SS2 Headphones Sound Upgrade

[665afd149d059Marvin]

If you want to.

[665afd149d23cMarvin]

I tested your Thief savegame and didn't have any trouble, at least none of the popping/tearing sounds you describe. Thief's reverb is indeed a lot louder than SS2's and with HRTF enabled, you get this sort of multi-reverb effect especially in the burrick tunnels (reverb-heavy EAX preset) because the footstep sound is played for both ears a little differently which, while realistic, can be a little bit jarring when you're not used to it. So it might be useful to tone that down as explained in the OP, but that's about it.

Now, whatever causes the problem for you is almost impossible to hunt down: Driver, OS settiings, driver settings, sound card, headphones, cables - could be anything. The least I can tell you is that it's neither due to the game (well, not my installation, but hat's just the original + TFix, same as you), nor the sound library. If I had to make a guess,  I'd probably blame some "fancy" 3D/quadrophonic/crystalizer/... effect enabled by the sound driver or a loose/broken headphone cable, but that's pure speculation. Oh, and I'm using a different HRTF table (IRCAM 1031, you have the default KEMAR).

[665afd149d325Fallen+Keeper]

Ok, thanks for your time. I will try different hrtf presets if problems persist.

[665afd149d58dOlfred]

Can you make a sound recording of the "glitch"?

[665afd149d6edcitadel]

Marvin, thank you very much for this post and the detailed install/setup guide, got it working with my custom HRTF table (IRC 1020 works best for me).

One note - my MHR is considerably smaller than the in-built default ones:

Code: [Select]

53,853 default-44100.mhr
53,853 default-48000.mhr
13,417 irc_1020_44100.mhr

The built process finished without errors:

Code: [Select]

c:\Users\citadel\AppData\Roaming\openal\hrtf_defs>makehrtf.exe -m -r=44100 -d=sphere -i=.\IRC_1020.def -o=.\irc_1020_44100.mhr
Reading HRIR definition...
Calculating diffuse-field average...
Performing diffuse-field equalization...
Performing minimum phase reconstruction...
Truncating minimum-phase HRIRs...
Synthesizing missing elevations...
Normalizing final HRIRs...
Calculating impulse delays...
Creating MHR data set file...
Operation completed.


[665afd149d7e7Marvin]

Good to know the guide helped you making your own table.
The default tables are built with the KEMAR set from MIT which provides 710 records at different positions whereas IRCAM only uses 187 positions, hence the smaller file size.

[665afd149d915citadel]

I see.
Does it mean the KEMAR MHR has a better "resolution", more detailed table, thus sound localization should objectively be better as OALSoft needs to "interpolate" the IRCAM samples more?
I did not hear much of a difference between the custom and the default MHR, just created it "because I could" - may as well revert back to the original.
Thank you!

[665afd149d9f6Marvin]

KEMAR has better precision, IRCAM has better accuracy since you can select a lot more different HRTFs. The choice is yours.

[665afd149db5dKolya]

I added this thread to MOTHER's collection of SS2 Help Guides and changed the icon of the first post to the same lamp symbol all guides have.
I will not move it to the Helpdesk where the other guides are that deal with immediate installation problems etc, since I feel this is more of an advanced topic.

[665afd149dcebcitadel]

I can't reiterate enough how I enjoy this mod/upgrade!
A bit of searching brought me to the personal page of this modest Swiss/American guy Toni Schneider (https://toni.org/a3d), who worked on A3D, explains how Creative buried the original 3D sound functionality after they had bought Aureal and shares some impressive A3D demo videos from an old CD he found.
The binaural experience was flawless already back then in game classics like the original Unreal (makes it even more unreal), Motoracer or Jedi, sadly Creative locked A3D up.

[665afd149de08Marvin]

Another happy customer,thanks. 
Yeah, A3D2. 0 is still ahead of anything we have today. I'm confident however that headphone sound will see a lot of attention soon thanks to VR.

[665afd149e0c9citadel]

I'm confident however that headphone sound will see a lot of attention soon thanks to VR.

Feel that in my guts as well and have started reading, checking out the HRTF frameworks etc. :] A binaural game is my upcoming project. I am playing A Blind Legend on my phone and it seems to be an interesting market, it just gives a completely new perspective to such a common thing as even some cheap earbuds.

[665afd149e369Olfred]

Yeah, A3D2. 0 is still ahead of anything we have today.

When it comes to an interface or sorts, yes.
Nowadays the good games have 3d surround for headphones built in.
As example, Battlefield 3 and the following titles have pretty neat sound atmosphere going on.

[665afd149e5faMarvin]

When it comes to an interface or sorts, yes.
Nowadays the good games have 3d surround for headphones built in.
As example, Battlefield 3 and the following titles have pretty neat sound atmosphere going on.

Isn't that just a variant of CMSS3D though?

[665afd149e745Olfred]

No.
Whatever produces sound does have one or several (like tanks) sound generators attached to them. You hear the sound relative to your position. Additionally obstruction of geometry and reflection of sound depending on the material is also calculated.

There was some feature video from the developers where they showed of and described a little bit about it. Sadly I can't find it anymore.

[665afd149e84fMarvin]

Interesting. EAX type effects making their way back into gaming should be a given since devs can do all those things in engine  in software now. Not sure what you mean with the positional audio. I'll try to look for that video.

[665afd149e9ccOlfred]

It's more close to A3D than to EAX.
What A3D did was to get (a simpler version of) the world geometry and simulated the sound playback relative to your position.
So when a sound was played around the corner, it was in lower volume.
EAX simply added effects on the sounds, disregarding your position in the world.
(later EAX was better, but let's just stay at the time of SS2).

Back to Battlefield (or the Frostbite engine). If you have a Tank there are somewhat like 20 or more sound generators on the tank itself.
Now if a tank drives by from behind, you can hear it coming from behind and driving past you.
It even sounds different if you are standing up or laying down as some of the sound generators are now closer to you and some other get obstructed by the tank itself. Now if there are walls around or the ground is made of a special texture, you even hear the echo comparable to the materials surface.
The sound engine is really rich in features when it comes to 3d emulation.

[665afd149eb2fMarvin]

https://valvesoftware.github.io/steam-audio/

Cool. It would certainly be nice to have an open-source, hardware- and OS-agnostic software audio engine available which is maintained by more than one guy in his basement.

[665afd149ec0dLightwave]

Warning: latest OpenAL Soft 1.18.0 breaks the ghost despawn.

[665afd149ecfbMarvin]

In what way, do you know why? I'm sure the project creator would like to fix that. Or me, if it's actually my fault. I'll look into it when I have the time.

[665afd149ee1bLightwave]

> In what way, do you know why
Maybe it's OpenAL bug like memory leak or something (because 1.18 implements many features) or bug in NewDark audio subsystem. Who knows~

I can try to make git bisect on OpenAL codebase and see what's the commit breaking the game.

[665afd149ef51Lightwave]

Ok, the problem was gone if you compiled library by yourself with a new Visual C++ compiler (2017 Community, for example).
OpenAL releases on kcat.strangesoft site was compiled with GCC MINGW-64w and I think it's has an issue with code generating on certain circumstances (well, I'm not an expert in cross-compiling).

Sorry for the disturbing (and for my english, too).

[665afd149f08eMarvin]

That was quick, and you don't need to apologise for your English, native speakers are actually in the minority here I think.

If you can describe the problem, could you post it to the project's bug tracker? Maybe it will be fixed eventually. I'll keep linking to the older 17.2 version for now.

[665afd149f1a3Marvin]

I updated the post to OAL Soft v1.18.2 which comes with built-in HRTF tables, making it much less probable to screw up the installation. I did notice a bug with the ghost despawn as Lightwave suggested, but that seems to be tied to a SS2 bug and happens regardless of the sound lib used.