Antoine's system

[Antoine]

Thx, that was somewhat my suspicion but wasn't sure.  

The referenced CA Thread, and it's positive results on my system, have made such an impression on me that I am also considering taking it a step further. The Melco device and it's DSM (Direct Streaming Mode via on-board DHCP service) caught my eye the other day and I have begun to contemplate if it would be possible to have a direct signal path via Ethernet from NAS to Music PC (hosting Roon & HqPlayer) and then to the mR via bridge. (no switch /router in between) Want to approach my local audio dealer this week and see if he has a Melco to test. Will let you know how things turned out if you are interested.  

Do you mean you will use the Melco as a NAS only device, connect it directly to your PC/music server and connect that to the microRendu? If so that sounds like a potentially great setup too. Though I have to say I'm not sure if streaming files from an upstream network device would be better than from local storage. Thus far my impression from others is that at least the type of local storage matters (SLC compact flash being better than SATA SSD, which in turn is better than PCI-E/M.2 SSD, better than HDD though there's all kinds of exceptions reported for both music file storage and running/storing the OS where running the OS in RAM seems to be best). I did find out that even the SATA cable used has significant impact... It's a weird hobby we have!!

Thanks Antoine.

What is interesting, in my opinion, is the fact that Romaz is busy trying to reproduce, on his server, all the improvements which have been brought by SGM to its SGM 2015, namely:

• high stability clock, powered by a low noise PSU, to replace the motherboard's clock
  
• low noise PSU for the motherboard and the CPU
  
• independant low noise for the SSD
  
• aso...
  

The SGM 2015 is supposed to be connected directly to the DAC, and we can therefore understand all the efforts brought on these items. On the other side, Romaz is using a sMS-200 as an NAA renderer, and I am therefore asking myself whether he is not doing the work twice, especially in the context of the near release of an sMS-200 Ultra which is scheduled to land in April, equipped with a very stable clock. I have read your answer to this question a few pages before and I am not asking this again. I still have doubts of my own.

So the point remains, in my opinion, to decide up to which point it is reasonable and useful to bring those optimizations on a music server connected to a high quality NAA renderer. I guess everyone will judge based on his/her own skills, and I appreciate that you are sharing yours, and his/her listening capabilities.

The first test for me will be to build a new server and to try to listen whether I can hear improvements in my system together with the Paul Hynes LPSU I am expecting in a few weeks now.

Thanks again.

I agree but of course the things you mention are and have always been the 'tools' we have (isolation, high quality clocking, high quality power, low noise hardware, optimised software for lowest, most predictable use of computer resources) to improve SQ. They just get implemented in different ways and differing degrees each time.

It indeed seems he is doing things 'twice' or better said after optimizing his PC based source he -may- find out the renderer (SMS-200 or microRendu) is no longer required or even detrimental to SQ. It's something he is probably going to find out as we all do along the way. He did BTW already upgrade his SMS-200 to SMS-200 Ultra specs using the SCLK-EX board and is using the extra clocks that board gives him (up to four independant clocks per board) to replace other clocks on his motherboard and/or NIC and/or elsewhere.

What's reasonable/usefull or not, is indeed for everyone to judge though along the way usually some consensus starts to form in the community. After that's happened it's usually the time audiophile press starts to write about these solutions. (they're always late ) When one waits even longer one doesn't have to resort to all these point solutions we employ but see commercial parties implement these solutions into their products.

Thus far I myself haven't felt 'diminishing returns' have set in yet but I also haven't jumped on the clocking thing yet (replacing clocks everywhere). This is higher risk/high cost though I have been looking at the different solutions for this (Pink Faun, SoTM, Tent, Dexa/NewClass-D, etc.) so it may be my next chapter too.

If the results you get with the SR7 are anything like mine I suspect you'll be surprised/amazed at the importance of a high quality power source, even when it's "just" powering a computer source. I wish it all made real sense to me and I or anyone could explain it in a sound way that's right in every aspect or way you look at it but so far I can't and haven't seen anyone who can. It doesn't stop me from experimenting (even though there's the constant doubt you too describe) and trusting my ears though.