Reading through this thread, there seems to be a lot of confusion as to what the differences may be in regard to encoding vs. mastering. I have a pretty big stake in finding an answer to this, and I think it's worth the time. For the most part, I switched to streaming, even when I have PCM files sitting right next to me locally. Streaming has become my access point for most music- it's just easier, and I have the kid in a candy store feeling when I can access almost every album ever made. With this in mind, I put aside my negative feelings about Tidal's associating with MQA and recent Jay-Z acquisition, and got a trial subscription for a month.
I began by capturing a WAV from both Spotify @ "premium" (320kbps 16/44.1 OggVorbis) and Tidal @ "HiFi" (16/44.1 FLAC) quality. I sampled Philip Glass' Sand Mandala from Kundun, Bach's Matthaus Passion by Ton Koopman, and the Beatles "Let It Be" Remaster. I tried my best to find the same masters of each, but can't absolutely guarantee it. With the soundcard in 44.1khz mode, I created an Audacity project @ 44.1khz, and captured streams from both music services. I then output both streams independent to a 16-bit/44.1khz WAV file, as to retain all data possible. My first step was ABXing. I could not differentiate beyond statistically random. This led me to approach things from a more concrete perspective, a nulling of the two WAV files I attained from each stream using the audio processing software Diffmaker. I have decided to post the files so that everybody can listen with their own ears:
Tidal HiFi vs Spotify Premium Null track: Philip Glass, Sand Mandala, Kundun -
(edit: Please do not adjust your volume for this test, the purpose is to compare noise and artifacts to regular listening levels. I should also warn you that there is a slight overhang at the end of the clip where the null drops off and the monks return, so please don't blow your ears or speakers out by turning up the volume.)
http://s000.tinyupload.com/index.php?file_id=86139679279655796028
Consists mainly of low frequency oms and didgeridoo. Doesn't seem to present any issue.
Tidal HiFi vs Spotify Premium Null track: Bach, Matthaus Passion, Ton Koopman
http://s000.tinyupload.com/index.php?file_id=51110490042415488335
Low level noise, seems to be amplified when the chorus comes in toward the end. Wish I had captured more of the chorus, but was worried about file sizes.
Tidal HiFi vs Spotify Premium Null Track: Beatles, Let It Be, Remastered
http://s000.tinyupload.com/index.php?file_id=00880002567388396767
Some low level noise around the vocals, but I don't know what to make of the extra piano hit. There are two remasters of Let It Be, from 2004 and 2009, and because Spotify and Tidal both list the albums' original 1970 release date, there's no way to know for sure which is which. Probably a mastering difference.
I hope this has been useful. To me, according to the ABX results, the differences based on encoding do not appear to audible enough to justify Tidal on their own. However, it does seem to me that certain masters on Tidal are of better quality. One example is Fleetwood Mac's Rumors, which only appears on Spotify as the brickwalled and distorted "Deluxe" version. Tidal has a standard non-deluxe master that sounds much better. Philip Glass' Koyaanisqatsi was in its original form on Tidal, whereas Spotify only had an inferior soundtrack version with lots of hiss. I was also able to find a Till Bronner album named Oceana on Tidal that I have been missing since the days of MOG. On the other hand, Tidal was missing the Beatles Anthology, which I never purchased due to price, and enjoy having access to on Spotify. And Jay-Z's obvious marketing influence over Tidal, along with MQA, makes me want to throw up in my mouth a little. MQA is very disturbing to me, and I find it difficult to give money to a company that promotes it. I am still in the process of deciding what to do, because I do think that Tidal's library is curated with better masters. I may do further testing at Tidal's reduced 320kbps ($10/month) quality to see if it contains any marked differences from their 16/44.1 FLAC. But even at $10/month (same as Spotify Premium) the moral qualms I have with MQA are deep, and are the biggest hurdle in me choosing Tidal.