Okay, I have definitely found a firmware bug, and it was a very frustrating experience. I was experimenting with different methods to achieve proper bass management with my Denon AVR-4800 receiver and Outlaw ICBM, and decided to try just setting the subwoofer channel trim to +10 dB and leaving the other channels at 0 dB (which combines the redirected base and the LFE at the correct levels in the ICBM). When I first entered these settings in the BDP-83, the calibration tones measured perfectly on my SPL meter, so I was a happy man and powered down my system.
However, the next day when I decided to check the settings again, the LFE test tone was measuring 10 dB too low. I checked the subwoofer trim setting in the BDP-83 speaker configuration, and it was still +10 dB. After many frustrating hours trying to figure out what was happening, I isolated the problem. The BDP-83 resets the subwoofer channel trim internally to 0 dB every time it is powered down (the setup screen will still say +10 dB or whatever). If any change is then made to the subwoofer trim (no matter how small), the output will be correct. I verified this theory by power cycling the player, measuring the LFE test tone (it was 10 dB too low as expected), lowering the channel trim from +10 dB to +9.5 dB, and then measuring that the LFE level had increased by 9.5 dB! I could repeat this experiment every time the BDP-83 was power cycled.
So, to get around this problem I left the subwoofer trim at 0 dB and lowered the trim on all of the other speakers to -10 dB (as previously suggested by others on the forum). This seems to work properly, as the BDP-83 always defaults the subwoofer trim to 0 dB internally when powered off. This firmware bug should be simple for Oppo to fix, but it was certainly difficult to find!