I am adding to this same thread since I am addressing the same problem, but I am going to describe the situation from scratch as the prior posts may be confusing. Here goes.
I have two ways of playing music through my stereo (Method A and Method B described below). Both work well, but there is a tiny difference with Method B in that it doesn't have small electrical snap problem when pausing or changing tracks that Method A has. This suggests to me that Method A is not set optimally and there might be a buffer or a setting somewhere (probably at the streamer level) that would be more optimal.
For both Methods A and B, I have a NAS which stores my FLAC files. The DLNA is disabled on the NAS. I have J River 26 loaded onto a pretty decent i7 chip laptop that runs Windows 10. I have imported a library into J River from the NAS. At the other end I have a DAC (Schiit Yggdrasil) that feeds into a preamp and power amp then speakers.
Here are the differences in the two setups.
Method A
Here the NAS is connected by ethernet cable to a streamer (and external clock, each with their own power supply) made by SOTM which is connected by USB cable to the DAC. J River is set for network audio with the player being my streamer.
Method B
Here I do not use the streamer. My laptop gets connected to the DAC directly with a USB cable. I am not using network audio here; I guess the way to describe it is that J River on my laptop is the player.
So here now is the issue. When I change tracks or hit pause using Method A, I often hear a small electrical pop sound for a split second. With Method B, however, there is never any such noise. Another way to describe it would be to say the Method B has a gentle fade-out when pausing the music (or changing tracks, which starts with a pause) over perhaps an eighth of a second. Method A has no fade-out, there is an abrupt shut-down of the music with a (small) electrical discharge sort of sound.
So I am wondering how I duplicate the "Method B gentle fade-out" in Method A. From a J River point of view, is J River doing something as the player (renderer?) that I can somehow duplicate by adjusting buffers or something else in the streamer.
The options to adjust are "Audio Buffer Size" (default is 4 MB), "Buffer Time" (default is 0.1 sec), "Buffer Before Play" (default is 0%) and "Period Time" (default is 4 ms). I have only the roughest idea what these mean and what I would really like to see is a "fade-out on pause" time option, but no such option exists (!).
SOTM documentation is weak but I have asked them the same question (with no response yet). I thought that since J River is doing something different in each of my two Methods, you might have an idea of how Method A doesn't cause the problem (i.e. permits a gentle fade-out), which information I might use to address the electrical snap sound with Method B.
All thoughts are welcome! Thanks for your patience.
Tom