The only "alternative" implementation I can think of is OSS, which was dead for a while now
I'm digging backward through highlights. Just wanted to point out the story about OSS.
When Linux was first rising to prominence in the early 1990s (i.e. the Softlanding Linux System days, early Debian, proto-Slackware, etc), there was only one sound card that everyone using Linux cared about: the Creative Sound Blaster 16. Drivers became available quickly, sound quality was allegedly excellent, this led to more people coming along to make drivers for other cards that "emulated" Sound Blasters. All these projects agglomerated together until they eventually morphed into the Open Sound System. No longer limited to Sound Blasters or Sound Blaster compatible, there was now an entire team writing up drivers for sound cards. They made 3 versions that were all libre, and it was "de facto" on all Unix-like environments from every Linux distribution at the time to the BSDs to Solaris and everything else in between.
I think it was 2002, the maintainers of the OSS project decided to make OSSv4 proprietary software. That pissed off a
lot of people, so a solution had to be made. The BSD and Solaris people decided "okay, we'll pick up where OSSv3 left off," and
that's what powers all audio on the BSDs and commercial Unices like Solaris, HP-UX, AIX, and so on
(a touch reductionist; each Unix-like OS does shit differently on some level, but they still contain an assload of OSSv3 code along with a hodgepodge of native stuff). Linux, for whatever reason, went in the polar opposite direction and decided to completely fucking rewrite the sound implementation. This gave us the "Advanced" Linux Sound Architecture... aka the ALSA that we all know, love, and loathe to this day. While ALSA is more than serviceable today, it should be remembered that ALSA in its infancy was fucking abysmal. The API was poorly documented at the time, tons of features from OSS were missing for a good long while (i.e. audio mixing), audio drivers had to be completely fucking rewritten because the API, ABI, or whatever the proper abbreviation is weren't compatible with whatever OSS had at the time. There was a nasty stretch of time in the mid-to-late 2000s where it was a real crapshoot as to whether or not your sound card (be it a dedicated DAC or your motherboard's built-in audio) would work properly. Tools like Jack and PulseAudio came to the "rescue," they solved tons of problems, but they also made tons of problems too.
I'm rather ambivalent on PipeWire as a replacement for tools like Jack and PulseAudio. I've been living with PipeWire for about a year since deciding to daily drive Linux again last year, and I honestly don't notice any conceivable difference. The last time I daily drove Linux like this was back in high school like 15 years ago during the Ubuntu 10.04-12.04 days, I'm running vastly different hardware now than I had back then, but there's one issue that persists to this day that won't ever fucking go away no matter what I try. It's an issue that PipeWire's fundamentally incapable of solving the same way Jack and PulseAudio failed:
sound's bloody quiet innit mate?
What I mean is that both 15 years ago and today, the built-in sound on my motherboard is too fucking quiet, and it's an
ALSA-specific issue that software layers like Jack, PulseAudio, and PipeWire all basically slap duct tape onto and say "job's done." I need to to boost my shit on Linux to like 40% to achieve what would easily be 20% on Windows. I know it's the same issue because the OEM sound on my old HP tower's motherboard and my current motherboard are recognised as some variation of
snd_hda_intel, and the codec is the same damn thing: some Realtek garbage that has eleventy billion fucking variations split across dozens of OEMs.
Technically, ALSA has the drivers for all of them. The problem is that ALSA
always defaults to a shitty generic driver that just cuts off 27-30dB from the volume I'm capable of.
You can do all sorts of shit with
hdajackretask to apply boot overrides that swaps your front sound with the rear sound jacks to get the gain to work better, you can run
alsamixer -c 1 to unmute and boost all other channels to 100% while keeping the master volume relatively low, you can set
alsa-base.conf to whatever OEM you have like
options snd-hda-intel model=auto, reboot your PC, and pray like hell it works only to be bitterly disappointed when it doesn't. It's all one gigantic exercise in frustration, and it persists
regardless of whether it's jack, PulseAudio, PipeWire, or whatever other newfangled thing Linux developers will adopt when PipeWire gets too tedious to maintain 10 years from now. It just fucking sucks, and you're only able to put so much mustard and hot sauce on a fucking turd sandwich before you gotta shove the thing down your gullet.
The only real solutions I have are suboptimal no matter which way I slice it:
a) Live with this suboptimal experience from my on-board sound as I have it now, with various tweaks and fixes that help but don't go the full nine yards.
b) Buy a new sound card, and pray that I actually did my research properly when I part with cash.
c) Use headphones and succumb to sweaty ear plague... because that's the #1 thing I've seen across LinuxQuestions, Yahoo Answers, Quora, and even Facebook. "Oh too bad for you mate, I use headphones and it's good good enough for me."
d) Roll the dice and see if I'm doing PipeWire wrong, switch to PulseAudio to see if that changes anything, git gud at learning Jack and its intricacies, or some other activity that puts my nose to the grindstone and still probably won't accomplish shit.
***
TLDR: Sound on Linux is garbo because ALSA itself was shit 20+ years ago and it's still shit now. You can only do so much with sound servers that run atop ALSA because ALSA itself is an unstable foundation. Oh and OSSv4 was later relicensed as free software, but no one gave a shit anymore because they all moved on.
