I have a wonderful sounding and highly revolving system with no room treatments or DSP. I know there are real issues with the room but to a large extent have been mitigated using very high quality player and optimising software on my server pc that result in very low jitter and latency. These are Windows server 2019 core, Ramdisk, Jplay Femto, Audiophile Optimizer and MinorityClean. The resulting time coherence has greatly improved bass and also edgy high vocals. These are not nano differences though I certainly take your point
@hulkss.
I believe DRC is the most promising direction to go and AL looks like the best choice for me. However I cannot get ALC to work in my system and I am very reluctant to use any other player than Jplay Femto. For example I have tried Jriver but it just does not sound as good, partly due to the fact it will not run in server core. GUI mode requires many additional processes need to run and they create jitter, so does the feature-rich Jriver. I'm also trying to avoid additional links in the audio chain like Asio4all as in my experience it reduced sound quality considerably.
If I do a music server build based on Jriver and got wonderful results, I'll always wonder how much better it could be using Jplay.
I was in a similar situation. As you have described, getting computer audio to sound great can take a ton of work. All sorts of O/S optimization and getting hardware and software to play nicely together to deliver a clean signal with the lowest possible jitter and noise to the DAC.
The problem is with the architecture. Unlike a CD player, a general purpose computer with a thousand or more background tasks running is a complex beast that does not behave anything like an audio component. Have a look at what the Aurender guys do to make their music servers get out of the way of themselves. Crazy stuff.
The best solution I found was to adopt a different architecture. Instead of asking the music server to act like an audio component, move the server to some other part of the house. Get it as far away from the listening room as possible, taking advantage of the inverse-square law to eliminate noise. Then, replace it with an actual audio component to bridge the playback system to your home network. Ethernet and fiber networks give you galvanic isolation for free and do a much better job of rejecting common-mode noise than USB and analog cables.
As you have seen, for my implementation, I have chosen Roon. The server (Core) is located two floors below my listening room. My network audio transport varies, but I have the ZEN Stream from iFi Audio and a number of VitOS for RPi4 builds, powered by an LPS. They are all extremely low noise and deliver a clean signal to the DAC.
This approach leaves all of the heavy duty processing on the Core machine, located far away. This includes uncompressing FLAC files and streams, volume leveling, Convolution of FIR filters, MQA Core Decoding, etc. The network audio transport has virtually nothing to do. It just copies audio from its network interface to its USB or S/PDIF output. This enables it to deliver much better noise and jitter performance than is possible with a general purpose computer at a fraction of the cost.
There are many other advantages to this approach that I cover in my 2-part article on Copper Magazine:
https://www.psaudio.com/copper/article/roon-done-right-a-user-guide/
I realize that you're not interested in making a change, but it took me a few years to come around as well. Hoping this information will plant a seed. Cheers.