Singxer posted earlier in this thread they will remove this jumper from future amps, since a DAC with considerable DC offset combined with cap bypass / DC coupling, could damage upstream gear. The only effect that could show up would be below 20Hz and it would be tiny. Those 'improvements' people write about are surely in their head lol.
The effects of the 330uF Nichicon bipolar electrolytic caps will be heard across the frequency band.
In this design, they are also bypassed with some cheapo 1uF film caps but its always best to never have any AC coupling caps in the circuit !
My Oppo source already has output coupling caps so removing the secondary caps in the SA-1 was a no brainer.
My previous amp was DC coupled by design so I'm always aware of the issue (don't want constant DC at the headphones).
if your source has more than +/- 5mv or so then keep the caps in the circuit (especially if you use high gain).
I only noticed a small difference but did not do a direct A/B comparison (Balanced in/out & low gain setting with volume at about 9:00).
I also soldered the pins together since I would never use those cheap PC jumpers etc.
One last thing is that the coupling caps are on the wiper side of the 10k volume pot Vs the input ?
The caps may have more influence on the sound if you have the volume pot generally set to a lower level.