Based on my past 5 year history of experimenting with and optimising my network I believe there are 3 elements to optimising sound quality in music streaming.
The better the quality of your audio network’s Physical Layer, the better the sound‘s ‘presentation’ ie. Your brain can construct a more realistic, live sounding soundscape of musicians spontaneously making music in a venue, recorded or engineered
For ALL network components, Better in = Better out. The better the stream quality going into a component, the better its output. Better = less noise, less jitter, less vibration, less cable loss, better power supplies, lower traffic, etc.
The less traffic on the ENTIRE audio network, the better the sound quality. The SOONER you can separate the audio related stream from the rest of the household demand, the better.
If you do the separation WITH the Taiko router, it will not sound as good as doing it before the router, simply on a better in = better out basis. In 5 years of experimenting and reading others’ experiences, I have never seen this Axiom disproven. Improve the input measurably and the output will improve.
So what that tells me is:
Ideally, your Taiko router and switch should be vibration isolated, have the finest power supplies and DC cables attached and have all non-audio related traffic separated off before either device.
In the end, the biggest influence on what you hear will be the power supplies employed, because they generate the bit stream and its the bit stream‘s physical quality that has the most impact on the output of your DAC. Assuming a bit-perfect stream of bits, its how the bits are structured and how much ’extraneous’ voltage and timing deviations are built-in to their structure that matters most.
In my network I had 6 streams of Sean Jacob’s ARC6DC4 power and extensive anti-vibration measures and the resulting sound was utterly gorgeous. That stream was upgraded from DC3 to DC4 to ARC6 and each upgrade brought huge improvements. Upgrading DC cables from really good Neotech to Mundorf silver/gold also brought jaw-dropping changes, mainly to purity, holographic imaging and a sense of reality and being present in the actual venue as the music was being created. Throughout those upgrades my perception gradually changed from listening to really good quality recorded music, to being at a live performance of the music. All the increased information I heard was the result of my brain being able to clearly resolve more information from the resulting soundwaves. The information was already there, but the accompanying ‘noise’ of varying sorts meant that my brain couldn’t resolve it from other parts of the music. Drop the noise and that impacts the brain’s ability to resolve more.
The BIG deal architecturally for a network built for sound quality is that as the stream progresses from incoming wall to final client, it should encounter constantly improving physical layer specs. For example, there no point having a low noise stage, low vibration stage or low jitter stage if the next stage or any downstream stages are worse. In audio, the network is a Stream Conditioning process, which should constantly improve the physical layer until what reaches your DAC or Server is as close to perfect as it can be.
Unfortunately I can no longer hear any of this due to complete hearing loss in one ear 6 weeks ago. I just wanted to pass on some of what I’d learned. And yes, my entire system is now for sale as I need to morph it into something more suitable and fun for a person with single sided deafness.