Absolutely!

Right, but if the song is already at -10LUFS, it most likely isn't going down any further. So, there isn't really any point sitting through the whole thing. The only time I do that, is if I am mixing my song and want to make 100% sure where it is. For this, he's just making a point.

@@musicwithbrad you're assuming the song has no dynamic development. If you measure short-term LUFS in a loud bit then the real integrated LUFS may be considerably lower - and vice-versa.

@@elmoreastro2929 Exactly. Inegrated LUFS can only be acquired by measuring the whole song.

Yeah, but if you have for example a track that has build up and breakdowns of 1 minute long and it's pretty quiet (-14lufs) and the chorus goes up to -5 lufs and it's only 20 seconds long, then by measuring the full song you won't get an accurate representation of how loud each section is.

What peaks were being subtracted to begin. That part confused my. The RMS peaks or DB peaks or what was that?

👌correct. However his method does give a good ballpark, Quickly. Granted its not the "correct" way to do it. An i agree with you here that for the purpose of the video that would have been a better choice. As more people are familiar with measuring integrated that way.

@@Unders Sorry, I have to disagree with you. It very much does not give a ballpark figure because he is only measuring the loudest bits. This is why he thinks mastering to -14db LUFS is wrong when it isn't. Measuring the loudest bits gives him much too high a figure which is why he thinks other people are not working to -14 when they probably are if he onlly measured them correctly!

yes it's a bit misleading. still most people take the short term LUFS and use the loudest part of the song as a reference for the whole song. that's what a plugin like ozone will do if you use the master assistant, it will only listen to a part of the song. so this is sort of a global misundestanding here.

Thank you. I was scratching my head watching this "fake news" as well. But I'm just a B-level bedroom artist, so what do I know compared to this professional.

But Izotope's article comes to the same conclusion. Are you suggesting that Izotope doesn't know how to measure integrated loudness?

speaking of clipping, any reference track I pull in is +12-15db in the red in rms and -6 or so on the lufs meter. the rms makes sense but this guy is right on lufs

-0.30 db always and forever baby

no, just before his own measurement, he went on the website where we can see again -10 / -6

BTW this comment is right. You don’t measure LUFS over part of the song, the quiet bits are included in the calculation. That’s how streaming services do it and it’s in the standard.

@@RobHarrison it's disgusting on youtube, there's -14Lufs normalization, and it still have gap between music uploaders, some tracks are quiet, and other loud..... my track is -14,8Lufs and is quieter when i compared it with some artists, famous or not

@@Fritz2824 would you like me to take a look at your source audio wavs to see if I can identify cause of issue?

Obvs KZfaq won't turn your volume up so that 0.8 is not going to come back but 0.8 isn't that different and should not be so noticeable to cause you concern.

If your mix is louder than targed level of streaming service, they will only be lowered in volume, nothing more. There will be nothing to cause distortion, or degeneraton. That alone will not affect quality of original, so there you are good to go. Transcoding may, but should not. The client, on the other hand is the one who ultimately pay your bills, so you do as they say, but mastering -8 vs -14 LUFS alone only makes one louder if you play them side by side. On the other hand if you lower -8 to -14 and play them side by side ( thats what streaming services do ) you have a lot more potential in your -14LUFS mix, because you have a lot more dB of headroom. It is up to you, if you use this potential or not, but it is there to be used. Client on the other hand knows best, and if client wants louder, louder it has to be.

​@@marcinfilip7405but if both songs peak at 0, but is -8 LUFS and one is -14, and let's just assume it's the exact same mix, the -8 master is more compressed, so even when turned down in volume to be at -14, it'll sound more like a "wall of sound", while the one mastered at -14 will have more dynamics. And depending on the song, one night be perceived as louder than the other. Could be the dynamic one with punchy drums, but also the crushed one with massive synth scapes.

Lol, its a fact bro...

askers?

Yup, when songs are super squashed they fatigue the listener much quicker. Same as youtube videos that are edited with ultra fast cuts. Within a minute or so I cant even follow whats happening anymore and just switch to something else.

Agreed. All these songs he went through are top of the charts modern pop. You think they care about dynamics lmao? I prefer breathing room in my music, so a -12 is where I'll be headed unless I'm mastering for a specific type of music

This is a great option! You can press your CDs using your -8dB LUFS master, and upload your -14dB LUFS to the streaming services. Radios play from CD so you'll get the radio mix for free. You can provide B side mixes at -14 so your CD customers can still get a decent quality listening experience. Just call the A side the radio edit, or whatever.

that's what I'm getting from this; the streaming service will analyze the whole song; and use the integrated LUFS for the whole song to determine the "penalty". Which means the loudest part will be higher. Still it means we can have a bit of dynamic inside tracks and have loudest part at -9db or so and quieter at -18db or so.

They are using the integrated LUFs which is basically taking the average of the LUFS over the whole track from beginning to end. Momentary LUFS is the instant loudness number that will fluctuate quite a bit. But I agree with what you are saying about dynamic range. Everyone is talking about LUFS-I and Loudness, but I think the more important number is Loudness Range, which if you look on the pro examples he had here on this video were all between 5 to 10 LU and up, and that seems like a pretty impressive numbers to me because that is the one number I struggle a little with getting that high. I know my mixes are not squashed, I mean none of them sound like Death Magnetic or anything, But I think there is still room for improvement and that last little 5% is the hardest to get and usually the thing that separates great from good.

Your thinking is off. The izotope article showed the dynamic ranges of those songs.

Exactly right Mastering to -14 LUFS / -5dBTP is no different than -10dB / -1dBTP; the crest factor is exactly the same; the LRA is exactly the same. This is literally what subtracting peak from LUFS in the video is unknowingly (or cynically) acknowledging: crest factor doesn't change if you simply turn it up

@@aholder4471 laughing at your Death Magnetic reference. That album really is a prime example of ridiculously bad squashed mastering. A shame, as it was musically decent.

I have noticed that when I measure the LUFS on an unmastered song (no master bus compression or limiting) and normalize the volume to around -1 true peak it naturally ends up very close to -14 dB LUFS.

they definitely aren't; look at the difference in peak in the two tests.

Nailed it. I don't understand why people make this more complicated than it needs to be.

I dont think music under -14 lufs will be turned up though. In fact I know it because classical music on streaming platforms is still very quiet (because it t3nds to be mastered under -14 lufs)

@@joe4570 It will be turned up to -14 LUFS, or as much as the headroom allows. Here is an excerpt from Spotify's mastering guide: "If a track loudness level is -20 dB LUFS, and its True Peak maximum is -5 dB FS, we only lift the track up to -16 dB LUFS" Adding to the example above, if the original true peaks had been at -7 dB or lower, the track could have been turned up to -14 LUFS.

@astralbrain8265 presumably everyone is mastering to at least -1 db true peak though. So yes you're correct but if we assume everyone is mastering to at least-1 true peak then -20 stays -20

@@joe4570 Correct.

What do you use on your master bus?

Happy to help!

Of course. Just try to post it there. Then listen it in your pc and compare to the uploaded one. Youll see for yourself.

they were obviously taken from Apple streaming service; as they had a peak at -2.5 or so. The service did apply a "penalty" but I'd say the -14 dB standard takes the whole song into account; not the loudest part; so the penalty would be a lot lower than the difference between the loudest part LUFS. or maybe there's something else going on as well I'm not sure.

DSPs only turn down your track IF the listener is using NORMALISATION

---- Garbage for years during the loudness wars era / 💯 ❤ !!!

And by the way, who wants to emulate the crap the music industry is putting out? Their stuff would sound a lot better if they didn't squeeze the life out of it. But, these days, for the most part, people don't care about crappy sound quality. If everyone else sounds like crap, what's the incentive to be different?

what do you mean with "If your Dynamics are in the green most of the time than in the red ...." why should they ever be in red?

@@snubdawg1386 MAAT makes a very good that will help when doing mixes and because it allows you to monitor your dynamic range. I will include a link so you can see how it works and you will understand what I mean UCoLCmff5dDTYVTdg4fJV4Bw

Why do you think your masters don't get turned down? Do your songs on, say Spotify, sound roughly equally loud as everything else? If they do, which I'm sure they do, they have been turned down to -14 LUFS as everything else.

@@astral_brain if your masters have good dynamics then when it gets normalize but the streaming services it will still sound loud. It is important to have a meter that monitors your dynamics. So when you are adding a limiter of clipper or both you can keep an eye on how the dynamics are being affected as you increase volume and depth. MAAT meter is perfect for this

@@astral_brain kzfaq.info/get/bejne/p7imndtmxLnUnmw.htmlsi=dGxZDL6OCDCblIl7 a link about that MAAT meter plug in

i use pro L-2

This is actually what many releases seem to do now and historically too. Back in the day you'd never see the same master being used for album and single releases or for use in cinema or video backing. MP3 you really need 24bit to get the best quality but for CD you have to dither to 16. If I'm listening to the album I don't want ear fatigue at all or I'll turn it off but if I'm listening to a single then I can stand a bit more fatigue cos only a few minutes long.

Set up a re-record from Spotify feed and compare the loudness of your uploaded -14 song against other similar songs. You will probably find it very soft in comparison. The bit of compression that limiting adds makes a mix sound fuller and louder, then there is headroom, EQ, True Peak, etc. Point is I’ve uploaded the exact same track to Spotify thrice, every time getting louder, and it literally does play through Spotify louder as well per documented re-recording. And that’s re-recording with normalization on btw.

Use clippers and saturation

Okay; @@bllymusic9614

Lol

-8 is not THAT loud lol, any good mix will easily get -8.

@@heddshot87 -8 means you have no dynamics.

All these tracks I tested were from Apple music streaming charts. I didn't test CD.

​@@DavidCamposComposer Don't streaming services normalize tracks? I was under the impression they analyzed the whole song and took an integrated approach to how they would normalize. I think what BassBusMusic is saying is that you would have gotten a different result had you bought the CD like Izotope did.

There also are settings, at least in Spotify that allow a listener to bypass their normalization. Which could be the case in his tests, I don’t have Apple Music so I can’t test it myself

Again, it's not 3 seconds he is fast forwarding through the video of the measurement, it's very obvious the songs don't sound like that and he is fast forwarding through the measurement so he doesn't get struck...

I would trust the Izotope article more than the meters of a guy running Apple music into his DAW via a soundcard and his own faders. If you buy the CDs or digital files of the chart-topping pop songs, they really do have LUFS values in single figures. The streaming services then turn them down to -14 LUFS so that they don't stand out as much alongside random songs on a random playlist. But the top pop songs really are compressed and limited to extreme levels, such that the actual masters (and CDs and WAVs) sound much louder and have much less dynamic range than music of the past. The point is that professional pop music is really loud. Spotify will turn it down, but the main reason our amateur tracks don't sound professional is that we don't compress, clip, and limit it as much as the pros do. A master at -14 LUFS just isn't compressed enough for the pop charts.

Nice take. What do you think our true peak levels should be for our masters. I was surprised to see some pro songs with +1db true peak when we are told to be -2db true peak from Spotify. I wonder if the mastering engineers are bouncing out separate versions of the masters, one louder for cd quality and one quieter for streaming services like Spotify and tidal.

I guess they just let it clip@@MaxRayMusic

I heard another KZfaq channel saying that the top engineers don't really pay much attention to True Peak. IDK if that's true but I'd like to find out.@@rmv9194

lmao at the last sentence.....

it will turn down the volume at playback depending on the user settings

I'm not really into mastering rock music but generally speaking, a premaster at -11dB LUFS seems very hot to me.

A brickwall limiter at the end is a good idea to catch intersample Peaks which can cause clipping. Transients might get a little softer that way, so it's always an individual decision. If your true peak sits at -1dB you should be safe.

if u ever touch a fader on a mixer or master channel in ur DAW just make sure it never goes above 0db aka "100%" on the fader. otherwise u would get both distortion and clipping

You can turn Spotify normalisation off in your Spotify settings (just Google it). Or set it to loud mode and everything will be normalised to -11 Lufs. So he's got the point. iTunes on the other hand does not allow users to turn off the normalisation (as far I know). The only thing is matter for iTunes is -1 dB true peak.

That's not even remotely what he's saying. At all. He never said anything about streaming services' normalization practices. He's analyzing hit songs and seeing what levels they're mastering at. It may be "incomplete" to only analyze a sample, but it's extremely unlikely that a song with a LUFS of -7dB at its loudest will be anywhere near -14 integrated. It may be a bit lower, but no lower than -9 or so. The point isn't that streaming services aren't going to normalize a track above -14LUFS, it's that mastering to -14 for the sake of matching the streaming service isn't necessary and possibly detrimental depending on the track.

@@ParanormalArson "but it's extremely unlikely that a song with a LUFS of -7dB at its loudest will be anywhere near -14 integrated. It may be a bit lower, but no lower than -9 or so." that's extremely dependent on the song; but it would be cool to have a proof of that.

@@valdir7426 I mean, it's really not, with the exception of MAYBE Classical and Jazz. In pretty much any form of "popular music" (i.e. Rock, Metal, Pop, Hip-Hop, EDM, etc), if you have a greater variation in dynamics than that you likely need to change something in your limiting or compression.

@@ParanormalArson this Billie Eillish song has 10 Db of dynamic over the song; I'd be curious to know what the average is over the whole song. Anyway it's clear that the streaming service does lower the level but from -8 Db LUFS or more for the whole track to 2,5 lower than that makes it -10,5dB LUFS; so how do you account for the remaining 4 DB? that's what I'd like to understand. so integrated LUFS is part of the explanation; but if it's not enough; what is it?

I´ve just checked that with a song I mastered for youtube and they didn´t raise it up. A shame cause I was couting on that. I took it for granted... It was my fault anyway. Second step then I compared my master to an oficial Post Malone video music and they definitely lowered it down quite a few dBs to lock to -14 LUFS integrated. So that´s it. The information this guy has given is not false, but the streaming services in fact lower down the songs that have beyond ceiling LUFS. It´s all your choice to work a master louder than you really need.

@@franciscoesmanhoto8445 KZfaq only applies loudness normalization on some devices, desktop/web player usually but not always, mobile usually but not always, smart TV's usually not.

So stupid, digital has an 80 db + noise floor, if you want it louder turn the volume, louder with the greater dynamics of less compression actually has a greater impact as the difference between the quiet and loud sounds are greater.

@@DerSilvano wdym?

What's your true peak on your limiter set at? For streaming, youtube/soundcloud/spotify you should set your limiter on your master bus to -1 dBTP. This should allow enough headroom so that your audio doesn't clip/distort during the transcoding process when you upload for streaming services.

the article and the test check the loudest part of the song. the standard is presumably for the whole song integrated. He did test songs from apple streaming services with already -2.5 db or so of "penalty".

@Beizeiten78 I agree, there was a Bruno Mars song, same deal in the red distorted way out of balance with the bottom end and it sounds like crap I could not believe they released this and someone said yup its good to go, As stated well many of the pros do it this way. Believe me many pros do not think it sounds good, they are just doing what the producer and label asked of them. I remember Bob Ludwig a mastering legend, said he would make 2 versions one how he felt it should be done and one that was asked of him so the client could then make the choice from there. I absolutely hate the loudness war mentality, loud is not better, sure if its done right it can sound passable but the true dynamics if the song had them will just no longer be there in the way it was intended. Take some classic music like Steely Dans Aja as an example ... the music is allowed to breath and have its dynamics kept intact, if this was mastered to -8 or some other insane LUFS it would sound like crap at least to those who know what it should sound like and are listening on a real system. So while yes its dependent on the music I do not like this whole every thing must be ridiculously loud if this one jumps off the bridge we must all jump of the bridge. Call me old fashioned but that's what your volume control is for you can make it as loud as you like. As for SAGE Audio that was ref in this video I have to say one thing and I have no affiliation with them at all, but have listened to his work and to my ear his mastering sounds fantastic open and punchy not some over compressed heavily limited mess, way more natural then most of the big release crap in my opin.

I totally agree with Beizeiten78; loud definitely doesn’t mean better. Has the loudness competition gotten so bad that we just put stuff out there without considering dynamics? What I here is a bunch of over compressed loud bass saturated music. Loud does not particularly make your music stand out. In my opinion and my opinion only a great song and arrangement makes the song. I’m not saying make the song where you need a hearing aid to listen to but at least in the competitive ballpark. I have no problem with -10 or 12 even. Tracks can still be loud enough but maintaining clarity and dynamics.

Glad it was helpful!

Thank you Carlos 🙏

Because the assumption is that the original song was mastered at 0 and it assumes that the streaming service turned it down by whatever the peak is. In reality it was probably mastered at -1 db to account for inter sample peaks so lufs - peak+1 is probably the right number.

The relationship between dB and LU is not linear, close but it is not exactly the same thing.

Glad it was helpful!

-12 LUFS-I is a good balance between loudness and dynamic range. Mastering engineer Ian Shepherd recommends -10 LUFS-S (short term, loudness over 3 seconds) peak during the loudest part of the song as a starting point. That will usually end up around -12 LUFS-I for the whole song... MOST commercial music is much louder than that, but it actually wrecks the dynamic range to do that, and more and more people are waking up to that. So your university's advice is good for fidelity, but if you do this professionally you'll deal with annoying bands/managers/producers that want you to master louder.

@@CyanideLovesong I asked one of the lectures about this, and in short he said exactly that, although, he did mention when he sends stuff off to likes of abbey road they come back generally -9LUFS . He said recently he was struggling to get a piano piece above -10 without destroying dynamics , he sent it to a guy at abbey road and it came back -8LUFS and sounded amazing (as would be expected).

@@SolarPeel Yeah, in the end it's subjective. Also what numbers are we talking here? When you say -9, -8 ... Is that LUFS-Integrated? Or LUFS-Short Term at the loudest parts? There's a big difference. Here's an example of a song that wouldn't have the big dynamic bass that it does if it was more smashed: kzfaq.info/get/bejne/fraEdcSBv-CWZ4E.html

@@CyanideLovesong LUFS integrated. I think he said you can have short term hitting -4/-5lufs , but I might be getting that wrong! Sure that’s what he said tho. And cool I will check that video out - thanks ☺️ edit: wonder if It’s worth taking the song and putting it through a compressor to hear the reduced dynamics.. might as well hah

@@SolarPeel Wow, that is in hardcore loudness war territory and I personally prefer music to be more dynamic. I think Ian Shepherd's advice is best, but he masters more dynamic than most engineers.