Several people have already hit on the situation quite accurately.
The digital processing in the XMC-1 is quite "deterministic" - there are very few situations where two different XMC-1's are going to play the
EXACT SAME digital audio stream and behave differently. (Most of the digital processing is essentially a computer running a program; if two different computers run the same set of calculations, and neither is outright broken, they will generally produce the same exact answer). However, depending on the digital audio you're playing, and on your source equipment, the digital data stream itself may be very different in several ways.
Most albums and CDs never actually have "pure digital silence" - because dither is normally applied to digital audio - including the track gaps. And some player programs will play audio continuously - without stopping and starting the data stream, while others will fill the gap between tracks with dithered silence, or with true silence; some will allow the clock to continue between songs (the two songs are played as "one track" so there is no gap). And some players will do one or the other of these depending on what options you have selected (jRiver has one mode with track gaps, one without gaps, and one where tracks change using an overlapping fade). The exact mechanism will also depend on the sample rates, and on the mode you have selected - for example, if you're playing in WASAPI mode, and a song follows one recorded at a different sample rate, then the audio stream
MUST stop and restart - because it must change sample rates. And, to put it bluntly, while we know most of what's going on inside the XMC-1, we didn't write every line of code used by the DSPs, and you probably don't know exactly what your player does under all conditions either.
I can tell you for certain that the output "hard" muting on the XMC-1 is activated when the XMC-1 is turning on and turning off, it is operated by "physically" shorting the output to ground, and it does
NOT enable between songs or tracks. I can also tell you that the DACs we used in the XMC-1 do in fact have an option to "mute on repeated digital zeros" - and that we have
DISABLED that option. I can also confirm that, at some point in the signal path, digital signals are still muted under some conditions when a series of digital zeros are received. This happens somewhere inside the XMC-1's DSP code - which we didn't write and have no direct control over. (And the reason that most of you never encounter this is that, most of the time, a digital data stream will never contain true zeros rather than dithered silence. However, we have no way of knowing, for example, if Sonos has chosen to fill blank spots with digital zeros when certain settings are in effect.) You will also find that many advanced players have options that may address this issue. (For example, when you RIP CDs with jRiver, it has an option to include a specified number of milliseconds of silence at the beginning of each track.)
However, to return to the original subject, since this is a detail of how the XMC-1 responds to a certain specific structure in the digital audio data, it is neither something you should expect a different XMC-1 to do differently (under exactly the same conditions), nor something that we can "fix" either in your current XMC-1, or by giving you a new/different one. We do continue to review and improve the code that runs the XMC-1 whenever we see an opportunity to do so, so we may in fact eventually succeed in further reducing such issues, or even eliminating them entirely, but in the mean time we can't simply "fix" it... and we have no hardware update which will make it go away.
Keith, is this issue in the new units like the XMC2/RMC1?
Thanks
Darin
