I've tested a lot of different settings and none of them really give me what I'm looking for.
Audio Mix:
Night Mode, Wartapes and Wartapes V.A.L. tend to have a narrow dynamic range, which is good for hearing quiet sounds like footsteps. Unfortunately it also means that the sound tends to become a little bit more muddy and less holographic. It's easier to hear footsteps, but it's harder to identify their exact location. The footsteps are still way too quiet though.
Default Mix and High Dynamics have better dynamic range. Default mix was just bad across the board. I found that High Dynamics offers the best sound quality and separation, but it does require you to crank up the volume to hear the quieter sounds. It's still feels too distant and not very holographic, you can't really pinpoint the location of the footsteps.
Sound System:
I haven't tried Mono, 2.1 stereo, quad, 5.0, 5.1 and 7.1 because those tend to be for multi-speaker setups, while I play with high-end headphones. Spatial 3D seems broken. Stereo/Headphones both offer the best sound quality as far as I can tell, but footstep loudness remain a problem.
3D Headphones in combination with something like Windows Sonar or DTS Headphone X seems to do well in terms of sound separation, directionality and allowing you to properly locate footsteps, but the footsteps still seemed way too silent. The reason why I don't like using 3D headphones is because enabling Windows Sonic or DTS nukes the sound quality in Windows to a maximum of 24 bit 48 kHz. This is significantly worse than what my system is able to output and it just sound so much worse.
So I don't really bother trying to fix it anymore. There are plenty of other games, even older Battlefield titles that have way better sound. I just default to Stereo/High Dynamics, because that's what sounds the best on my system in terms of sound quality.