Anything with DD decoding should really have the equivilent of night mode - it's a built in part of the DD spec and the reason for the dNorm value existing.

Night mode should compress the dynamics using the dNorm value encoded in the data stream as the normalization point. The entire idea is there so that you can reduce the dynamic range without grossly altering the perceived average volume of dialog.

It's not always called night mode though. Dig around - it might be in the surround decoder settings or something.