By skyforbes 
It is episode 23, and this marks a level where i can finally tell that i reached somehow reference level.
I would claim that the system now sounds as good as if it was 10x as expensive simply because i combined the components in a specific way.
The point is not bragging, but showing that some knowhow is more relevant than handbuilt boards. Not to questions their legitimacy, but if you aim at a certain goal of ‘physically correct’ sound, not necessarily the most ‘musical’, then there are cheaper ways to get there.
The last missing 2 pieces were related to the electricity supply layer, which was solved by adding a Supra MD06 SP SPC power strip and suitable power cables (LoRad SPC 2.5mm) to all connections, with the exception of the RME device. Those form a path of silver plated cables from the wall plug to the devices, and an audio resolution tweak.
The first measures gave some very relevant, missing properties to the system:
- Static, stable stage
- Clearly drawn instruments
- Smooth heights
- Airy overall signature
- Better integration between base band and upper range, speaking of, sub and monitor bands
- Overall much finer texture
- Calmer, relaxed and easy listening experience
While those things sound not so spectacular, they made two things very clear: the NF / RCA cabling does have less effect on the quality than the electrical power cables, if one would have to weight it. And the second aspects is that the monitors, given a certain degree of quality of course, will only be about 10–20% of the overall sound signature compared to the other components. Now the 8030c have a slightly ‘thinned’ lower mid range, or, lack some body. I personally like this, but there is some minor room for improvement.
With this new electric power supply improvements, physically, the following aspects have been optimized:
- 1. RF interference suppression
Reduction of high-frequency common-mode and differential interference
Less RF interference in DAC, clock, and amplifier stages. - 2. Ground potential stability
Lower drift/modulation of ground potential
Fewer compensation currents via USB - 3. Reduced sheath currents/common-mode currents
Better control over GND return paths
Lower RF load on NF shields - 4. Lower network impedance during load jumps
More stable power supply for subwoofers and monitors
Less interaction between devices - 5. Reduced cross-interference between devices
Laptop RF less effective on RME/monitors
Internal filter load distributed in a more structured manner
To make this very clear: those are measurable properties, and as you can see, the question is not ‘if the cable sounds better’, but that you see what it does from a physical perspective. Those are higher frequencies suppresion (1), which means, narrowing the bandwidth that is transmitted towards the devices, creating a more stable ‘ground’, which marks a stable relative 0 voltage level (2), reduced amount of noise coming to the NF cable shielding (3), and by reducing the sequential impedances of the different conductors, the inductance is reduced and thus the current changes are less affected(4), and it is also important that any kind of RF/HF interferences are reduced, for example does the power supply of the laptop, which has to be connected to the strip for phasing, necessarily create RF artefacts and capacitive electricity junk, which would otherwise just enter the current network unfiltered (5); by careful plug order in the strip, meaning, the power supply of the laptop will be the last in line, the internal filters of the Supra MD06 will absorb lots of those if the flow back from the laptop end towards the RME devices, but also towards the monitors (and the sub), which are farther away in the plug order.
The last aspect is very detailed, but all those electricity network aspects support a effortless, fine grained and opened stage.
The other puzzlepiece was finally understanding how resolution flows from Tidal over osX and the RME Fireface UCX II. Tidal can play back up to 192kHz. But what exactly happens to this ‘resolution’? Just because you have a nice yellow badge in Tidal does NOT mean you will hear 192kHz right away, and swithing to exclusive mode does not do the job either. It is worthwhile looking into the conversion flow, specifically for osX here.
There are 3 layers:
- Tidal potentially offering a resolution up to 192kHz (2025)
- Under osX, all audio runs through CoreAudio SRC (Sample Rate Conversion); note that those are optimized for 96kHz (multiples of 48kHz) (macOX 15)
- RME SRC; these are also optimized for 96kHz since the DAC family in current RME DACs are 48/96kHz optimized as well
Setting Tidal (under osX!) to exclusive mode will not auto-change the sample rate neither in the osX layer, nor in the RME device. But it is still relevant for avoiding extra conversions, which can happen per default (eg. osX system rate conversion -> SRC CoreAudio Target Ratio -> RME ..).
This leads to a setting of 96kHz being the sweet spot for the RME device under osX as static setting, optimizing the conversion flow by minimizing the conversion steps and the up/downsampling process. The only ‘tiny conversion path not even’ is 44.1 -> N, but this is handled perfectly by the osX CoreAudio SRC.
Also note that upsamplig is cheaper than downsampling; one could conclude that it may be cheaper for the converters (especially the DAC converter in the RME) to only convert at lower rates, but that is not the case (any more). They are optimized for 96kHz. And as mentioned, upsampling is cheaper, it basically means to just add ‘some steps at the same height’, instead of having to interpolate: if one had to resample a 96kHz signal to 44.1kHz sampling rate in the RME (or also in the SRC of CoreAudio), this means that ‘2 steps need to be interpolated’, which is more error prone and costly.
I know, this is all very detailed, but it makes sense from a logical standpoint, in contrast to wobble about bright sounding cables, if you can follow me.
So as mentioned, the result is what i would call a reference state. I will document all this in a handbook, possibly also in the form of a step system such that other people can emulate the process for themselves. At least in theory. Not using a streamer does give me the comfort of using tidal also on my laptop and controlling it; however, that would be the cleanest solution in the end since it eleminates the CoreAudio SRC layer. It would probably not make too much of a difference.