couple of decibels compared to Creative's lacklustre implementation.
The DACs from TI are simply superb.
For example, the distortion of the PCM4220 ADC is highly dependent on the capacitance of the capacitor in the high-frequency input filter.
Or interference introduced by the charge pump depending on the ESR of the flying capacitor?
But who cares?
Whoever shows the best overall result for this PCB, I am ready to admit defeat.
As an analogue circuit design engineer.)))

Aopen MX3S, PIII-S Tualatin 1133, Radeon 9800Pro@XT BIOS, Audigy 4 SB0610
JetWay K8T8AS, Athlon DH-E6 3000+, Radeon HD2600Pro AGP, Audigy 2 Value SB0400
Gigabyte Ga-k8n51gmf, Turion64 ML-30@2.2GHz , Radeon X800GTO PL16, Diamond monster sound MX300

Reply 912 of 941, by shevalier

User metadata
Rank Oldbie
Rank
Oldbie

By the way, here's a HINT.
If you take a film capacitor in your hands and touch each of its terminals with an oscilloscope, you can determine which external contact should be connected to the lower impedance part of the circuit, i.e. to ground.
This way, you can get a few part of decibels of noise. 😀

Aopen MX3S, PIII-S Tualatin 1133, Radeon 9800Pro@XT BIOS, Audigy 4 SB0610
JetWay K8T8AS, Athlon DH-E6 3000+, Radeon HD2600Pro AGP, Audigy 2 Value SB0400
Gigabyte Ga-k8n51gmf, Turion64 ML-30@2.2GHz , Radeon X800GTO PL16, Diamond monster sound MX300

Reply 913 of 941, by MattRocks

User metadata
Rank Member
Rank
Member

Yeah, that looks vaguely familiar - if you replaced all the caps!?

I don't see myself modifying hardware, unless forced by a repair, but I'll remember what you did. I do have a different sound card that needs new caps, but I don't have an oscilloscope - can I use a phone app?

Reply 914 of 941, by OM606

User metadata
Rank Newbie
Rank
Newbie

I'm spending quite a lot of time with my old stuff lately and ended up playing Far Cry a lot again and found out that there seem to be an issue with EAX through Alchemy and maybe also with native EAX under XP. It seems like some various subtle sounds are missing. For example sounds coming from messing around with objects. There are parts of the game where all sounds that should come from interactions such as grabbing ammos or opening doors are completely absent. It happened to me in the Research and Regulator in specific rooms or areas.

This is with an X-Fi Titanium with driver version 2.17.0008, Alchemy (can't remember which version), Win7 x64 SP1 and i seem to remember that exactly the same thing happened to me with an Audigy and WinXP last time i played the game about a year ago. Everything is just fine on another computer with onboard sound and Win7 x64 SP1. Can anyone confirm this bug?

What is funny is that i've decided to test Doom 3 on the same PC with onboard ALC1150 and thought that this game sounded quite better and much clearer without EAX. I don't think EAX makes that big of a difference in Far Cry.

Win7 x64 - Xeon E3-1271 v3 - Z97X-UD3H - GTX 960 - SB0880
WinXP - Q6600 G0 - P5B Premium - 8800 GTS 640 - SB0880

Reply 915 of 941, by NeoG_

User metadata
Rank Member
Rank
Member
OM606 wrote on 2026-01-15, 22:33:

What is funny is that i've decided to test Doom 3 on the same PC with onboard ALC1150 and thought that this game sounded quite better and much clearer without EAX.

Based on what was described here, where Doom 3 applies EAX on top of a fully software rendered audio environment already you wouldn't really want to play it that way IMO. The demo recording someone posted also didn't sound very good to me.

98/DOS Rig: BabyAT AladdinV, K6-2+/550, V3 2000, 128MB PC100, 20GB HDD, 128GB SD2IDE, SB Live!, SB16-SCSI, PicoGUS, WP32 McCake, iNFRA CD, ZIP100
XP Rig: Lian Li PC-10 ATX, Gigabyte X38-DQ6, Core2Duo E6850, ATi HD5870, 2GB DDR2, 2TB HDD, X-Fi XtremeGamer

Reply 916 of 941, by MattRocks

User metadata
Rank Member
Rank
Member

Isn't BioShock the exemplar Vista->OpenAL->EAX game?

Reply 917 of 941, by shevalier

User metadata
Rank Oldbie
Rank
Oldbie
MattRocks wrote on 2026-01-15, 21:18:

Yeah, that looks vaguely familiar - if you replaced all the caps!?

I don't see myself modifying hardware, unless forced by a repair, but I'll remember what you did. I do have a different sound card that needs new caps, but I don't have an oscilloscope - can I use a phone app?

Yeah, and half the resistors.
For example, in a Pierce oscillator, using a series resistor greatly reduces low-frequency jitter.
This resistor is on the PCB, but it's jumpered.
The X-Fi Ti HD designer had a broad vision, which he embodied in the printed circuit board.
But he failed to put the final circuit design into production.
Or why use the AMC1117 LDO instead of the LM117?
LDOs have a bug—voltage oscillation at low frequencies of 1-2 mV, visible even on an oscilloscope.
And more noise.
Or why isn't resistor C194 installed to bypass the this regulator reference voltage?
Even though it's listed as recommended in all circuit design textbooks.
Why use a different voltage regulator on the TPS with high pulse currents in the charge pump?
when a simple and inexpensive -5V linear regulator would do.

If developer Creative joined this forum thread, I'd love to hear his answers.

Aopen MX3S, PIII-S Tualatin 1133, Radeon 9800Pro@XT BIOS, Audigy 4 SB0610
JetWay K8T8AS, Athlon DH-E6 3000+, Radeon HD2600Pro AGP, Audigy 2 Value SB0400
Gigabyte Ga-k8n51gmf, Turion64 ML-30@2.2GHz , Radeon X800GTO PL16, Diamond monster sound MX300

Reply 918 of 941, by shevalier

User metadata
Rank Oldbie
Rank
Oldbie
MattRocks wrote on 2026-01-15, 21:18:

I do have a different sound card that needs new caps, but I don't have an oscilloscope - can I use a phone app?

Sorry, but no. I achieved incredible sound detail when the headphone amplifier was exhibiting parasitic oscillations at 50 MHz. About 10-20 mV, even without overheating.

Aopen MX3S, PIII-S Tualatin 1133, Radeon 9800Pro@XT BIOS, Audigy 4 SB0610
JetWay K8T8AS, Athlon DH-E6 3000+, Radeon HD2600Pro AGP, Audigy 2 Value SB0400
Gigabyte Ga-k8n51gmf, Turion64 ML-30@2.2GHz , Radeon X800GTO PL16, Diamond monster sound MX300

Reply 919 of 941, by MattRocks

User metadata
Rank Member
Rank
Member
shevalier wrote on 2026-01-16, 16:07:

Sorry, but no. I achieved incredible sound detail when the headphone amplifier was exhibiting parasitic oscillations at 50 MHz. About 10-20 mV, even without overheating.

I am admittedly jealous. Maybe one day I'll get better at electronics.

Have you posted your modified card here? https://www.reddit.com/r/SoundBlasterOfficial/