Hmmm... maybe they are both using the bloody apodizing filter then, ripple spec +/-0.075 dB. FAIL.
Both DACs could do much better in this regard - the ES9038Q2M has a whole slew of attractive filter options to offer, from a traditional linear phase fast roll-off (+/-0.002 dB, 120 dB ultimate) over minimum phase fast roll-off (+/-0.005 dB, 100 dB, pre-echo no concern but group delay deviation possibly is) to brickwall (+/-0.015 dB, 100 dB starting at 0.5fs already).
This should be brought up in the D10s and Tone Board threads, as a firmware update may be feasible.
CS4382 (single speed) digital filter ripple should be about +/-0.008 dB according to datasheet graphs, with a 90 dB ultimate, so that's not even anything super-duper special.