One thing I wanted to add that MIGHT help/point to a direction regarding popping only at the beginning. With the IanCanada stuff, I believe I read somewhere that it has to do with sample rate switching, as the popping only occurred at the beginning (I don't recall it happening at the end of the track, unless it required a sample rate change as well). Based on this, I believe one solution was to insert a few milliseconds of silence via player software (I used JRiver) so it would "play" silence as the board switched sample rates/clocks.
Edit - another point to make is that the popping only happened when using i2s out of the IanCanada, and did not occur when using USB (from same Pi) > Matrix SPDIF 2 > i2s. I SUSPECT it MAY have been something to do with the rudimentary circuit of the IanCanada boards. As Pi boards do not have an MCLK output, an intermediate board is used between the Pi and the i2s transport board (FifoPi). The FifoPi uses two clock chips for the different base clocks frequencies. This differentiates from the i2s out of the Matrix - not sure what the design difference may have been, as I never tried opening the unit up to see.