I'm not sure if it's because I'm doing it wrong, but I deliberately set my ReplayGain compensation to be fixed a little bit higher than most of the tracks I've got (+18dB, most songs I've got average between -10 to -15dB for replaygain) and then set Internal Volume to be less than full (in my case -7.5dB). I am occasionally able to hear clipping (or well, what sounds like clipping to me), so I did some investigation.
The audio path shows resampling first (as dictated by DSP studio too); Volume levelling is applied (+18dB, which will make the nett volume change to be +2 or +3 dB); then Internal volume applied (-7.5dB).
Is this the correct order? I'd assume that either JRiver will combine both values together to give an overall volume change before applying it so that set ups like that won't cause clipping. Or you could tell me that the algorithm is indeed correct and that my ears are playing tricks on me...