Most of the reviews I’ve seen on sound quality RPI3 vs RPI4 vs another much more expensive streamer with ‘audiophile’ digital outputs have been done with sighted evaluations i.e. the reviewer knows what equipment he’s hearing. And so the expensive streamer always wins! When done as true blind test with a good USB dac I would be prepared to bet that a set of trials would not give a statistically significant preference.
Measurements with high grade analysers such as the APX555, show no difference even when looking for differences significantly below the threshold of hearing.
Look at websites like Audio Science Review or Archimago’s Musings where some of the myths about digital audio, overpriced cables etc are laid to rest. Archimago did offer some useful thoughts on reducing cpu noise on a RPI though. He suggested reducing the clock frequency, which then allows a reduction in supply voltage to the cpu. A couple of his postings
The content does assume a certain level of knowledge and so may may not be very easy reading so it is for this reason that we get subjective reviews of ‘what it sounds like’ which are much easier to understand, are much easier to do as a blog, and are at the best much more prone to error!