Hi Brian,
Are you a JRemote dev?
It's been well discussed in various forums and hifi magazines over the last couple of years. The effect of high frequency noise on DACs is measurable and very tangible. RF noise inter modulates with the analogue parts, creating noise floor modulation, which makes it sound harder and brighter. As such, excuse me if I sound blunt but it's probably best not have another thread here arguing about the ins and outs of USB noise transmitted through common ground. If you google for it then there is good info out there. For example this product is a heavy duty galvanic isolator and filter:
http://intona.eu/en/productsThis is just a low pass filter:
http://www.audioquest.com/jitterbug/jitterbugThe problem is summed up well by Rob Watts who is an electrical engineer who has designed the latest high end Chord DACs (as well as the custom DAC chip they use):
"There are two problems that USB has against toslink - and one benefit. The benefit is that timing comes from Mojo - but with toslink the incoming data has to be re-timed via the digital phase lock loop (DPLL) and this is not quite as good - but you will only hear the difference via a careful AB test, so it's in practice insignificant.
The downside with USB is the common ground connection. This will mean RF noise will get into Mojo, making noise floor modulation worse. Now I go to very careful lengths to remove this problem by using lots of RF filtering, and double ground planes on the PCB, but even minute amounts of RF is significant. The other problem is down to the way that digital code works - which is in twos complement. So zero is in 24 bits binary is 0000_0000_0000_0000_0000_0000. If the signal goes slightly positive then we get just one bit changing to: 0000_0000_0000_0000_0000_0001. But if it goes 1 bit negative all the bits change to: 1111_1111_1111_1111_1111_1111. Now the problem with this is that when a bit changes, more power is needed, and this injects current into the ground of the PC - and the ground will get noisier. Unfortunately the noise is worst for small signals. Now the problem with this is that it then couples through to Mojo's ground plane, and the distorted signal currents will add or subtract to small signals - thus changing the small signal linearity. This in turn degrades the ability of the brain to re-create depth information, and so we hear it in terms of depth being flattened. What is really weird about depth perception is that there seems to be no limit to how accurate it needs to be, so the smallest error is significant.
So with toslink we do not get these problems as there is no common ground - so no RF noise, no distorted signals on the ground, and it will sound smoother with better depth against a noisy PC. But the problem can be almost eliminated by using a power efficient USB source that is battery powered - such as a mobile phone. But with noisy PC's the only way of solving it is to use galvanic isolation on the USB - but this draws power from the source, and we can't do that with mobile devices. All of Chord's desktop DAC's have galvanic isolation on the USB, and then you can't hear whether its a noisy PC or a mobile phone. In this case, USB sounds slightly better than optical, because we have the (tiny) timing benefits of USB.
I hope that explains - its a complex subject.
Rob"
I think my suggestion in the first post would be a really valuable feature of the JRemote software. Also it would allow a person to use an iPod touch for example as a very low noise, low power, cheap DLNA renderer with USB output to DAC. Not to mention perhaps easy to implement given the current feature set. Perhaps you could pass on to me the contact details of the lead JRemote dev so as I could have a discussion with him?
Cheers