There is one other thing that could cause the phenomenon you are describing. If the OP has used a room correction feature in his receiver or processor, (or the Room Correction module in DSP Studio) there may be "distances" applied to each speaker.
These distance measurements really correspond to delay. The processor adds delay to the channel with the shorter distance, in an attempt to ensure sounds from both speakers reach the listener at the same time.
If these distances are a little bit imbalanced (wrongly, as in deviating from actual physical distances) then the perceived stereo image can be widened or contracted.
However, if the distances become too imbalanced, then a psychoacoustic effect called the "precedence effect" comes into play. This is a form of temporal masking, where the brain deemphasizes the signal that arrives later and focuses on the first sound that arrives (this is a biological form of echo cancellation that aids in locating the source of the sound). This will be perceived as a lowered volume of the sound from the speaker with greater delay. The listener will need the volume of the delayed channel boosted, by perhaps 3 to 6 db relative to the other channel, to perceive them as having equal loudness. This is not an audio system malfunction but an innate human response.
So you might want to check your speaker distances.
There's no inherent flaw with digital audio with regard to stereo imaging being affected by small amounts of volume change, other than humans generally perceive louder music as being "better".
Author
Topic: What are the different ways to reduce clipping in JRiver? (Read 1259 times)