dennistdk
100+ Head-Fier
- Joined
- Aug 9, 2015
- Posts
- 161
- Likes
- 97
So... I really don't want to pay for two streaming services. I like Spotify because it's integrated with Carplay, my Onkyo surround amplifier in my living room and almost any other device I have.
A year or two ago I started spending (too much) money on CIEM's, headphones, dacs and amps as most people on here. I wanted to get the best sound possible, so I switched to Tidal Lossless believing I could hear a difference.
For practical reasons I went back to Spotify about 6 months back (ease with CarPlay being the primary reason). But ever since switching back to Spotify I just thought I could hear some things missing. Like less dynamics, less sound stage etc. Probably all in my head.
So I switched to back Tidal again. I did a lot of testing between the two services (not blind, but just a lot of 5-10 seconds switching with each player with many different tracks). First of it's really hard to volume match - and I always picked Tidal as the best sounding one. But again - it could just be because I knew it was lossless.
And let me just say that it is pure hell switching back and forth between streaming services when you have tons of playlists, favourites etc.
Anyways - I really wanted to do some ABX-testing with the two services. I know that I would probably fail an ABX-test if I tried doing it myself with own high quality encodings (haven't tried it yet), but I often read on forums that people suspect Spotify of not using high quality encodings (some suggesting that the 320kps ogg are just re-encoded from a lesser bitrate when they introduced the premium subscription). So I needed to grab some samples from the players to do the blind test, but before that I wanted to inspect the grabs...
I just did the first "inspection" (and replicated it to be sure I wasn't wrong). And compared it to the CD version I had. The tracks should be the exact same ones (same version/mix/production/master/recording etc.).
The audio was recorded from Tidal and Spotify using Audio Hijack and compared in Audacity (all free programs, so anybody can try).
So based on the initial findings I'm pretty sure I'll stick with Tidal for now, but I need to run a lot more tests with several different tracks as this one track (won't say which yet) could just be a bad rip from Spotify.
Some pictures:
First the sample length difference. Tidal on top, Spotify and then the CD.
Then the loudness (taken from a non-peaking part). Tidal is blue, Spotify red.
Any thoughts?
I'll try on some more tracks over the next couple of days doing the same tests. Maybe also test with Tidal's 320kps.
A year or two ago I started spending (too much) money on CIEM's, headphones, dacs and amps as most people on here. I wanted to get the best sound possible, so I switched to Tidal Lossless believing I could hear a difference.
For practical reasons I went back to Spotify about 6 months back (ease with CarPlay being the primary reason). But ever since switching back to Spotify I just thought I could hear some things missing. Like less dynamics, less sound stage etc. Probably all in my head.
So I switched to back Tidal again. I did a lot of testing between the two services (not blind, but just a lot of 5-10 seconds switching with each player with many different tracks). First of it's really hard to volume match - and I always picked Tidal as the best sounding one. But again - it could just be because I knew it was lossless.
And let me just say that it is pure hell switching back and forth between streaming services when you have tons of playlists, favourites etc.
Anyways - I really wanted to do some ABX-testing with the two services. I know that I would probably fail an ABX-test if I tried doing it myself with own high quality encodings (haven't tried it yet), but I often read on forums that people suspect Spotify of not using high quality encodings (some suggesting that the 320kps ogg are just re-encoded from a lesser bitrate when they introduced the premium subscription). So I needed to grab some samples from the players to do the blind test, but before that I wanted to inspect the grabs...
I just did the first "inspection" (and replicated it to be sure I wasn't wrong). And compared it to the CD version I had. The tracks should be the exact same ones (same version/mix/production/master/recording etc.).
The audio was recorded from Tidal and Spotify using Audio Hijack and compared in Audacity (all free programs, so anybody can try).
- The Tidal version is 100% the same as the CD version. I found no differences at all.
- Even though both Spotify and Tidal was at 100% volume - the Spotify track is lower in volume. Tidal has the same volume as the original CD.
- The Spotify track peaks at -3db Tidal is at 0db (same as CD).
- The Spotify track is further 3db lower on the right channel (so peaks at -3db left and -6db right). Don't know why there is a channel difference, but it is there.
- The spotify track looses 1-2 samples per 30 seconds (when aligning the waveforms Spotify is behind) - maybe they are using a different sample rate or just an artifact of the compression or the recording I did (however it was the same 2nd time I tested).
- When looking closely at the waveforms at non-peak level the Spotify track is usually "louder" than the Tidal track even though the Spotify track peaks below the Tidal track. This can almost only mean that the Spotify track is further loudness compressed loosing some dynamic.
So based on the initial findings I'm pretty sure I'll stick with Tidal for now, but I need to run a lot more tests with several different tracks as this one track (won't say which yet) could just be a bad rip from Spotify.
Some pictures:
First the sample length difference. Tidal on top, Spotify and then the CD.
Then the loudness (taken from a non-peaking part). Tidal is blue, Spotify red.
Any thoughts?
I'll try on some more tracks over the next couple of days doing the same tests. Maybe also test with Tidal's 320kps.