Importance of source quality with asynch USB DAC?

I have been plowing through this technology for about six months with the following observations:

1) A good asynchronous USB -> SPDIF interface box (I use the ART Legato) gives me the best sound for RBCD (better than either an SPDIF or AES/EBU dedicated audio card). The things the interface has to do (besides low-jitter buffered and re-clocked output) include isolating the computer from the DAC. I find streaming-mode USB interfaces and interfaces that use USB bus power sound inferior (almost always, jitter) to an asynchronous interface with its own clean power.
Interestingly, USB cables *do* make a difference; the best sound I get is from an ART cable that carries the signals but not the power lines. The presence of the power wires in most USB cables (most of which, excepting audiophile types, make no effort to shield data signals from power lines) had a deleterious effect on the interface, even though the interface made no electrical contact with the power lines. I also had an improvement from an optical USB (self-powered) cable that provided complete electrical isolation.

2) I have built three servers; one is a 'loss' in that it does not produce SQ at the level of the other two, whether by sound card or USB. I thought it was the power supply, but upgrading it did not help. The difference could be in the motherboard construction or power handling or it could be that the two much more powerful servers sound better because they have more RAM, more CPU power, and SSD for the OS (Win7). The 'loss' is a Core2 E5300 with 4GB RAM. The two better-sounding boxes are I5-650's with SSD and 8GB ram. All three use Corsair power supplies, wildly over-spec'd (650 and 850w).

Not sure we understand yet why one computer sounds better than another; laptops at least can be kept off the power lines and in many cases sound better than desktops. While I am no fan of the 'fanless' server with linear DC power (too many clocks in the PC for this to matter much), I do suspect that the more powerful processors produce better sound, especially in a Windows environment when the operating system always (always) has some busy-work to do. De-configuring services, indexing, and restore point processing is tedious and error-prone. Reduced O/S busywork in may have a lot to do with the excellent reputation of Apple boxes as music servers.

High power cpus usually do get a bad name in some circles. While it doesn't take a ton of cpu to playback 192kHz material for example, system latency may be an issue, and with a more powerful cpu maybe their are less DPCs. I have noted how driver problems caused high system latency using J. River when trying WASAPI vs. Lynx ASIO using their awesome library server over LAN. ASIO would cause high system latency and strain my low power ION board's gpu when minimizing and maxing the 1920x1080 screen - causing one click. WASAPI with same material, would result in super low system latency and no clicks or pops. It realy is amazing how well implemented WASAPI seems to be, for the most part.

This was with a dedicated gigabit wired lan connection, not wireless (which exhibited the same behavior but MUCH worse). There is this big anti-buffer, anti-power commmunity and I think it may be like Mike said, that with more crunching power you can actual better lock down everything Windows is doing. No one posts ABX tests, but hey maybe the difference when they add a linear is "night & day".

I think it's a balancing act which must be struck because at some point you are pushing the PC, such as streaming high bitrate video (mkv at 100000kbps or 192kHz audio at 3-5000kbps), closer to the edge of data shortfalls which results in buffer overruns that a DAC can't always handle (loss of sync or white noise_ because WASAPI lets the hardware and driver decide what to do with such data shortfalls (as opposed to ASIO).

Hi Mryan,

I've built a few servers as well, and similarly, one of my servers doesn't sound as good as the other two. I don't think the cause is specifically CPU power, as the one that doesn't sound very good is more powerful than one of the machines that sounds very good. Hence, I am interested to know what motherboards you are using in your various servers that both sound good and don't sound good.

The machine I built that does not sound quite as good uses the same style Seasonic power supply as in my other machine, an Asus P5W DH Deluxe motherboard, an Intel E6600 CPU and Corsair RAM; all reasonable quality stuff.

Alan

The systems I liked use Intel DH55TC mobos with I5 680 cpus. The system I have trouble with is an Asus P5QPL-AM with an Intel E5300. The Intel boxes carry 8GB and the Asus 4GB.