First impressions of new MH-DA006, Musetec flagship

At some point this winter I'll likely purchase Pink Faun I2S card, or hopefully JCAT comes out with I2S card.

Comparison then becomes:

1. Present setup is server-Roon core, streamer-Sonore OpticalRendu-Roon endpoint. Server>Sonore OpticalModule>Optical Rendu>Denafrips Gaia via usb>006 or Harmony via I2S (all components fed power from external lps).

2. Server/streamer with JCAT or Pink Faun I2S card-now contains both Roon core and Endpoint>006, Harmony via I2S. The Pink Faun or JCAT card will be utilizing external lps vs motherboard supplied power.

So No. 2 setup so much simpler, get rid of three components and their attendant lps and cabling, no more usb.

Question is simpler better? Less complexity means less chance of added noise, but then optical conversion completely eliminates transmission of upstream noise. How does Pink Faun (or possible JCAT)  I2S compare to Gaia I2S?  Gaia has added burden of optimizing usb.

This also brings to mind attempts to generalize about usb vs I2S performance. I suspect individual setups could have quite different outcomes, optimized usb at both streamer and dac could surpass less than ideal I2S outputs from streamer. I do believe I2S at dac end is inherently superior to any usb as its native protocol in dac, one is simply adding all that usb complexity to signal path.

Someone over at the other forum has compared PF I2S (w/ Ultra PF ocxo) with their USB counterpart (PF & JCAT) a while back and found they are at par - they sound different but none is better than the other. They both are of such high quality that it’s a matter of taste at this point. This tells that I2S is not always the superior input for DACs regardless of implementation. Some DACs will sound better over I2S, others not so much.

IMO, two things are needed to optimize USB from the server/streamer side - first is an excellent power supply and the secondly a good quality clock, preferably a good quality ocxo. Another thing that is often overlooked is the network side of things, like network interface, Copper vs Fiber Optics, downstream switches, etc. These play a major role in contributing to the SQ. Moreover, if you have a DIY server, the power supply to the server is important, so are the choice of dram module, low latency drive, OS optimization all contributes. If these things are not in order, the 2nd best solution is a well implemented DDC, like Denafrips Gaia or the Audio-gd DI-25HE.

Recently PF has introduced an Ultra USB Bridge (with in-built Ultra ocxo) that comes with a substantial hefty price tag (over $3k). Someone who tried it said it sounds substantially better. Well, at that price I won’t be trying it myself but I do plan to modify my JCAT XE to feed a good quality ocxo or get the JCAT ocxo board which is also a bit expensive for what it does but it has two 20Mhz outputs - so if you are using both USB and Network card from JCAT, then it might be beneficial. I use fiber out straight away from my server and hence I can’t reuse the 2nd clock as it.

Agreed with the post above, the USB format still offers SOTA sounds. I have owned all 3 of them and from my listening, I found this in order of preference for the Pink Faun bridges:

1. Ultra USB bridge (built in OCXO and UC power)

2. USB v2 bridge (with the new firmware)

3. USB v1 bridge with OCXO card plug-in

I've also listened to the PF i2S bridge too but not in the same system. I think it slots in around the 2 level. i2S is good but is apparently very sensitive to connection disturbances. 

To get the most out of USB, one has to experiment with external power and cabling, both of which makes a difference.