Integrated LUFS on streaming services

[BitMallow]

Hi! It is said that most streaming platforms normalize tracks to -14 LUFS; however, I've recorded some tracks from Spotify into my daw and measured them, and they all show very different numbers for integrated LUFS. Actually, none of them are -14, but significantly higher. Could someone explain to me why this happens?

[Vortifex]

You can disable the normalisation on Spotify, are you sure that's not what you've done?

[BitMallow]

You can disable the normalisation on Spotify, are you sure that's not what you've done?

Totally. Moreover, it's not only Spotify, but also other platforms. The measurements for the integrated loudness are usually way higher (and by that I mean louder) than -14.

[osiris]

They are all over the place. Apple and Itunes are the worst. I also find the -1 True Peak very hard to do.
https://www.loudnesspenalty.com/
If you upload your music there it will play it back how the services will encode it. They have a VST and AU version but it's easier to upload and analyze it.

[Opaque]

I think it's best not to obsess over this too much. Yes, there are different normalization targets between platforms, and one can disable or adjust (I believe) that level in the settings of some of them. On top of that, even after something louder (with higher integrated LUFS) is turned down, it will often still feel louder than something mastered to around -14.

From what I understand, it's best to basically master things as loud as they go without ruining the sound, and then check the measurements just to get a rough idea where it is (and it may be a problem if it's something commercial and below -14). Also, to get something really loud, like some of those pop tracks that still manage to sound good and relatively clean, one has to start producing, arranging, and mixing for that from the start (choice of sounds, saturation, bussing, etc.), can't just slam the limiter at the end like a monkey. And, from what I've seen, a lot of those tracks don't even seem follow the -1db true peak principle.

[BitMallow]

I think it's best not to obsess over this too much. Yes, there are different normalization targets between platforms, and one can disable or adjust (I believe) that level in the settings of some of them. On top of that, even after something louder (with higher integrated LUFS) is turned down, it will often still feel louder than something mastered to around -14.

From what I understand, it's best to basically master things as loud as they go without ruining the sound, and then check the measurements just to get a rough idea where it is (and it may be a problem if it's something commercial and below -14). Also, to get something really loud, like some of those pop tracks that still manage to sound good and relatively clean, one has to start producing, arranging, and mixing for that from the start (choice of sounds, saturation, bussing, etc.), can't just slam the limiter at the end like a monkey. And, from what I've seen, a lot of those tracks don't even seem follow the -1db true peak principle.

I think that's true enough. I'm just curious why the actual numbers, not the percieved loudness, but the actual analyzer readings are, for instance, around -9 or -7, whereas the music was taken from a streaming platform where everything is supposed to be around -14? That's what I'm struggling to understand.

[Opaque]

That is weird, yea. Have you checked what your spotify settings are, regarding loudness/normalization or however they call it?

[kritikon]

Almost nobody actually masters at the stated -14 and go far higher than that. Some providers normalize down or say they do, but nevertheless the song is mastered/limited at considerably higher than -14 and the "standard" is completely ignored. Normalizing is not the same as compression/limiting. A while back I spent a bit of time analyzing loads of different tracks on YT and very very few were anywhere near -14. Even sparse-ish pop tunes that you wouldn't think needed squashing that much were mastered very hot. I had some Adele songs sitting at -9 or so. Just ignore -14 like everyone else. Unless your track sounds good, which is what matters. I'd say if you make anything at all with drumbeats, then you're going to have to disregard a -14 target if you don't want your track to sound weedy compared to almost everything else out there. Sad but true

Well, not sad really. It's just the way it is, things change. Dynamics has an awful lot of crap spouted about it. Having -100s of headroom is actually pretty irrelevant if you're making electronic doof doof music that probably doesn't even use velocity on synth notes. Not denigrating, cos I often make that kind of music too. I often don't need expression of any kind on in-yer-face fat synth hooks. . Not just electronic either...what use is dynamics on a metal guitar riff that's been crunched up with distortion so it's flat-lined?

[osiris]

Mine usually run -11 and TP of -0.01 and Spotify has yet to tell me no we can't use that.

[Benedict]

I avoid Spotifry like the plague it is and wish Indie musos also did too as otherwise only supports the damage being done to ALL Music and those very aspirants(sic). However...

On YooBoob if you right-click the vid and choose Stats for Nerds, you can see how much (if any) the bot turned that video's level down overall. Most of my vids, where content is averaged around 0 VU which is (for me) correlated to -12 dBFS, are played back at 100%. It sounds as good as when I mixed it.

It is very common to hear vids that feel trapped, mushy, and other kinds of unpleasant: the Stats for Nerd tell me (as expected) that they content is being played back at around 50% of its exported levels. In reality, the more dynamic material is not only peaking higher - by about 6dB even (several times more louds my droogs)! - but it feels better, brighter, sparklier, easier, more enjoyable... All good things that are lost when master(bait)ing to -0.0001 dBFS.