"High density" audio refers to the dynamic range, not to the LUFS-I.
The LUFS-I doesn't matter. You could export your -9 LUFS-I lowering it by 10dB for example and you get a new file that is -19 LUFS-I. And you cannot say -9 sounds better than the -19 one because they are exactly the same. It is just that you have to turn your listening device up on the 19 one.
On the other hand you could have created different versions of the mix, one at -9 with low dynamic range and one at -19 with high dynamic range. then export the -9 mix to -19. then yes, you have different versions, both at -19 and you could argue one sounds better than other and that depends on the music style etc...
I hope it is clear now.
When discussing LUFS and loudness in general it is implied that the top of the waveform reaches 0 dBFS or close to it. I.e we are discussing the density of a piece of music.
I hate over crushed music as much as the next person, all I'm saying is that there IS a place for it. If there wasn't, I'd immediately lose half my clients. I say this as an ME who has a rep for high sound quality and dynamics, and built that rep on deliberately NOT pushing things nearly as hard as some.
One thing I've started to do is master for a target of -18LUFS, and then limit to -14LUFS with 1.5dB TP (so basically 12.5LUFS with a TP of 0dB, same difference). This allows me to keep a dynamic master on hand for those who want to do an HD release in the future, while still targeting for streaming with the -14LUFS master. I find if the music isn't sounding good at -18LUFS with that dynamic headroom, you should probably work on the mix a bit more.
And you cannot say -9 sounds better than the -19 one because they are exactly the same.
If you limit/compress so that you can actually hit -9 LUFS (or similar in RMS for that matter), turning *that* down to -19 LUFS doesn't make it "-19 LUFS" unless the peaks still reach up towards zero which they won't. That's what the FS means in LUFS - Full Scale. If the peaks aren't near full scale, the LUFS number is mostly meaningless unless you do the math to account for the difference.
mlprod is exactly right when he uses the term "music density" because in your example, the density isn't changing, it's only compensating for the perceived loudness, it is not changing the narrower DR that was created to get it to -9 LUFS.
I also agree it's 100% content dependent. I land anywhere from -14 LUFS to -10 LUFS for stuff I record, not because someone told me to, but that's what it ends up as when I make it sound like I want it - long before I started measuring to see what the number was.
__________________ Music is what feelings sound like.
Last edited by karbomusic; 07-03-2022 at 12:47 PM.
-9 lufs mix will be like normalizing to -5 true peak when it is played at -14 by the streaming services.
I understand you all. I originally only wanted to say that maybe it was better to speak about the loudness dynamic range value, instead of the LUFS-I.
Who says you need to normalize to 0dbFS? There is no need. And in fact it should be maximum -1.0 true peak for compliance. -1.5 better.
No one said that (did they?). We use 0dBFS (full scale) for no other reason than to simplify the discussion because it's referenced that way in the specification. This is yet another reason to use one's ears to make the final decision instead of a meter.
__________________ Music is what feelings sound like.
I originally only wanted to say that maybe it was better to speak about the loudness dynamic range value, instead of the LUFS-I.
You are correct but that is what we are doing. That's what LUFS-I indirectly tells us about the entire song. It's an attempt to measure DR as perceived by humans (aka loudness) for the entire song. Before that, we'd take hundreds of RMS snapshots and average them all together to try to solve the same problem. That's what TT-DR meter did before LUFS was all the rage.
__________________ Music is what feelings sound like.
Every mastering engineer I've ever spoken to says they don't even think about LUFS. They use their ears and then ask the client to use their's when giving feedback. Being so precise with LUFS targets feels backwards to me. The music itself should dictate how compressed it is, not the other way around. Also, why -1.5 True Peak? Why not 1.3 or 1.7? Unnecessary precision in my opinion.
I only think about LUFS in the final limiting stage - most of the tracks I mix and master just happen to arrive around -18LUFS for full dynamics while LRA might be between 4 and 6. -1.5dB TP is only due to to the way lossy compression works. With a TP of -1.5dB you are guaranteeing that there will never be any intersample clipping when lossily compressed to MP3, AAC, Ogg Vorbis, OPUS, etc. If you do -1dB TB then probably 99.9% of the time it will be fine.
Lossy compression is all about the math, and while I’d love to simply mix and master to what sounds good and never use a limiter, at the end of the day, streaming standards and the proliferation of lossy formats delivered dictate the final limiting process which is why it’s SO important to have as transparent a limiter as possible. If you as the mixing and mastering engineer don’t take those things into account, you’ll be at the mercy of whatever algorithms each streaming service employs to fit your carefully mastered artwork into their obligatory box.
Every mastering engineer I've ever spoken to says they don't even think about LUFS. They use their ears and then ask the client to use their's when giving feedback. Being so precise with LUFS targets feels backwards to me. The music itself should dictate how compressed it is, not the other way around.
As a side-note, for me personally I prefer dynamic tracks and so have certainly run into scenarios where -14LUFS is not dynamic enough to express the emotions involved in the music itself. I’m a huge fan of powerful amplifiers, large speakers that can take it (Cerwin Vega CLS-215s anyone? AMAZING when EQd flat using a MiniDSP and given power to breathe) and turning up the volume control. Unfortunately, we live in a world of phones, soundbars, Bose car systems, and cheap convenience.
Every mastering engineer I've ever spoken to says […] They use their ears and then ask the client to use their's […]
I mean… have you ever worked directly with clients before? Most either don’t have a trained ear or worse, want their audio to sound “live”. Ya know, how it sounds to the drummer from his drummer position, and how it sounds to the guitarist from her guitarist position, etc.
Most clients have no clue what they want, but most of them do agree on the desire to make money. That comes from the mixing/mastering engineer and gets dictated by algorithms. Welcome to Technocratic Capitalism
I understand all of this already. My question wasn't about why we leave headroom for crappy lossy conversion (though again, plenty of top-end professionals don't bother, and lossy conversion has come a long way since the 2000s). My question was why do you so specifically target -1.5dB, it seems weirdly precise to have a value correct to one decimal place. I think people who fuss over this too much remind me of those who spend hours worrying about which noise shaping algorithm to use when dithering. The reality is, no recording has ever failed to make money because of intersample clipping. From a psychoacoustic perspective, it just doesn't matter.
Because -1.5 dB is the middle approach between the research and guidance for streaming services that suggests -1dB and -2dB of headroom. It’s easy to target; use Reaper to render with an LUFS of -14 and TP of -1.5 dB in the render dialog. Easy Peasy, no fuss needed. Doing that with any other limiter just take two renders and some easy math. Literally maybe 5 seconds to do the math and what, three minutes to render? It’s not a focus thing, it’s literally the last thing I do to release a finished product.
Quote:
Originally Posted by The Frown
I like soundbars and phones. I don't think we should be snobby about how people choose to listen to music.
We aren’t snobby about this at all. Merely pointing out the reality that your music will likely be played back on seriously compromised systems using psychoacoustic processing that will limit what you’re able to reproduce.
Quote:
Originally Posted by The Frown
I actually will come to your clients' defence a bit here. I think you might be misinterpreting them. Have you considered that what they're really asking for isn't for the track to sound literally 'live', but rather that it maintains the energy that it had whilst they were tracking it? Those aren't quite the same thing.
Actually no, I use active listening techniques to verify what my clients actually want. I’ve had clients who literally want a live sound with live energy. I have to explain to them that live sound requires some pretty extreme dynamics, big speakers, and lots of power to recreate. Otherwise their music will sound tiny, thin, and weak on average systems.
[QUOTE=The Frown;2574781]
I simply meant most (well, probably all actually) mastering engineers ask their clients for aural rather than mathematical feedback on the music, they don't ask their clients if they're satisfied with the LRA or if they'd prefer the song to be 0.3 dBFS hotter. They ask them if they think the music slaps or not.
[QUOTE]
Of course I don’t get into the technical aspects of things with clients, unless they actually want to know and then I’m more than happy to oblige. The only feedback negatively I’ve ever received is if the drummer wants louder drums or something like that. Each band member tends to have an idea of what they want and generally it gravitates toward the instrument(s) they play.
Quote:
Originally Posted by The Frown
Just to clarify the spirit of my questioning. I'm not knocking you for leaving -1.5 dB of headroom, it's good practice to and I agree in principle. I'm just querying the level of precision. I want my mastering engineer to focus on the bigger picture rather than decimal point details, speaking as a one-time client who has no idea what they want.
Totally get that; I always focus on the big picture. The precision is just super-easy to get on the final render as stated above. I barely think about it since it’s the very last part of the process and like I said, takes little-to-no time to do. Some people want the information though for final delivery since they want the streaming service to not touch the audio they’ve put so much work into. Hell, it wasn’t until a year or so ago that Spotify finally decided not to use a compressor and limiter on quieter tracks (unless you have your settings to “Loud” which in that case the algorithm targets -11 LUFS, and to those people I say, have fun). Hence the information
I don't have anything to add other than 99.8% of the time your clients, and the consumer, are going to prefer the louder one.
It just be like that.
Loudness wars aren't over it's just a lot more complicated now because you have to predict how different services are going to tamper with your mix and then try to predict how it will sound loudest within THAT paradigm.
CD era was nice because we rendered the final product ourselves. Now the various services take your mix and re-render the final product.
I don't have anything to add other than 99.8% of the time your clients, and the consumer, are going to prefer the louder one.
It just be like that.
Loudness wars aren't over it's just a lot more complicated now because you have to predict how different services are going to tamper with your mix and then try to predict how it will sound loudest within THAT paradigm.
CD era was nice because we rendered the final product ourselves. Now the various services take your mix and re-render the final product.
Exactly. Fortunately, it ain’t getting any louder than -14 LUFS on most streaming services so once I volume-match anything louder then that and play it back, the clients usually prefer -14LUFS version over the -9LUFS version. Any monkey can stick a limiter on something and “make it louder”, but when played back at the same volume, it suddenly sounds weaker.
I understand where you're coming from more now. I was under the mistaken assumpption that you were maticulously crafting your masters around the target loudness rather than aiming for it at the very end. Thanks for clarifying.
Ahh. Was wondering haha. Definitely not. It just takes longer to explain the process than to actually do it
I'm sure I'm not alone in thinking that loudness normalisation is and always was a terrible idea and should never have become a normal part of streaming platforms. We coped with using the volume knob on our amplifiers for decades, why is it all of a sudden a problem? If anything, it keeps listeners engaged. The loudness wars were never that big of a deal…
I respectfully disagree. People in general are lazy fooks who complain about volume differences, and to the clients, if it’s louder it sounds better. Try this sometime. Take your master, play it for the client, then play the exact same master but turned up 4dB. They’ll pick that one every time even though it’s the same file. The average listener simply disengages when a song comes on that’s noticeably quieter then those which came before it. Most listeners are using music as background noise and aren’t sitting in the dark with their eyes closed.
The loudness wars were and still are an annoying nuisance. Check out Apashe’s albums. I couldn’t figure out why I was getting a headache listening to his stuff since I like the music and composition is brilliant. Then I “ahem, found” the FLAC web version online and analyzed it. -4LUFS and 1.7LRA. Yea. It’s a brick wall of audibly clipped and compressed sound. A real shame since the music itself is brilliant at times.
But in the wild, as it were, it was never a real problem. Listeners weren't complaining that music isn't all the same volume, it wasn't something people were crying out for.
Ahhh, but you are forgetting, good Sir, that Nerds, not Artists, have inherited this earth, and Nerds got tired of the overly-compressed BS
Quote:
Originally Posted by The Frown
if people just stopped worrying about it and started using proper gain-matching when showing their clients their work, we could all go back to making music without worrying about LUFS targets, use our volume knobs as intended when listening, and hope the streaming platforms get the hint.
I don’t think anyone exactly—worries— about LUFS targets. LUFS is a relatively reliable tool at measuring perceived loudness by humans, nothing more (works great to normalize all tracks to say -26 LUFS before mixing btw since faster mixes usually equals better mixes since ears aren’t getting tired)
I DO think people worry about over-compression, which is the right approach. Gotta remember these young kids grew up with ever-increasingly-compressed garbage and so their ears are tuned to it.
Hitting a LUFS target is pretty straightforwardly simple now and most modern limiters are pretty transparent as long as you aren’t limiting more that ~4dB peaks
Yes but what system? Someone had to standardize something somewhere; the question is, where do we build to from here? Unfortunately the big tech companies actively stifle creativity from smaller people (usually guys, let’s be real here) or buy out their ideas and shelve them. Very anti-competition to be sure, all while touting competition to be welcome lmao.
I personally love digitalization if done correctly from a mathematical standpoint.
Honestly the only times I’ve found the same LUFS to sound louder between songs is when, say, there’s a track with nothing but stringed instruments… or when there’s a propensity around certain frequencies such as 1kHz, but then yea the way LUFS works, it only has two filters involved instead of a more advanced curve such as a Munson curve (and yea I know that curve depends on volume but that could be taken into account based on target LUFS, StereoTool for example has implemented Munson loudness curves with pretty reliable results)
When it comes to authoritarianism… are you living in the US? There’s an element of authoritarianism here; and social media certainly so, which is why I deleted all of it. Better to not have social media than to be passively supporting it… but that’s a topic for another thread
Honestly I wouldn’t doubt that it was the artists that wanted their music to be louder in order to try and sell more records. Also radio presence (although these days radio gets smashed to bits, check out StereoTool here: https://www.thimeo.com/) probably had something to do with it, although I haven’t looked into it enough. I’m actually glad the streaming sites turn down music that’s too compressed; -14 LUFS allows for PLENTY of compression. In my view, anything that’s mastered hotter than -14LUFS (well, really -12.5LUFS if using a -1.5dB headroom) is simply, and likely, too compressed. I mean, sure, most people seem to master EDM and it ends up around-8LUFS or so, but I’ll take the same tracks and mix/master to be more dynamic and it’s just that: more dynamic. Dynamics are really fun; especially when you have the speakers and power to drive it, just like, ya know, the clubs that EDM will be played in
In the loudness wars, I believe the straw that broke the camel’s back was Death Magnetic by Metallica. It’s tiring to listen to. And sure, according to the mastering engineer, he received tracks from the band that were already brick-wall compressed beyond reason, and was not proud of that master AT ALL, but that’s when you have to have the balls to get a hold of the band or mix engineer and walk him/her through what the problem is. Unfortunately the mastering engineer was probably working in a corporate setting where they didn’t give two shits about his professional opinion…
Not everyone is watching meters and waveform shapes. The point where you start getting the brick wall look when zoomed out is around -13 LUFS. Even if you aren't watching meters or waveform shapes, you'll hear that. And to be able to hit -13 LUFS overall level but still see waveform shapes with rock music takes some compression work in the production or you'll be lopping off your snare hits with a limiter.
Some left field stuff like fully blasted out sounds found in some EDM or truly punk rock kind of stuff do kind of need to be blasted out and those parts can hit -6 LUFS give or take. If your music isn't bionic blast house like that, it will suffer from the "Hulk smash!" technique of brick wall limiter and 10k high shelf until it matches that one Metallica CD. And it will be turned down by the current popular streaming services.
As far as rogue consumer devices like Blose ear buds and shitbars go, these are still designed to reproduce music from standard formats. The better the mix and mastering presentation, the better it will sound on a shitbar or phone speaker. Garbage in, garbage out still applies. There's no skewing of the mastering or mix that works better on a phone. Sparse arrangement music can be hyped up louder and tends to come across clearer on shit devices. There's this genre of basically a click track along with a speaking (or sometimes yelling) voice in vogue right now that can be hyped up loud. Knock yourselves out if you're doing that!
The fact that some of the big names in the industry are being self destructively lazy right now should be taken as an opportunity to show them up, not follow along.
I love you guys soooo much. Typing until the hairs are all well split.
I'm not sure if it's been mentioned yet, I've been a bit skimmy in most of this thread, but it seems to me mastering in the same room with the same monitors where you mixed and tracked everything is a bad idea, and possibly the main reason we send things to mastering engineers?
And btw, I'll toss another quick vote for lufs levels being song dependent. I always shoot for -11-12 but it usually winds up anywhere from -9 to -14 depending on the song.
My Merzbow 'Pulse Demon' CD has LUFS-I in positive figures, True Peaks of nearly +6dBFS, and sounds great, cos that's how it's meant to sound. OK, it's an extreme example of Power Electronics, but it exists...
simply pure sound, viciously unadulterated static. Literally.
Yea that stuff sounds the same at any volume. I personally can’t enjoy that stuff as it’s an assault on the senses vs actual music (boy I wouldn’t want to take mushrooms and listen to it, that’s for sure!) but to each their own.
Some left field stuff like fully blasted out sounds found in some EDM or truly punk rock kind of stuff do kind of need to be blasted out and those parts can hit -6 LUFS give or take.
Yea but remember, Blasted Out is what your amplifier, speakers, and volume control is doing. Also, yes some parts of the song are always going to be quieter or louder than others (well, except noise metal or the like) because if they aren’t, your ears/brain get bored pretty quickly
Yea but remember, Blasted Out is what your amplifier, speakers, and volume control is doing. Also, yes some parts of the sang are always going to be quieter or louder than others (well, except noise metal or the like) because if they aren’t, your ears/brain get bored pretty quickly
And that is precisely why one needs to be careful in mastering with final levels! Pushing up into heavy limiting past -13 LUFS and especially into the -7 to -5 range should be a special case. Otherwise you are delivering program that will be pushing the playback system and this leads to inconsistent results across different systems. If you had an element of your own mixing/monitoring system being pushed when you were doing that, no one else is going to hear that the same way.
And that is precisely why one needs to be careful in mastering with final levels! Pushing up into heavy limiting past -13 LUFS and especially into the -7 to -5 range should be a special case. Otherwise you are delivering program that will be pushing the playback system and this leads to inconsistent results across different systems. If you had an element of your own mixing/monitoring system being pushed when you were doing that, no one else is going to hear that the same way.
Agreed. In the extreme case mentioned above with Japanese Noise genre, that cannot be healthy for your electronics pushing those highly clipped digital signals through your speakers…
Some music doesn't seem to suffer from being squashed to "death"! Here's some brutal death metal, that stays around -3.3 LUFS long term, but jumps periodically to -2.8 LUFS. It is definitely louder, at least based on the LUFS figures than Metallica's Death Magnetic and it DEFINITELY sounds way better mixing and sound wise!
(It sounds the way it should sound like, even if you may not like the music itself)
Some music doesn't seem to suffer from being squashed to "death"!
The positive side-effect of that much compression/limiting is it can sound like it's cranked at low and comfortable volumes. The negative side-effect is that if you turn it up loud it is literally painful to the ears due to the distortion created (this song is a great example) - therefore listening for any length of time at higher volume is very fatiguing to the ears comparatively. That's not disparaging the song, just observing the results of all this.
The opposite is true for music with suitable dynamic range. It sounds better and better as you turn it up (until you reach the limits of the playback system), but not as impactful at lower volumes, details can get lost in the noise floor of the environment, and in all cases causes less ear fatigue comparatively.
Problem is you kind of can't have both.
Now which one chooses to master to is their prerogative and decision, but I find this aspect of LUFS/DR et al nearly non-existent in these discussions, while at the same time being one of the most important factors to understand while making the decision.
__________________ Music is what feelings sound like.
Last edited by karbomusic; 07-07-2022 at 09:42 AM.
Except that we could have both if portable devices included a limiter. Call it "loudness" perhaps? Digital audio format is happiness and light. People are just being idiots about it working in the last 4db like it was an 8 bit system or something.
If you make a proper full dynamic range master - which probably means no louder than -12 LUFS - you can produce anything more compressed from that when you need it. Then you can have whatever you think you need.
The positive side-effect of that much compression/limiting is it can sound like it's cranked at low and comfortable volumes. The negative side-effect is that if you turn it up loud it is literally painful to the ears due to the distortion created (this song is a great example) - therefore listening for any length of time at higher volume is very fatiguing to the ears comparatively. That's not disparaging the song, just observing the results of all this.
The opposite is true for music with suitable dynamic range. It sounds better and better as you turn it up, but not as impactful at lower volumes, and in all cases causes less ear fatigue comparatively.
Problem is you kind of can't have both.
Now which one chooses to master to is their prerogative and decision, but I find this aspect of LUFS/DR et al nearly non-existent in these discussions, while at the same time being one of the most important factors in the decision.
I do agree, but with a little caveat of, "if you are using good headphones and really want to "slam" some dynamic music, your pain tolerance mileage may vary". In fact I just listened to Eric Clapton's Journeyman and tried to turn it up (on my headphones) and those drums (on the first track) just made it almost unbearable after a certain point. They used a drum machine on the track and it can't really vary the playing dynamics on its own (they probably didn't program it enough or there wasn't any possibility to do so), so to me it'll sound good at a certain level and then it can become painful and overbearing if played louder really quickly!
People are just being idiots about it working in the last 4db like it was an 8 bit system or something.
+1. You could hit conservative peaks of -48dBFS in 24-bit before it equates to the same fidelity as 16-bit! Hence, suggestions of -10, -12 or even -20 dBFS peaks being entirely reasonable when stepping into the unknowns of live recording.
I land anywhere from -14 LUFS to -10 LUFS for stuff I record, not because someone told me to, but that's what it ends up as when I make it sound like I want it.
Although there are no rules, I do believe in common sense and I'm with Karbo.
I maintain a -1.0 Peak while I shoot for -14LUFSI. I almost never go beyond a -13LUFSI, but there are certain conditions I will accept -12 to -10LUFSI.
I can't handle music that is flatlined or overcompressed.
I can't say enough for Reapers ReaLimit and JS Loudness Meter, they make mixing and mastering so much easier.
I do agree, but with a little caveat of, "if you are using good headphones and really want to "slam" some dynamic music, your pain tolerance mileage may vary". In fact I just listened to Eric Clapton's Journeyman and tried to turn it up (on my headphones) and those drums (on the first track) just made it almost unbearable after a certain point. They used a drum machine on the track and it can't really vary the playing dynamics on its own (they probably didn't program it enough or there wasn't any possibility to do so), so to me it'll sound good at a certain level and then it can become painful and overbearing if played louder really quickly!
I'm sure we aren't far apart.
Growing up I found there are plenty of mixes that don't sound that good loud regardless of limiting et al. But the dynamics vs added distortion thing isn't as subjective as some might think, it's something that really occurs to our ears with volume, "what" is loud is comes into play.
Yea I agree, there is still a pain tolerance but it's a lot louder to get there. The main thing I'm trying to get across is the amount of time a person can listen comfortably goes down as the distortion from limiting goes up. If one listens to highly compressed music at low volumes, there is at least some illusion of it sounding loud (a good thing) and the ear fatigue isn't an issue; making much of a subjective thing at that point.... until you crank it and keep it cranked so to speak.
^This was something I noticed on my own long before I knew very much about compressing and loudness wars. Then I started creating playlists for my car (had a nice system). And I had a high DR list and a crushed DR playlist of stuff I really liked. I would Guinea pig unsuspecting riders by playing the high DR list followed by the low DR list. Without fail, every single time the crushed play list started they yelled at me to turn it down within 10 seconds, though the previous hour of high DR material cranked as loud as I could without distortion, they were absolutely fine with and bobbing their heads to prove it.
So I'm not knocking either per se, just describing observations about how they behave and affect.
__________________ Music is what feelings sound like.
Last edited by karbomusic; 07-07-2022 at 05:12 PM.
I should have also mentioned that I don't really follow the method prescribed in the link contained in my earlier post. I do, however, put a hard clipper on almost every track, set to reduce peaks with no audible difference. Using this method I can easily achieve -10 LUFS with my limiter kicking in only on snare and kick hits, and even then only reducing the transient by less than 1.5 (and typically less than 1) db.
There is a fair amount of compression on each track, but for sonic purposes only (i.e., keeping volume levels consistent, glue, or for special effects). So I'm not intentionally trying to make it easier to get higher LUFS, I'm compressing to make everything sound the way I want it to sound. Serr had made a statement earlier that anything lower than -13 LUFS would necessarily sound squashed (or something to that effect); this is simply untrue.
Serr had made a statement earlier that anything lower than -13 LUFS would necessarily sound squashed (or something to that effect); this is simply untrue.
There is a threshold around there (or lower) where things change for a lot of music, whether one calls that smashed or not is opinion, and there's a context dependent margin, but the threshold he's attempting to explain does exist in that general area, and it's easy to get a little loose with the adjectives. Good or bad is in the ears of the mixer's intentions though.
But that's irrelevant to the grand scheme of things, just make it sound like you think it should sound.
__________________ Music is what feelings sound like.
Last edited by karbomusic; 07-08-2022 at 02:50 PM.