Hopefully with better eyes can find Waldo.
Really nice diagram plus some really cool toys. I might try to replicate this.
I wonder though if you ever walk across the sitting room just to turn the television off. 
@Progisus , I suspect the MScaler and other âfiltersâ involved (in K30 and Hugo) maybe a factor in noise reduction in this set up. I am stumped how the network creates what you are suggesting is immunity.
The MScaler is a noise generator. I mean that almost literally as it is well known to pass a considerable amount of RF on to DACs. Significant audio benefits can be gained from taking step to address that noise. Those steps can also make it a lot easier to notice improvements from network enhancements.
I didnât suggest the MScaler as a culprit behind he doesnât notice differences in the devices he mentioned as he also listens without the MScaler so I assume he was not hearing those differences even with MScaler involved but in hindsight I should have confirmed that.
Yes, hence why I said âŠ
Objectively, and not meaning to criticise anyoneâs beloved system, it seems to me a system has various sources of noise, interference, distortion etc and the cumulative result of those might be such that changes in ethernet are too small to detect. But if all those other sources are adequately addressed/reduced, then the impact of changes to ethernet will likely be more apparent.
I didnât realise/expect the MScaler would be a source of noise. I was assuming its algorithms/filters, like many filters, were reducing or shifting noise so its impacts are not/less audible.
It doesnât need to be cumulative as a single component can have the noise floor at the level where addressing network noise brings no apparent benefit. Remove that one component and that could be enough change the equation.
Itâs an upscaler that uses a FPGA to run the algorithm - and RF generated by that FPGA wasnât addressed sufficiently at the ouput. Itâs purpose is to reconstruct the analog signal more accurately. It does have the ability to filter but thatâs more about reducing the quantization noise that results from digital reconstruction. Itâs not something that would make it more difficult to hear benefits of network optimization. Quite the opposite, when the RF is well addressed, one is able to hear deeper into the music.
Thanks for sharing.
You are streaming when you talk about not hearing differences? No local files, correct?
In the diagram it seems it did not replace but was added instead?
I wonder what Netgear?
I shared your curiosity and tried successfully. So it confirmed I had to work on that part of the network. Unfortunately, I couldnât justify the expense of the Phoenix in relation to other components.
If possible, you could replace the Netgear (and G-Tek and Sonore) and give it a try. I would be surprised if you couldnât hear a significant difference.
Edit: Continuing with the idea of isolating your audio from other home devices, it might be beneficial to separate the listening room network immediately after the modem, as many have done. An EtherREGEN switch with an SFP cage is one option to consider.
At this stage, it seems difficult to say that SFP ports should automatically be standard across all Antipodes servers. The more interesting question is whether the benefit comes from the SFP interface itself or from the way a particular implementation reduces electrical noise before the signal reaches the server.
Many of the listening impressions being shared point toward lower noise floors, better instrument separation, improved bass articulation, and a more natural presentation. Those are all worthwhile improvements, but they donât necessarily prove that SFP is the only path to achieving them. A well-optimized Ethernet connection may still perform exceptionally well depending on the overall system design.
A good analogy is the difference between cleaning methods in other industries. A company such as RMS Cleaning can achieve excellent results with different tools and processes, but the final outcome depends on the entire approach rather than a single piece of equipment. Network audio can be similar, implementation often matters more than the connector type alone.
Until there are more controlled comparisons between high-quality Ethernet setups and equivalent SFP-based configurations on the same Antipodes platform, it may be premature to conclude that every server would benefit from adding SFP ports. The recent reports are certainly interesting, though, and they suggest this is an area worth continuing to explore.