The verdict for indie developers (not using Unity/Unreal) seems: just bite the bullet and license FMOD (or Wwise or any of the popular proprietary audio engies).
I was actually searching for a good open-source audio library to use, and found out that my options aren't that good. SoLoud is a pain in the ass to install and integrate into an existing codebase, and OpenAL doesn't have any good implementations available (either proprietary or LGPL). I'm now just using a simple single-header audio library in cute_headers (https://github.com/RandyGaul/cute_headers/blob/master/cute_s...), but will probably switch to MiniAudio once the high-level API is finished (https://github.com/mackron/miniaudio/issues/196)
Somehow, it is completely counterintuitive to people that a coherent soundscape is important to establish a feeling of user presence in the rendered environment. Ambient sounds, footsteps, tiny impact noises etc. are essential to make a virtual world seem physical.
Partially I blame the focus that game marketing and game journalism puts on the quality of the game visuals. If your game's PR media package doesn't have good screenshots, you can pack up and go home. Maybe, the reason for that is that historically, games were marketed by getting articles about them placed in game journals and they could only present images. This has carried over to some extent to the web (screenshots are easier to make and easier on the server bandwidth than gameplay videos).
I think it’s just that humans are highly visual. You need to be able to see the thing you’re making to make it much more than you need to hear it. So sound is always an after thought and good sound also blends in to the whole.
Engine programmers are even more susceptible to this natural bias because there is a vanishingly small part of engine development that needs aural feedback. It always amazes me how many engine projects have fancy renderers but can’t do more than play one shot sound effects.
This audio bias extends into web video as well. The vast majority of videos get played without any audio.
On a side note, I've been going through games I've played that had excellent sound and some of them (e.g. the Half Life series) are among the better games of all times. The one prominent exception in my library would be Elite: Dangerous - it's a decent game overall, but not mind blowing. I believe that good audio is mostly correlated with the developer's attention to the complete experience and not just to a certain wow factor like stunning graphics or a particular gameplay gimmick.
When I worked in AAA we would get periodical reminders not to mute the game when playing. People still did.