I just did take a Studio Premier Qobuz trial, and listened to its offered 24/96 version. It sounds considerably less dynamic than the lower rated 44.1/16 Tidal version. Through Roon. I had to put in +20dB to match sound level.
There’s one version on Tidal in 44/16 labeled “Remastered… released in 1990”. There are two versions on Qobuz, a 44/16 Remastered from 2004, and a 96/24 version released in 2014. The two 44/16 versions are in fact identical.
Here’s the kicker though, the Qobuz 96/24 release is considerably more dynamic than the 44/16 versions, you might be confusing loudness with dynamics
Let me show you objectively (via Adobe Audition). Images: Tidal 44/16, Qobuz 44/16, and Qobuz 96/24 respectively.
The 96/24 is much lower in volume than the 44/16 on that album, I have both version i nQobuz and this might be one of the biggest changes in volume I have heard on “the same” album between different masterings.
Much of that 96/24 “Heaven or Las Vegas” track is mastered with peaks at about -6dBFS, with just a few peaks reaching as high as -3dBFS, leaving I believe plenty of headroom for upsampling D-to-A conversion without risk of Intersample overs.
By the way, the 96/24 version has a measured dynamic range of 127dB while that highly compressed/loudened 44/16 version is 59dB
As a general observation with digital recording where the maximum attainable level is fixed at 0.0dBfs, the wider the dynamic range the lower the mean volume level must be, ceteris paribus.