If LUFS is actually a thing on Spotify, then why do...

[JKN]

...why do so many recordings on Spotify not sound normalized? So many tracks (especially older ones) are significantly lower; even new tracks often very cut to cut (at least, album to album). On any given playlist or random skipping around, I'm forever adjusting the volume of many, many songs. So, if LUFS is actually a normalizing thing, why do I still have to keep adjusting levels? And given the constant need to adjust, does submitting to Spotify at -14 LUFS truly matter? I'd appreciate any insights that can quell my confusion.

[RichardT]

LUFS is quite a new thing. It was first defined in 2010.

Many old tracks prior to that are mastered at much less than -14 LUFS, and some from the 90s and later are mastered much higher than -14 LUFS when highly limited tracks became the norm. So historically there is massive variation in loudness.

But even for new tracks, a lot of people still master at much higher LUFS than -14 for various reasons.

If you listen to loud tracks with normalisation on, Spotify will reduce their volume until they reach -14 LUFS. Because they have very limited dynamics they will sound relatively quiet.

Very quiet songs will be raised to -14 LUFS but only if their peak volumes stay below -1 dBFS. So very dynamic old material will not be raised that high because it will reach -1dB peak long before it reaches -14 LUFS. So these will sound quiet too.

Just to check - you are listening with normalisation on and set to -14?

-14 is just a guideline for releasing music. Up to about -10 LUFS to my ears, music sounds OK. At higher LUFS it sounds dreadful. So unless you’re going for that kind of effect, I’d say don’t go above -10 LUFS.

On the low side, the level that’s most appropriate depends on the material. If you want to capture a realistic recording of acoustic instruments with a wide dynamic range, LUFS way below -14 are fine. I’ve released some solo piano pieces with LUFS < -20. This is unavoidable if you want to capture the full dynamic range, or something that sounds like it’s got the full dynamic range.

Mix your music so that it sounds right to you, see what loudness it is, and then take a decision whether to adjust it - depending on the genre, and on what other music you’re releasing it with.

The AES actually recommends normalising to -16 LUFS for a track, and for an album, normalising the loudest track to -14.

Here’s their document, which also gives a lot of background.

https://www.aes.org/technical/documentD ... ?docID=731

[JKN]

Richard, thank you so much for the info -- and the quick reply. Yours is one of the clearest explanations I've ever come across. Thanks, too, for the link to the AES document.

As for how I'm listening, I do have normalization selected on Spotify but can only choose a volume level of Quiet, Normal or Loud (I choose Normal), and not any LUFS level.

As for future productions, I now move forward much better informed.

Thanks so much again.

[Matt Houghton]

Yeah, first check what normalisatipn settings you're using. Spotify has three settings - loud, medium and quiet. The middle one is the default. I think the low one is for really dynamic stuff eg classical. The loud one is for listening in npisy environments (building sites, factories etc!) and is the only mode that employs a (not great sounding) limiter to raise the level.of quiet tracks that would otherwise have true peaks above 0dBFS. In other modes, the max loudness is limited by the true peak value. So if mastered quieter than the normalisation level but with high peaks, then it won't be brought up all the way to the normalisation level.

Second, normalisation is different for albums/collections than for shuffle. Third, even if you have I think three tracks from the same album in a playlist their relative levels will be preserved, as per album mode.

This was some of the 'nitty gritty' Ian Shepherd talked about when I recently interviewed him. But he was quick to point out that these details arw constantly changing!

[RichardT]

Yes, they are constantly changing and they vary between streaming services as well.

I think the middle / default setting on Spotify is -14 LUFS while the loud setting, with limiting applied, is -11 LUFS.

[tea for two]

I hover between -27 and -20.
I get alarmed if my stuff starts reaching -16

[RichardT]

I just read in the AES document that they believe their recommended LUFS values for music will drop to about -23 in future, as music files start to include loudness metadata. This will bring streaming in line with broadcast standards

[tea for two]

Had a gander at article on Loudness & Dynamics in this month's SoS by Ian Shepherd and Matt Houghton.

These two handy links in the article.

https://www.loudnesspenalty.com/
https://www.meterplugs.com/

::

I had a relook at my some of my Drum N Bass and EDM pieces yep they go up to -10. Lol.
So it is genre style dependent for me.

On a few of my Non EDM pieces theres parts of choruses also parts that are meant to be a bit louder goes up to -14 whereas the rest of the piece is around -27 to-20.

[JKN]

Quick follow-up question: apart from setting normalization levels in your app, does anyone know how Spotify processes a file that comes in above the recommended TP level of -1.0.

For example, if I submit a file at -15.0 LUFS and a True Peak of -0.2, what does Spotify do? Raise it to -14 LUFS, but put a limiter on it so TP is -1.0?

Thanks!

[muzines]

The loud one is for listening in npisy environments (building sites, factories etc!)

I think "Music For Building Sites" is Eno's forthcoming project...

[RichardT]

So far as I know, Spotify will convert incoming files to its internal format exactly as they are presented to it, so if you peak at -0.2dBFS the conversion is likely to produce artefacts (distortion of some kind) but it will produce an output.

It’s the conversion process that leads them to ask for files peaking at no more than -1 dBFS. If you don’t follow this rule, the playback level is not going to be your biggest problem!

[The Elf]

And yet there are videos like this...

https://youtu.be/h8MoK9vKl-M

[RichardT]

Hmm, has he read the Spotify guidance? Somehow I don’t think so.

However, plenty of tracks are mastered to peak at -0.1dBFS and with inter-sample peaks routinely hitting 0.25dBFS. The latest Muse album for one, which averages > -6 LUFS into the bargain.

[Aled Hughes]

And yet there are videos like this...

https://youtu.be/h8MoK9vKl-M

This guy’s videos has been popping up all over my Facebook and Instagram recently. Absolutely contemptuous and ignorant, not to mention damaging.

[JKN]

Thanks to everyone who took time to answer my latest question. I wasn’t able to get on here sooner. I think there’s a login glitch somewhere. Thank you again!

[Rich Hanson]

And yet there are videos like this...

https://youtu.be/h8MoK9vKl-M

This guy’s videos has been popping up all over my Facebook and Instagram recently. Absolutely contemptuous and ignorant, not to mention damaging.

Jeez, the presentation style of that video is enough to put me off. You really don't need to highlight every other word with a silly picture, I'm not f***king twelve!!

And "the only professional mastering engineer on YouTube"? Really?

[stevepowell7253]

Am I missing something here? The songs I record get murdered when posted to Spotify because they insist on making them way too loud -NB yes I'm talking about wanting music to be softer not louder! My typical mixes (and I suppose I would reluctantly describe them as soft acoustic or singer songwriter much as I detest genre pigeon holing) sound the way I want them to coming in at -24/23 LUFS and peaking around -6/5 dbs, Spotify will raise the volume alarmingly (+8.5 for example) so they sound diabolical. So saying mix a song so as to be how I like it doesn't cut the mustard if I want to post then to Spotify figuring this is the most ubiquitous platform. Aside from not using Spotify any suggestions about how I should proceed?

[Kwackman]

Surely the "mix" won't be any different, just the volume? It'll be like someone turning up their volume control, but your mix isn't changed?
{EDIT looks like what I wrote above isn't always true, please ignore!}

[amanise]

From what I can gather I think the general advice is aiming at -1dBfs True Peak - which is quite a bit higher than your -6. If you set your loudness as you like it - but with TP at around -1 its probably less likely to get mangled by the automatic normalisation and compression algorithms the platform uses to keep tracks around the same volume for listeners. When you hear your track back and dislike what you hear - that's probably what you're hearing? I think you can switch that off when you upload the tracks though - although I've never done it.

I'm not sure streaming is particularly friendly to subtle dynamics, generally.

[RichardT]

Am I missing something here? The songs I record get murdered when posted to Spotify because they insist on making them way too loud -NB yes I'm talking about wanting music to be softer not louder! My typical mixes (and I suppose I would reluctantly describe them as soft acoustic or singer songwriter much as I detest genre pigeon holing) sound the way I want them to coming in at -24/23 LUFS and peaking around -6/5 dbs, Spotify will raise the volume alarmingly (+8.5 for example) so they sound diabolical. So saying mix a song so as to be how I like it doesn't cut the mustard if I want to post then to Spotify figuring this is the most ubiquitous platform. Aside from not using Spotify any suggestions about how I should proceed?

It depends on the settings you are using for listening. Some listening settings will results in limiting to raise the LUFS, but some won't.

I'm not saying it's wrong, but why do you want your tracks to peak at -6 dbFS?

Here's Spotify's guidance. See the bits in bold. When they say -14 dB LUFS, I think they just mean -14 LUFS.

We adjust tracks to -14 dB LUFS, according to the ITU 1770 (International Telecommunication Union) standard.

We normalize an entire album at the same time, so gain compensation doesn’t change between tracks. This means the softer tracks are as soft as you intend them to be.
We adjust individual tracks when shuffling an album or listening to tracks from multiple albums (e.g. listening to a playlist).
Positive or negative gain compensation gets applied to a track while it’s playing.

Negative gain is applied to louder masters so the loudness level is -14 dB LUFS. This lowers the volume in comparison to the master - no additional distortion occurs.
Positive gain is applied to softer masters so the loudness level is -14 dB LUFS. We consider the headroom of the track, and leave 1 dB headroom for lossy encodings to preserve audio quality.
Example: 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.
Premium listeners can also choose volume normalization levels in the app settings to compensate for a noisy or quiet environment

Loud: -11dB LUFS
Note: We set this level regardless of maximum True Peak. We apply a limiter to prevent distortion and clipping in soft dynamic tracks. The limiter’s set to engage at -1 dB (sample values), with a 5 ms attack time and a 100 ms decay time.
Normal: -14dB LUFS
Quiet: -19dB LUFS
Mastering tips
Loudness normalization means we don’t always play your track at the level it’s mastered.

Target the loudness level of your master at -14dB integrated LUFS and keep it below -1dB TP (True Peak) max. This is best for lossy formats (Ogg/Vorbis and AAC) and makes sure no extra distortion’s introduced in the transcoding process.
If your master’s louder than -14dB integrated LUFS, make sure it stays below -2dB TP (True Peak) to avoid extra distortion. This is because louder tracks are more susceptible to extra distortion in the transcoding process.

[Wonks]

I think there may be mixed units being used here. StevePowell, can you clarify some of your values?

If your peak value is at roughly -6dBFS, then how can Spotify raise that by 8.5dB, as you'd then be at +2.5dBFS, (which is obviously impossible)? You might get a true peak reading above zero, which basically means it's clipping.

Or is it the LUFS value that's increased by 8.5dB (and if so, which of the three LUFS values).

You know all the details, but for use to comprehend what's going on, you do need to carefully spoon-feed us all the data.

[Hugh Robjohns]

....sound the way I want them to coming in at -24/23 LUFS and peaking around -6/5 dbs, Spotify will raise the volume alarmingly (+8.5 for example) so they sound diabolical.

I think we need a definition of diabolical.

Spotify uses a target of -14LUFS — so 9 or 10 dB higher than your target mix level.

But, although it wants to raise the level to -14LUFS it will only raises the level until true peaks hit -1dBTP.

So, if your true peak is hitting -6dBTP Spotify would only raise the level by 5dB.

Unless... you've selected LOUD mode when it forces the level to -11LUFS and hard limits as necessary to -1dBTP.

That would probably sound diabolical...

Ideally, you would re-master your -24LUFS mixes for -14LUFS... then they wouldn't sound diabolical.

[RichardT]

....sound the way I want them to coming in at -24/23 LUFS and peaking around -6/5 dbs, Spotify will raise the volume alarmingly (+8.5 for example) so they sound diabolical.

I think we need a definition of diabolical.

Spotify uses a target of -14LUFS — so 9 or 10 dB higher than your target mix level.

But, although it wants to raise the level to -14LUFS it will only raises the level until true peaks hit -1dBTP.

So, if your true peak is hitting -6dBTP Spotify would only raise the level by 5dB.

Unless... you've selected LOUD mode when it forces the level to -11LUFS and hard limits as necessary to -1dBTP.

That would probably sound diabolical...

Ideally, you would re-master your -24LUFS mixes for -14LUFS... then they wouldn't sound diabolical.

Hugh and I are saying exactly the same thing.

(According to Frank Sinatra, Hugh and I are just like a couple of tots, but I don’t think that applies in reality).

[tea for two]

And yet there are videos like this...

https://youtu.be/h8MoK9vKl-M

This guy’s videos has been popping up all over my Facebook and Instagram recently. Absolutely contemptuous and ignorant, not to mention damaging.

Jeez, the presentation style of that video is enough to put me off. You really don't need to highlight every other word with a silly picture, I'm not f***king twelve!!

And "the only professional mastering engineer on YouTube"? Really?

I think this is becoming the standard utube presentation style directive: hamming it up, biggin up yourself, ramming it down people's throats.