Ken, this is the volume normalisation that I mentioned in post #171. It all depends on how it has been implemented.
JRiver for example uses the "R128 analysis method which is the international standard for loudness normalization. The R128 standard targets -23 loudness units below full-scale (equivalent to -23dBFS) as the volume reference level. Media Center's implementation of the standard targets an 83dB output at the reference level."
If a streaming service or replay software uses this then there should be no quality loss when implemented. Unfortunately, not all streaming services or players implement it - I could be wrong but I think Spotify uses ReplayGain which has different levels, so you can't directly compare a service using R128 with one using ReplayGain as they will output at different levels.
In any service if you turn off volume normalisation then those tracks that are more heavily compressed will sound louder. And they may even be mastered differently for different services - unfortunately this makes comparing apples to apples nigh well impossible.