To clarify some points made in this video: - The true-peak meter is to catch inter-sample peaks, or the peaks that happen between two sample points that otherwise go unnoticed on a regular meter, since the two samples are both under 0dbFS but not the crest that rises between the two samples. The drawing he was making made it look like de-clipping, or that a clipped waveform will get re-drawn with a new crest on top in the analog domain. This is not the case - he meant to draw the crest between two sample points, not to say that these will be added later during playback, but rather to show that these crests may already exist between various samples and are simply unnoticed by a regular meter. The -1dbFS headroom allows for these inter-sample peaks, as well as gives headroom for transcoding. - Spotify does not attenuate any of your files that you upload. The only loudness attenuation happens when the user's client has "normalize volume" enabled in their preferences, as well as a "volume level" setting below the normalize option. A simple test will confirm that when normalize is disabled, loud tracks are very loud - far beyond -14 LUFS or anything close to that. For example, try "Hibou - Above Us" on Spotify and then grab a demo of Sonnox ListenHub and observe the LUFS. It's like -5 LUFS or something crazy if I recall. Extremely loud. If the listener has normalize disabled as I do, and you upload a -14 LUFS track, your songs will sound quieter in comparison to most louder tracks since nothing is modifying the music. However, when normalize is enabled, the louder tracks can lose impact as they are attenuated in comparison to other songs that are not as smashed already. Also keep in mind that there are no loudness settings on the web version of Spotify and so web versions will all play as loud as they were uploaded! You can test this to confirm as well. If you have the Spotify app open while you also have the web player open, the web player will inherit the loudness of the app until the app is closed. Close the Spotify app, go back to the web player, and the song will be played back at full loudness.
100% with you on the Normalize Volume settings on Spotify. As long as Spotify (and other streaming services) have this settings as an option, I cannot think why people would want to master and export their final track at -14LUFS for example. Mind you, i am not hating on those who do, i just don't find it attractive as an option since electronic dance music (in which i specialize) competition is so far away from this standard still. Good insight on the Web and App based version though. Did not know this.
Good mix won't lose impact even after normalisation, although it might be genre dependant. In busy metal, which I mix most of the time, if you balance and process tracks correctly (mostly dynamic-wise) I start to limit peaks only around -9, so when I'm near -6 I'm limiting some sources just close to -0,5. Anyway I use normalisation on spotify and believe me that none of the extreme mixes lose their impact. The only scenario where I'd believe that impact could be lost (in metal at least) is the case where someone mixes really loud from the start and loses perspective.
Wait, the true peak thing. Isn't he demonstrating what analogue converters do to the flat part of the signal? AFAIK, the flat part is like a square wave which depending on the duration, it can have frequencies higher than Nyquist. Some analogue converters are not designed to handle this, so it creates artifacts (both from not being able to handle it and having a low pass filter or shaping it to be inside Nyquist).
Daft Punk threw a wrench in the Loudness War with their last album- basically mastered it for vinyl and left it alone. Beautiful record, zero ear fatigue.
I can say, that one thing, that drives people to get stuff loud to the level you mentioned is a competition of other releases on labels, that they are trying to get signed on. For example standard loudness on a lot of Liquid DnB labels is around -8 to -7 LUFS. People listening to the demos take the loudness into account. It is hard to steer away from pushing stuff loud when you are basically conditioned to do it since the beginning.
Loudness does matter tho At work they have to crank it just to hear and some songs don’t cut it We are not going to keep adjusting the level to hear at least not while working Especially some of the dynamic range differences Too quiet and it disappears
I believe that labels look for dynamics/loudness ratio, like the dynamics preserved within that loudness target. A better example can be seen in Neurofunk, where generally you see more filters "movement" compared to Liquid.
@@davidpereira4455 I agree and perhaps I could have write that in a better way. Although I still think, that releasing electronic music at the top tier level labels slowly conditions one to push it loud - sometimes way too far. I have been in discussions on GS forums and there was a interesting talk about this topic. Some of the artists and engineers there said, that not shooting for the hot masters basically means career suicide in certain genres like hip-hop for example. I have fallen in the same mindset when making my own music - coming from trance and house genres, all thanks to the competition and release ambitions. They - the labels - need DJs to propagate the releases and they often test the track in a mix, and if it dips in loudness to much, it is a problem - DJs might consider that a bad sound and not suport it.
There's just one big true here.....When BIG top charts engineers stop doing their masters at -8 / -9 LUFS and begin to follow the platform specs, everyone else will follow (including me)! Until then, we'll always be in the loudness war! That is.
We deliver all of our music at approximately -9.8 to 10.5 LUFS and between -1.4 to -1.2 true peak (it varies a little depending on what the sounds best on the song). This gives users some room to turn it up a tiny bit in case they need it. We've had zero problems or complaints. I know our loudness is a bit higher than requested but as streaming is a smaller part of our business, I'm perfectly happy going this route.
Hi, I am exactly in the same range of values. Seems to be a nice compromise. I tried to be at -14LUFS, but then, on sites where there is no normalization... my hiphop and metal tracks were so quiet they were skipped by the listeners...
@@AboveEmAllProduction We aren't making psytrance so I don't see the problem. We mostly compose ambient, ethereal, new age, some classical/cinematic pieces with other things thrown in when we get other inspiration. Writing for real estate videos or for corporate videos for example, we don't need -7 lufs. They're going to turn it down substantially anyway to get it below dialog levels.
Absolutely love this video!!! When i started out over 20 years agao i never gave a crap about being loudest. When i got back into music production it was all about being loud especially in the edm realm. I jumped all over it. I ended up making some of the worste sound music i had ever made trying to keep up with being loud. Over the last few years ibe gone back to just making music i like not caring as much as how loud it gets. I focus now on my tracks sounding good an more coherent.
There was a great Sound On Sound (magazine) article on exactly this issue some few years back 'The End of the Loudness War?' (February 2014).. and in my opinion one of the best articles to appear in Sound On Sound.. an absolute fantastic read. And yes! The ending of the loudness war does mean that producers can breathe easy and produce music that has wide dynamic range once again.
It is not only about loudness. Loudness has it's sound. Driving the clipper and limiter harder produces more aggressive sound due to pops and cracks and also side result is lower dynamic range - compression. So if you master at 6 lufs and upload it to spotify no problem, it will automaticaly turn down whole song to average -14 lufs... but you will still retain aggressivnes of a mastering pushed hard.
there are so many songs that I love that happened to be -7lufs or louder...and I'm talking about major artists... good songs with awesome production and mix will sound good on any loudness threshold
There are great tracks that hit even -3 LUFs... plenty. This is a conversation about genres and taste... either you like these genres or you don't. A number on your LUFs meter shouldn't be what guides decisions. If you are making loud music, it's going to be loud. And yes, there is a such thing as loud music... has very little to do with mastering limiters lol but this debate never ends.
I got a question about this topic. My friend is an engineer for Grammy artists and he told me to put the loudness on -9 db and the true peak at -2db. When I told him that everybody says -1 is enough for true peak he said it's true but that's if you wanna stay at -14 on the loudness. when you wanna go louder you should go down on the true peak and put it at -2db. How true is this and why?
I guess it's about the conversion algorithms, as Wytse mentioned in the video as well. Specifically compressed audio file types like OGG and MP3 will alter the sound significantly and to avoid any clipping after the converted files, you need to have the headroom. The louder your track is on average, the more likely it is to go beyond the 0 dB true peak after conversion if -1 dB is used on your master track.
I check the website tunebat to see what the lufs are for my favorite tracks usually they all land somewhere between -8 & -6 lufs, the random access memories album which is one of my favorite albums ever actually had an average of -10/-9 While artists like Kaytranda come in hot at about -7/-6 so if the pros are loud I’d say be loud as well just don’t go overboard
I agree but will also say that there is nothing stopping you from making different mixes for streaming per their recommendations and a loud one for physical media. If you only want to make one- go the RAM route and you'll be safe.
It's funny that it's actually something I found out by myself when I was still a rookie at music production 10 years ago. Back then I'd make my tracks as loud as I could, but I would always make sure I peaked at -0.1 instead of 0 because I hated seeing pixels above the limit when I opened them in Audacity. Since then, I stopped brickwalling my music, so I don't have to care about competition and how overlimiting will affect my tracks (almost always in a negative way).
I have no interest in mastering louder or quieter than anyone else. I LIKE the sound of loud as shit, and I LIKE the sound of very dynamic. I just do what the track needs, or what I want to do. I make intense aggressive music a lot of the time, so I master very loud. Of course I take a lot of care of transients, but I saturate a lot, and I mix so that there isn't much gain reduction needed in mastering. But when I master a smooth jazz track, or drone or something, there isn't a whole lot of dynamics reduction. So, for me, it's more of an emotional resonance with the music, and an intuition, rather than some kind of competition. The competitive aspect of loudness is so so so dumb. It's all ego. Screw that stuff.
With dance/club music, it's important to match the loudness of other producers because 1. It's not really comfortable for DJs to always match the volume for every song; 2. Your song, no matter how good it is, if it's not mastered properly, it will sound bad being played in the set after a louder and better-mastered one. I've been a producer for more than 10 years and also always been obsessed with the loudness issue. However, eventually it's very difficult for amateurs to mix/master tracks properly and make them sound good and loud, so in a way, sounding loud but good is like a threshold that separates pros and amateurs. I always had to match the loudness level of my productions to famous tracks, or they just wouldn't work when I play them, and it hasn't been easy. I've been a DJ much more than a producer. I get a track I love and it's quieter. I'm thinking, ok, just to remember to raise the volume up, however, it just doesn't worked after most tracks, even with the volume up. I know, you'll say, oh but the mix was not good, if the mix were good, it wouldn't have to be loud. But having to match the track's loudness to famous songs actually forces you to do a good mix and master because there's no choice. Also if I enjoyed the track, then the mix was not necessarily bad. I think, there is much more to commercial loudness than just standing out on the radio.
1. Noone listens to radio. 2. I hear a cheesy dance track that's too loud with not dynamic range I turn it down to match my vibe, then it feels lifeless because it has no dynamics and I skip it. 3. You'll only get love for your music if you have good ideas, melody, structure, philosophy etc. 4. Streaming is set to auto level then my ipad auto levels etc. 5. Get off the loudness train.
if you as a DJ are so concerned with loudness matching every track then it sounds like you should add a limiter to your signal chain and destroy every track’s dynamics on your end
@@sqcaraudio He's not talking about radio. Not liking some music is fine, but the point of the OP remains the same. Some genres have loudness-competition built into them due to the nature of the industry and the venues we play at. Some genres are worse than others, mainly bass-oriented EDM, but loudness is important for most DJs that play at venues larger than 500 people. Clubs skimp on the systems and you have to work with the limits they have. Sadly that means some tracks won't work in context of others. I'd love higher end PAs and more dynamics in the tracks, but c'est la vie. Optimally you'd have masters for clubs and masters for streaming, but who can afford two masters of every track.
@@bananermat3798 can’t djs just use a limiter to loudness match on the fly instead ? i imagine that’s what most french DJs in the early 2000 did when incorporating 70s funk records into their sets
The ONLY reason streaming platforms request -1.0dBTP is due to intersample peaks caused by SRC algorithms in streaming codecs. In reality, these can exceed 6dB or more in some circumstances, but Apple pragmatically chose -1.0dBTP as part of Mastered For iTunes and streamers followed suit. Many argue that intersample peaks are inaudible, which is why some mastering engineers continue to use -0.3dBFS (i.e. some still rely on Full Scale and ignore True Peak, which is essentially an oversampling algorithm to predict intersample peaks). The main reason for adopting -1.0dBTP for WAV masters is due to subsequent issues with lossy codecs (i.e. SRC algorithms when converting WAV files to other formats). This doesn't completely eliminate all intersample peaks, but it's sufficient for most. In other words, adopting -1.0dBTP is a very useful, pragmatic metric for multiple reasons, but some diehards disagree.
Intersample peaks are only reproduced during analog conversion. This will happen at any digital level and it simply cannot be avoided. The DAC can handle a much higher voltage than the nominal corresponding voltage for 0 dBFS, and since the current is alternative, it will always oscillate to reproduce the true peaks. Of course if you push the true peaks very high (artificially, because it’s technically impossible to do that with just music), then there will be a limit to the maximum voltage the DAC can output. AC cannot be a fixed voltage unlike DC, so it will not clip flat either, instead it will oscillate very fast and create unwanted frequencies. Please leave the true peaks alone, this is what gives you that extra volume extension that none of the streaming platforms can detect. In any case, one sure thing is that if the loudness of your song is already squashed to -9 LUFS, it isn’t a true peak of +0.5 dBTP that will increase the max volume…
@@JakeyWakey Limiting at -1dB Digital has nothing to do with the true peaks... True peaks will still happen when your speakers are reproducing the sound, this is inherent to the physical properties of an analog waveform...
@@louisrmusic yes and those have been happening since the dac and no one's been comparing or noticing audible differences, but that 1db apple chop is offensive.
The basic vibe I feel here is people's insecurity in wanting to be famous or signed. Having been signed 20 odd years ago and now being retired I'll give you my take on this. All A&R peeps will use a volume knob when hearing a bunch of demos. They just will. They are not idiots. They won't hear a bullshit track that's -6db rms in old terms and go "wow that's the loudest track sign it nowww".. give me a break. Personally mastering some later releases I chose -14 and -1tp and it gives me much more satisfaction in hearing crisp snares and punching kicks and in general more sparkle.
Very well said, I just downloaded the pdf and I'm gonna read it and seek further info because it's a very interesting and dividing subject. Keep up the amazing work, your channel is GOLD!
@@Whiteseastudio You are right, but imagine being told you are doing it somehow "wrong" and get caught at being ignorant... Guess for some personalities this is hard to swallow. 🤷♂️ Florian Camerer was my teacher back then in 2007 (the guy who co-invented those new loudness norms and measuring) and mastering at decent- lower- levels was so normal for me that I used to watch that loudness war from a distance, scratching my head.
@@Whiteseastudio Yeah I totally agree, but people who have been doing it their own way for years and actually know what they're doing, a lot of times claim that those ''rules'' are ''crap'' because they could possibly ''kill'' individual preference and signature and stuff like that. Tbh i don't understand no of that, I think most of the times giving shit to these ''regulations'' is a way to boost your own ego and to feel like you ''know better than them..
You mentioned A&R managers towards the end, and I truly believe they're the reason the loudness wars is still a thing. When they receive demos, the demos aren't played in end-user streaming services, but via some file-sharing or cloud storage services, which very much does NOT do loudness normalization. It should also be noted that soundcloud, which is where the loudness people mainly upload their music, doesn't offer loudness normalization to all its users
Interesting, I noticed there were huge differences between songs loudness on SC but didn't know it was due to them not offering normalisation to all of its users. I now use Soundcloud as a place to test and compare my music to pro reference tracks.
I have tried several times to follow the recommendations of the streaming platforms and doing that...the tracks sound far below the references that I use (Artists that I admire). Of course you have to turn off Spotify's compression that evens out the volumes... to be able to perceive that. I think that the benchmarks for ''BIG'' productions are guided towards what is a CD Mastering level... between -8 and -11 LUFS (approx.) and between -0.5 and -0.1 dbTP. When I started using these values everything sounded at a much more acceptable and satisfactory level to me! (it is also much more difficult to achieve without distorting). The platforms do not reject files with these values outside the recommendations...because the big industries that take most of the profits would make them undo that decision right away, they want the volume war to continue...and separate themselves from the ''amateur audio productions'' (ironically...there are missing dynamics in the ''PRO'' way) Also...all these recommendations and streaming platforms could change or disappear one day, and I don't want my entire collection of music productions to be based on transitory things. Cheers!
I wouldn't call streaming services transitory. They pushed CDs out of use for good (check how selling CDs dropped). They became the vinyl of the industry now. People even stopped having any CD readers at home (especially on laptops and PCs - they're not even there anymore). Everybody is accustomed to streaming, and that's a custom very difficult to turn back from now on. And it's been like that for years now. I wouldn't call it transitory. It's been well adopted.
@@---pp7tq I didn't mean that streaming as a way to consume music is transitory (but it could be) or that CDs will be imposed again. I meant that streaming services platforms and their guides values can change, even appear something newer than streaming. I mean that everything changes, and that's why I don't follow to the rules the ''moment''
I follow the specifications to the letter and it’s funny because most of my tracks seem much quieter than most other comparable tracks, but my stuff won’t get penalized and still has dynamic flow. It’d rather keep the integrity of the music than JUST worry about loudness.
As a mastering engineer myself, I have been mastering to -1 true peak for a while, but 3 out of 5 times the client comes back to me and says it's too quiet. It's so frustrating, but you have to keep the client happy. I'm amazed how many people in the music and tv / film industry still don't understand the reason behind this, and still insist on having that extra dB.
In the TV industry too? they don't demand -23LUFS-i? Are you sure? Because here in Austria they obey that norm pretty strictly. I think same goes for other European countries.
@@Mansardian tv mixes are -23, yes, but when I master music for tv (trailers mostly) they still want it maxed out at 0dBfs. Crazy, as they'll have to turn it down considerably.
I just master to -0.3. Apple has a plugin in tool on macs (and I’m sure there’s others) that can analyze your audio in various compressed formats for peaks and overs- the more compressed the codec the more overs. I find -0.3 true peak to be safe enough, it’s not like any streaming service is really going to use 96kbps these days.
One important point which missing the true peak get mostly in effect when music file will be encoded in a other format mp3,aac,oggvorbis etc. The encoding can also add little bit volume this is also why you not can see any peaking if you check your Orginal pcm wave file in DAW.
yeah, but why can i never hear the difference? If you listen to it in FLAC or WAV, it doesn't peak above 0, but when you export it in MP3, there it says for example on this song I just checked, Sample peak 0.65. On some songs even higher but I never hear a difference? Also True peak is always a little higher than the sample peak, i've never seen it the other way around.
@@asd2640 you can hear difference if you compare wav to mp3 if the peak is pushed to hard e.g. -0.3dB or -0.5dB. But this always can depending on kind of music you are listening e.g. Rock or Metall music where you often use tool like clipper. I personally notice the effect on releases which I download from juno, of some of the release was more tired after listening. When I use the metering tools at wav and mp3 version it shows that they was pushed to hard for encoding.
Excellent video and much needed, thank you! A discussion is also required regarding engineers recording too hot, trying to track close to zero dBFS for some reason, daring a loud hit or note to clip, when 24 bit recording gives us so much dynamic range (144dB theoretically) that recording with a peak level of -20dBFS and then compressing that signal 20dB still leaves us with way more dynamic range to play with than any studio's noise floor. You may say this is a hangover from the analog days when hitting a preamp and tape head hard was an optimal way to work, but some younger, digital-native engineers seem to have a mania for trying to get near zero (FS), as if it's some sort of race, like they will have a weak, feeble sound if the meter never hits orange. The loudness war starts at the preamp. And starting a mix session with an iZotope RX De-clip party ain't no fun.
My masters aren’t loud even when I try lol but I do understand that perceived loudness really makes a difference when djing live. I’ve killed energy on the dance floor with bad masters or when i export a wip with bad limiting. But I agree, no need to fight the loudness war for streaming any longer. Should we just export two masters?
Until every top 100 song on spotify really commit to loudness normalization nothing will truly change. I would love to follow the guidelines but I have to follow the popular songs that are the references and follow the client desires to match their favorite artist songs.
@@Whiteseastudio heheh great analogy… the endless cycle….. I hear you, but really my hands are tied. Many of my clients do reggaeton, and those are played on clubs here in my country…. All of the clients want to match the loudness of Maluma, Camilo or J Balvin, if I don’t, they find someone else that do.
In principle I agree with everything you’re saying BUT…. my competition at Sterling Sound and other Grammy Nominated mastering engineers are cranking out everything hitting-8 to -6 LUFS -8 hitting a limiter still sounds better than -14 on Spotify regardless if the normalization is on or not.
I work with Sterling Sound, and the masters I got from Ted are obviously louder than my mix, but also feel more dynamic and just sound better, and I can tell you, he's not shy on limiting!
Loudness is more than the number on your meters. Our brain recognizes compressed/limited signals as being loud because our ears do compress themselves if what we hearing gets too loud. So perceived loudness will always be a part of music production.
The thing nobody seems to want to talk about is lossy encoding and what the associated EQ changes do to peak levels. I've personally observed as much as 3dBFS additional peaks added by 192kbps mp3 encoding. Imagine what distortion that adds when your mp3 file is converted directly from a source that slams the 0dBFS mark with no gain reduction....
This conversation is tiring, and it's hard to explain to people that -3 or -4 LUFs is literally the sound of the arrangement and sound selection in these loud genres. It's not just a trick you do on your 2 bus... so it's a completely different conversation, one about style and tastes. You will never be able to make a Beatles track as loud as a tearout track because the arrangement and sounds won't allow for it. When we use synths or whatever to make our sounds and then process them to make them sausagey and fat... we are preparing to make a loud track from the very start with very full waveforms. If there's any "secret" to hitting those ridiculous loudness levels and still sounding clean and dynamic, it is to limit your subs and clip everything else as a general rule. You can do this with something like Kilohearts snap heap/multipass. If you arrange properly in heavy bass music, choose the right sounds (saturated and clipped/limited @ the track and bus level), you can hit -6 LUFs and still have dynamics and punch without much of a difference in quality with or without a mastering limiter/clipper. Getting to -4 LUFs from there isn't difficult with certain bass music genres/tempos. People that gripe about Loudness usually seem to mix everything but heavy bass music genres and aren't producing bass music themseleves. Done right, one can get most of the way there with arrangement and proper sound selection alone. There are many amazingly produced tracks out there hitting even -3 LUFs during the drops and it still sound clean, as proof.
Here the thing about loudness, I reference other tracks, and each genre has different loudness; I attach loudness to the style, which makes my track closer to the style I am looking for. Example Dubstep = -1 to -3 Luf, Pop = -8 to -12 Luf, Rock= -10 to -14 Luf. Once, I tried master dubstep to around -14, and It didn't sound the same; it sounded weak and hollow, which is why now I prefer to mix and master my song to a reference track, making life more straightforward rather than just wasting my time and money debating on something that clearly made to be broken because rule or recommendation is created to follow but also broken.
Ohh danm oldschool dubstep producers are crying now... is -1 to -3 LUFS really common in the newer stuff? I don't listen to that genre so idk but that seems very excessive, even for dubstep. Must sound horrible at that level. Up to around -6 LUFS I can still understand it, even if I prefer more dynamics. Goes the other way too though, beyond around -12 LUFS things become a bit too dynamic for most music to me and mix can fall apart with too little compression/saturation.
@@frogify_music yeah the dubstep scene is that loud now but just from my perspective it change to make it more dynamics now. I check most of the .flac version of most top dubstep producer the loudness is around -1 to -3 luf and nothing under that. I use to obsess with loudness nad master song around +2 luf which is dumb but good experiment.
@@frogify_music Also dubstep producer have a technique to work around the "over compress" sound which is the rule of play "one sound" at a time. Fun fact we only use compressor when need , most of the time we use a really good clipper " The Newfangle one". Also majority of the time they spend time sculpting space for the different sound. ✌🏾✌🏾
when i would master to -14LUFS i was always told the music was quieter than others on a given streaming service. i've had that listed as the reason for being left off playlist etc. if you keep enough headroom you can have a lot more perceived loudness. ALSO, the playback is only normalized in certain situations. Also, listeners have the option to turn OFF normalization... bottom line, do what needs to be done
For bass ppl i always recommend Tipper - Mariscos it's a -10 LUFS track and great example to explain clients that by not getting -4 LUFS loud we can have more bass.
A lot of interesting information. My only complaint is that making everything at -14 means the end user must use significantly more amplifier power on literally every playback device just in order to hear the music at a reasonable level for enjoyment. I certainly don’t appreciate “Loudness Normalization” toggles and I always turn them off. I want to hear the music at its intended listening levels, not have to crank it all the way up on my car or headphones from my phone. You can’t hear it loud enough in most situations. When done properly, you still have plenty of perceived dynamics. Don’t mistake me, I HATE overly compressed music, but when mixed properly, it can be very loud and not sound over compressed. -9 or -10 is still plenty dynamic. In fact, almost all my mixes before being mastered are sitting at -11 or -10 and barely need any final limiting to reach any sort of level and they still sound plenty punchy and dynamic. I think the loudness normalization is a scam.
Totally agreed. I don't think I can enjoy the dynamic and quiet songs in a noisy subway station. Also, even if applying the loudness normalization, songs may not sound even due to different energy distribution in frequency and those psychoacoustics stuff. So why bother turning the normalization on?
Mastering to a higher than -14 LUFS is not always about loudness itself. It has to do with dynamics perception. You want to listen to the dynamic classical record, or a medium-squashed pop song maybe a heavily compressed / limited metal or EDM record. Yes, at the end of the mastering chain you can technically turn down the file to -14 LUFS, but nobody does it bc then you'll need to use a separate master for almost every streaming platform, and loosing loudness when normalize is turned off by the end user. But they definitely don't want to hear a dynamic metal song or a squashed classical piece just because the recommendation is -14 LUFS so you'll give the snare that extra +3db peak when normalized. Also true peak -1dBFS is (as far as I know) only for the compression. When the platforms are streaming in low bitrate mp3 formats, the algorythm needs more headroom to not distort the audio.
And yes.... this message concerning loudness very worth repeating again and again... since even though the new recommended protocols came into existence as long as 8 years ago.. still too many fledgling producers know nothing of a) the importance of this and b) the meaning of it.. This whole issue is actually an absolute First and guiding issue. No music producer should be without understanding LUFS... and why LUFS.
Hey Wytse, thanks so much for great and informative video's mate. One thing I would raise as a counter for discussion, is that in my experience a lot of client's still want physical production of their work and are not necessarily wanting different versions for streaming, once they approve the physical versions of the mixes. Often these client's are not new to music creation and have previous releases at a given level of loudness. In this instance you don't want yours to sound quieter than the previous album. To client's it won't make sense. This is especially present in genres like rock and metal, where loudness IS still a thing. Also I would argue that it is an art in itself to get a master to a very loud level (lets say -6 LUFS) whilst still retaining dynamic punch. At lower LUFS this is a little easier to achieve as you're not crushing the life out of it. Anyway, thanks again for the great content!
As a DnB (Drum and Bass) artist, I usually master my stuff at -8 LUFS because it will translate in a club and I'd rather that streaming services turn my stuff down than up. I also use Pro L in True Peak mode to minimise overshoots.
For the loudness war to truly end, normalization needs to be forced on all playback mediums, and not just commercial streaming services etc. Because as you mention, the war is still in full swing when sending songs to A&Rs that they play from their mailbox or computer folder. And since it's actually very difficult to take a dynamic mix of let's say -16 LUFS and master it to -7 LUFS without major shifts in especially the low end, many producers start pushing way early in the production/mix process, meaning the production itself is often designed around having a loud master. It's actually super common for producers to have a massive limiter on the master fader on their template, before even starting a track. And because of that, the loudness war in the A&R room is also felt in the final releases since the A&R-approved production require a mostly intact loudness to not fall apart. So even with a mix engineer that knows his/her shit, they are still trapped.
The minus 1 dB peak point is there to give the system "wiggle room". If somewhere in the chain the gain drifts above unity by .5 dB , it drifts up to -.5 dB, NOT .5 dB 'over'.
So in the past few years, I was a mixing & mastering engineer in my own studio, a mixing engineer for live band, an a system engineer for both audio company/live event/live band. I used to question my self this : how an old song like Queen's song, Beatles, Etc, can have a nice & pretty enough loud master, while keeping the dynamics, the sound quality, the staging of the mixes ? And for me, I found the answers by learning & being an audio system engineer. I finally understand how to mix & master at -14 LUFS, -1 True Peak, loud enough without losing the quality. I don't know if its the correct way, but when i compare my final master with those old songs references, i think i make it. And i think that, yes, mixing is about art, taste. But we should also think about the logic, the science behind it. Without understanding the science, we could make a good art, but not a masterpieces. That's my opinion.
i always aim for -10 to -9 LUFS. -14 always seems so weak to me and whenever i check the popular songs in the style none of them have been that quiet so i haven't had a real reason to do it
This is the way. You can also still get decent dynamics at that level. I also think some level of squishing stuff sounds good as long as it's not totally hammered.
Yup. I look at -14 as “don’t go below that”. Sometimes with a ballad type song, since it’s lufs of the whole program, you’ll be lucky to get it to -12….with the loudest parts hitting -10 or -9, and that’s fine.
that's the range most pros go for. They respect the -1 dB, but they don't care about TP or being at -14 dB LUFS. They'd rather be turned down by spotify. They could turn it down themselves, but all that means is now they'd have to export an additional version of every mix.
Then Chris Gehringer shows up with it’s Dua Lipa masters at -0.05 dBTP / -6 LUFS clipping on all True Peak meters and it sounds fantstic. Same applies to other ME such Mancini, Hutch, Merrill, etc… I think music should not follow any recommendation/specification. Love ❤️
As long as you listen to those songs on devices that can handle inter sample peaks you should be fine. If the converters don't have enough headroom you will get distortion. Loudness should be chosen by the listener, not the mixing engeneer. Having a standard to follow is a good thing.
@@victorgarcia6912 the problem is consistency. Some people will get distortion, others won't depending on their devices. Usually you want a good sound for everyone.
When you encode an mp3 from a minus 1db true peak wav file the generated mp3 will result in -5 db true peak depending on the encoding options like 320kbps o 128kps etc, and they do encode it in different formats because, for instance, when your internet connection is slow they will play the smallest file etc. Spotify, for example, has an option on the player that you can find in settings called Normalize volume that is ON for default and will normalize your track to -14LUFS but if this option if OFF you will hear the file exactly as it was delivered. Cheers
Loudness normalized playback is great for listening to a bunch of unrelated tracks, but it ruins albums so being able to disable it is still important.
I use transparent clipping, because it sounds better. Sometimes it even sounds better if i clip it more than transparent. I do it because the sound changes. If i compare my older tracks, when i didn't use clipping at all, i notice, that they have mixing problems that wouldn't be, if i had used clipping. Clipping is not bad per default. It depends heavily on the style. I even used clipping to some extend while mixing for vinyl and it came out perfect. This vinyl sounds exactly like i've intended it to be. I didn't clip the master for vinyl. But when it is not for vinyl the tracks sound so much more powerful on a club system, when i clip almost everything. I had the chance to compare it. My fucking loud tracks weren't only louder than the quieter ones - they sounded better, even if i turned down the volume by a lot. I do it always how it sounds best. If it sounds better with -12 db lufs i leave it that way.
12:05 The answer is that in a lot of genres, the djing has become more basic, so the dj will play with the gains in the same place. So for instance in drum and bass, if you gave them a track at -14 instead of -6, you would find that on the beatport preview, it would be 8db quieter and the dj playing that out would need to turn the gain up. Now when you do that, there will be more lights showing on that channel in the mixer as the less limited track is more dynamic.. or if you "matched the lights" the more dynamic track would be quieter as the lights dont show LUFS. The interesting strange thing is that -14lufs is great for people who have lost control, such as listeners on youtube. The loud ones stand out, with the users who actually do have control, such as the djs, who you would think would turn up and down the tracks. Correct, louder masters sacrifice quality. Loudness war ends if you took away the control of djs, clip preview volumes and Label/ A and R people. Hence, Ed Sheeran Celestial distorting at the start of the track. That team of control listens to volume rahter than quality (see a and R room).
As a DJ/Producer who is trying to break through in my genre which is asking for -8 minimum LUFS, it comes down to being able to mix my own originals with “industry standard” (in my genre, not AES standards.) releases. Yes I can turn them up but the difference in compression changes the dynamics so drastically on these huge P.A. Systems that we run, that my tracks end up clipping before they can punch as hard. Perhaps this is due to something else I am missing in the mixing/mastering process but I’m not sure what that something would be.
Excellent speech Seriously One thing I noticed when I first started mastering to -1dBTP was the stuff I uploaded to SoundCloud simply sounded better, especially in the higher frequency range Also, if someone wants to listen to loud music, just turn the damn volume up But don't complain about your tinnitus problem...
The big reason to stop with the loudness gimmick is because it's making millions of people deaf, that includes us engineers. Taken from that same AES document : "If Portable Media Players (PMPs) are played too loudly for too long, they will cause hearing loss. The Scientific Committee on Emerging and Newly Identified Health Risks estimates that between 2.5 and 10 million people in the EU are at risk of developing early hearing loss as a result of listening to PMPs [8]. Europe was the first region to implement regulations to protect the hearing of PMP users: any PMP and headphone sold in Europe must comply with CENELEC EN 50332. In the original 2013 version, compliance was achieved by limiting the maximum gain of the device. This successfully reduced the maximum SPL from PMPs sold across Europe, but prevented quieter audio content from being played loudly enough to be heard under demanding listening conditions. A second adverse effect was that content louder than the EN 50332 Reference Test Signal still played at dangerously loud levels."
I don’t think i would like spotify to remove the normalize toggle - It’s really useful to be able to reference between your daw and spotify, if you’re like me and mix into a limiter ofc. In any case it’s interesting to hear what the mix/master for a certain track actually sounds like before the loudness penalty :)
@@att6915 No. Just no. Limiter is NOT a fix for bad mixing. Limiter is a LIMITER for the individual small peaks that good mixing cannot get rid of without introducing other problems. It is the last resort. Never do the easiest thing. Always strife for doing the correct thing. And, of course my comment is context sensitive. I'm not talking here about ALL usage of the limiters. I'm talking about it ONLY in the context of getting the correct true peak value which is just a very small segment of use cases.
I think people put too much stock in streaming services. The loudness war is not being fought on spotify. It is being fought in live venues on huge PA systems. If you want your edm track to compete in a live set back to back with other tracks mastered up at -8 to -6, your track better be hitting that target as well.
I love this subject as I had been discussing about it on forums as many loudly mastered EDM tracks are placed around -4.5dB to -5.3dB LUFs and are thusly exceeding that -1 true peak. I always refer this back to the medium as the tracks we are making are intended for live DJ performances, but are sacrificing this standard on streaming platforms, but I'd personally prefer my tracks to sound linear to the tracks I play than having to level match using MASTER level faders on a DJ mixer. Again, one needs to be aware that the sacrifice it gives when not following the guidelines.
Having loudness normalization was about the best benefit when switching from DJing with CDs to using my laptop. Though I still think using CDs it was far easier to navigate through my library.
All I can say is: download some successful tracks from wherever. Open them up in Audacity (it's free) and take a look at the red that appears. Red is where the reconstructed waveform goes above 0bB. You might be surprised. Note, however, that if you open the wav file vs the mp3 file, you'll likely get slightly different results. Some wav files will not clip when the equivalent mp3 will. In any case, you'll learn a few things. Like, for example, that you probably can't hear the distortion of the true peak overs. Or, you might even *like* that distortion (because we're used to it at this point).
===> means streaming/converting below 256kbps encoder may overshoot at levels above -1.0 dBTP "Codecs Safety limiting should take into account the peak signal level anticipated at the decoder’s output, which may be higher than the peak level at the encoder’s input. High-rate (e.g., 256 kbps) coders may work satisfactorily with as little as -0.5 dBTP for the limiting threshold. Typically, peak overshoot increases as the bit rate decreases, so the limiting threshold may need to be reduced below the recommended -1.0 dBTP."
The problem is now is not so much just conditioning but the systems we play on are not all build for dynamic range. Cheap amps and speakers (ie phones, laptops, TV speakers, little Bluetooth speakers) can’t handle dynamics well. I don’t like mixing/mastering for specifics but if the majority are playing on those systems what choice do we have! Get everyone on Dolby Atmos and we have a solution, however I can tell that is an even more confusing subject for most punters who still don’t really understand that it’s not just another surround sound medium.
The problem for me and why I tend to master very loud is because I produce and mix/master drum and bass music. Sometimes I get a reference and even tho I am at - 6db lufs, it's still not enough with some tracks.
@@Whiteseastudio the problem comes with Beatport and dj mixes, a lot of djs (and I know personally some) won't play tracks that are not loud enough compared with other tracks in the same genre. I know they have a gain knob on the mixer but some don't know that.
Sonible True, says my mix has +1.5 db truepeak. I have a 2x FF-L2 as last on my chain. How is it possible the limiter passes audio above the threshold lvl?
Interesting to see a mainstream discussion about this, I came to this epiphany about 3 months ago and realized the respect for dynamics over volume. I hate limiters. In essence leveling is the most important fundamental of mixing.
I know that I ask for a lot of my rock tracks to be not as loud, because I want to subconsciously inspire the listener to turn it up, a quintessential part of rock n roll to me. Better that than to have them reach to turn it down, anyway, imo.
I used to turn the toggle off on Spotify in regards to loudness but the amount of engineers who ignore the recommendations drove me crazy. It's on now and I am not looking back. My next step is to send a complaint to the regional TV station because well, it's way to loud in comparison to any other commercial TV station. Maybe I'll forward this video to them.
In any genre that uses alot of sub frequencies like EDM, Trap, Hip Hop, etc -9 LUFS or louder is standard. End of discussion. These genres are played in clubs and in car stereo systems that are about loudness and impact. You can still have good dynamics and not overly squashed waveforms.
Noob question. How do we deliver a track at -1 dB True Peak? Setting the Output Gain on the master limiter to -1 dB? The last plugin in the chain on the masterbus
By using a true-peak (oversampling) limiter, and by using a true-peak meter. You can set the output gain, but you have to check the true-peak meter for -1dBTP
If you’re using a DAW like Logic, its 32 bit float internally. You cannot clip a 32 bit float signal. But you might clip a analog style plugin. I also doubt you can clip a 24 bit signal. I think we’ve all mastered files so loud that the tops are flat. You don’t hear any clipping. But that’s not a good thing to do.
the interesting thing with TPM is that every measurement method produces slightly different results. even if you look at the loudness meter in Cubase vs their own insight tool...
I just checked a track that's been bothering me due to its loudness (Tides - Ed Sheeran) and as expected it is normalised to 0dB and true peak shows as 0.7dB on the meter I'm using. This was ripped from CD as 320kbps mp3 so PCM may be slightly better. A quieter track on the album is normalised to -1dB and keeps the true peaks to 0dB. It seems that they've used brickwall limiting and clipping as a creative decision.
@@Whiteseastudio I don’t know what that is. Maybe a topic for a video. 😉 But let’s say you have a file at 0dB true peak. Can’t you just turn that down 1dB? And if so, can’t the streamers do that?
I see a lot of pro’s talk about -8db lufs. According to them a lot of popular songs are mastered like this. So why are people still sticking to -14db lufs??
Cause it's the most compliant for a good variety of dynamic ranges among most genres, that's the main reason films are around -23 and such. And still as you can see in the linked paper, recommendations are at -16...
Music at -8db LUFS is not good audio, it is a distorted mess without clarity or making deafness as marketing. "We have the loudness song ever because we have bad monitoring equipment, musicians playing live with lackluster audio quality and producers that have deafness" .
I do here some people say that. Worst though is some who would do this and say one should reference at -8 LUFS and when I look at their meter, they are using momentary. Like really??? why? I thought its already been cleared that when it comes to this topics and this tech specs and requirements are always LUFS Integrated. they might as well just use peak metering if they just keep using short term as the basis for their LUFS.
@@gregrodrigueziii8075 Short term and momentary are better indicators of if your pushing too loud. You should check Intergrated to see what normalisation will do though.
i loved what u said about (considering to look inside of urself conserning the obsession over loudnes), cz really...it aint what can make ur track STAND OUT.
Major 🔑:: the true secret that no on talks about is dynamic range. If your your dynamic range is between 6-8db you can make the music as loud as you want. Once you start to fall outside of that range your music will sonically sound like shit. So. If you can get it to - whatever db and your dynamic range stay in between the range I just mentioned it will be okay.
As someone who doesn't have the money to pay someone *else* to do this stuff for me, I appreciate your clear explanations on how to keep things in spec so they don't sound worse online. Thanks!
I've seen several respected sources say to keep your peaks when recording no louder than -12 dB for the best results. I'm just getting back to recording, and up until now I would always set just under clip, and it never has sounded that great
I Master louder than -14 , like about -10 because drum and bass just needs a bit of squashing, but I always check that the true peak stays between -1 and -0.5
I studied for live audio engineering (among other event technician stuff) and had just one small course about studio audio engineering. This thing about true peak was one of the first things we had to learn. I'm genuinely surprised if these are not a well known facts. Granted, probably most audio engineers are self taught so there can easily be some specific gaps in their knowledge, but still...this is the basic stuff. Nothing advanced knowledge.
true and I self educated myself, but since I started, one of the first topics I stumbled upon was loudness, how it's acquired, why there is loudness, what loudness means and I also learned about the loudness war going on since years. So if I will recieve some audio from customers one day, I'll know what to do, to grant them a well balanced result and not an overboosted mix, that peaks, clashes and completely destroys the endusers ears...
I found that -1 dbtp is plenty to keep your files under 0 when converted to other formats… most of the times I lost headroom between 0.4 and 0.8 dBs when I did some backtesting on converting files to aac mp3 and ogg vorbis at different rates…
To clarify some points made in this video:
- The true-peak meter is to catch inter-sample peaks, or the peaks that happen between two sample points that otherwise go unnoticed on a regular meter, since the two samples are both under 0dbFS but not the crest that rises between the two samples. The drawing he was making made it look like de-clipping, or that a clipped waveform will get re-drawn with a new crest on top in the analog domain. This is not the case - he meant to draw the crest between two sample points, not to say that these will be added later during playback, but rather to show that these crests may already exist between various samples and are simply unnoticed by a regular meter. The -1dbFS headroom allows for these inter-sample peaks, as well as gives headroom for transcoding.
- Spotify does not attenuate any of your files that you upload. The only loudness attenuation happens when the user's client has "normalize volume" enabled in their preferences, as well as a "volume level" setting below the normalize option. A simple test will confirm that when normalize is disabled, loud tracks are very loud - far beyond -14 LUFS or anything close to that. For example, try "Hibou - Above Us" on Spotify and then grab a demo of Sonnox ListenHub and observe the LUFS. It's like -5 LUFS or something crazy if I recall. Extremely loud.
If the listener has normalize disabled as I do, and you upload a -14 LUFS track, your songs will sound quieter in comparison to most louder tracks since nothing is modifying the music. However, when normalize is enabled, the louder tracks can lose impact as they are attenuated in comparison to other songs that are not as smashed already. Also keep in mind that there are no loudness settings on the web version of Spotify and so web versions will all play as loud as they were uploaded! You can test this to confirm as well.
If you have the Spotify app open while you also have the web player open, the web player will inherit the loudness of the app until the app is closed. Close the Spotify app, go back to the web player, and the song will be played back at full loudness.
Agree. There are some concept errors in this video.
Very valuable information. Thank you for sharing 👍
100% with you on the Normalize Volume settings on Spotify. As long as Spotify (and other streaming services) have this settings as an option, I cannot think why people would want to master and export their final track at -14LUFS for example. Mind you, i am not hating on those who do, i just don't find it attractive as an option since electronic dance music (in which i specialize) competition is so far away from this standard still. Good insight on the Web and App based version though. Did not know this.
Good mix won't lose impact even after normalisation, although it might be genre dependant. In busy metal, which I mix most of the time, if you balance and process tracks correctly (mostly dynamic-wise) I start to limit peaks only around -9, so when I'm near -6 I'm limiting some sources just close to -0,5. Anyway I use normalisation on spotify and believe me that none of the extreme mixes lose their impact. The only scenario where I'd believe that impact could be lost (in metal at least) is the case where someone mixes really loud from the start and loses perspective.
Wait, the true peak thing. Isn't he demonstrating what analogue converters do to the flat part of the signal? AFAIK, the flat part is like a square wave which depending on the duration, it can have frequencies higher than Nyquist. Some analogue converters are not designed to handle this, so it creates artifacts (both from not being able to handle it and having a low pass filter or shaping it to be inside Nyquist).
Daft Punk threw a wrench in the Loudness War with their last album- basically mastered it for vinyl and left it alone. Beautiful record, zero ear fatigue.
I can say, that one thing, that drives people to get stuff loud to the level you mentioned is a competition of other releases on labels, that they are trying to get signed on. For example standard loudness on a lot of Liquid DnB labels is around -8 to -7 LUFS. People listening to the demos take the loudness into account. It is hard to steer away from pushing stuff loud when you are basically conditioned to do it since the beginning.
Loudness does matter tho
At work they have to crank it just to hear and some songs don’t cut it
We are not going to keep adjusting the level to hear at least not while working
Especially some of the dynamic range differences
Too quiet and it disappears
-7 Lufs is easy
I believe that labels look for dynamics/loudness ratio, like the dynamics preserved within that loudness target. A better example can be seen in Neurofunk, where generally you see more filters "movement" compared to Liquid.
@@davidpereira4455 I agree and perhaps I could have write that in a better way. Although I still think, that releasing electronic music at the top tier level labels slowly conditions one to push it loud - sometimes way too far. I have been in discussions on GS forums and there was a interesting talk about this topic. Some of the artists and engineers there said, that not shooting for the hot masters basically means career suicide in certain genres like hip-hop for example. I have fallen in the same mindset when making my own music - coming from trance and house genres, all thanks to the competition and release ambitions. They - the labels - need DJs to propagate the releases and they often test the track in a mix, and if it dips in loudness to much, it is a problem - DJs might consider that a bad sound and not suport it.
@@iamgeorgesears Yes man i agree 100%. Sucks that our creative mixing gets limited like this. I wish you all the luck and ambition you can summon up 💪
There's just one big true here.....When BIG top charts engineers stop doing their masters at -8 / -9 LUFS and begin to follow the platform specs, everyone else will follow (including me)! Until then, we'll always be in the loudness war! That is.
We deliver all of our music at approximately -9.8 to 10.5 LUFS and between -1.4 to -1.2 true peak (it varies a little depending on what the sounds best on the song). This gives users some room to turn it up a tiny bit in case they need it. We've had zero problems or complaints. I know our loudness is a bit higher than requested but as streaming is a smaller part of our business, I'm perfectly happy going this route.
Hi, I am exactly in the same range of values. Seems to be a nice compromise. I tried to be at -14LUFS, but then, on sites where there is no normalization... my hiphop and metal tracks were so quiet they were skipped by the listeners...
That's quiet. Most psytrance are -7 to -8 lufs.
@@AboveEmAllProduction We aren't making psytrance so I don't see the problem. We mostly compose ambient, ethereal, new age, some classical/cinematic pieces with other things thrown in when we get other inspiration. Writing for real estate videos or for corporate videos for example, we don't need -7 lufs. They're going to turn it down substantially anyway to get it below dialog levels.
Bruh, if you ain’t clippin’ you ain’t pimpin’
Cuz pimpin’ ain’t easy 😂
😂😂
I'm clipping, but I got no b.....😑
For real. I mix loud af. But still export at -1cause eff appul
You are a silly boy. Period.
Absolutely love this video!!! When i started out over 20 years agao i never gave a crap about being loudest. When i got back into music production it was all about being loud especially in the edm realm. I jumped all over it. I ended up making some of the worste sound music i had ever made trying to keep up with being loud. Over the last few years ibe gone back to just making music i like not caring as much as how loud it gets. I focus now on my tracks sounding good an more coherent.
This! 😘
There was a great Sound On Sound (magazine) article on exactly this issue some few years back 'The End of the Loudness War?' (February 2014).. and in my opinion one of the best articles to appear in Sound On Sound.. an absolute fantastic read.
And yes! The ending of the loudness war does mean that producers can breathe easy and produce music that has wide dynamic range once again.
that was a good article
Good article.. Agree too, dynamic range is where it's at
This article is just a classic.
It is not only about loudness. Loudness has it's sound. Driving the clipper and limiter harder produces more aggressive sound due to pops and cracks and also side result is lower dynamic range - compression. So if you master at 6 lufs and upload it to spotify no problem, it will automaticaly turn down whole song to average -14 lufs... but you will still retain aggressivnes of a mastering pushed hard.
there are so many songs that I love that happened to be -7lufs or louder...and I'm talking about major artists... good songs with awesome production and mix will sound good on any loudness threshold
Nevermind songs there's whole genres at that loudness
There are great tracks that hit even -3 LUFs... plenty. This is a conversation about genres and taste... either you like these genres or you don't. A number on your LUFs meter shouldn't be what guides decisions. If you are making loud music, it's going to be loud. And yes, there is a such thing as loud music... has very little to do with mastering limiters lol but this debate never ends.
No, they won't. As soon as inter sample peaks appear the fun is over.
I got a question about this topic. My friend is an engineer for Grammy artists and he told me to put the loudness on -9 db and the true peak at -2db. When I told him that everybody says -1 is enough for true peak he said it's true but that's if you wanna stay at -14 on the loudness. when you wanna go louder you should go down on the true peak and put it at -2db. How true is this and why?
I guess it's about the conversion algorithms, as Wytse mentioned in the video as well. Specifically compressed audio file types like OGG and MP3 will alter the sound significantly and to avoid any clipping after the converted files, you need to have the headroom. The louder your track is on average, the more likely it is to go beyond the 0 dB true peak after conversion if -1 dB is used on your master track.
Spotify does indeed recommend to deliver at -2dBTP for higher loudness deliveries
BRAVO man! I like when you do longer videos and talk about stuff other than plugins.
I check the website tunebat to see what the lufs are for my favorite tracks usually they all land somewhere between -8 & -6 lufs, the random access memories album which is one of my favorite albums ever actually had an average of -10/-9
While artists like Kaytranda come in hot at about -7/-6 so if the pros are loud I’d say be loud as well just don’t go overboard
Hi Brain, where I can see the lufs on the tunebat?
I agree but will also say that there is nothing stopping you from making different mixes for streaming per their recommendations and a loud one for physical media. If you only want to make one- go the RAM route and you'll be safe.
It's funny that it's actually something I found out by myself when I was still a rookie at music production 10 years ago. Back then I'd make my tracks as loud as I could, but I would always make sure I peaked at -0.1 instead of 0 because I hated seeing pixels above the limit when I opened them in Audacity.
Since then, I stopped brickwalling my music, so I don't have to care about competition and how overlimiting will affect my tracks (almost always in a negative way).
I have no interest in mastering louder or quieter than anyone else. I LIKE the sound of loud as shit, and I LIKE the sound of very dynamic. I just do what the track needs, or what I want to do. I make intense aggressive music a lot of the time, so I master very loud. Of course I take a lot of care of transients, but I saturate a lot, and I mix so that there isn't much gain reduction needed in mastering. But when I master a smooth jazz track, or drone or something, there isn't a whole lot of dynamics reduction. So, for me, it's more of an emotional resonance with the music, and an intuition, rather than some kind of competition. The competitive aspect of loudness is so so so dumb. It's all ego. Screw that stuff.
If it fits, do it. Just an aesthetic that some people and producers enjoy
Facts.
Sometimes its too dynamic and leaves me turning up the quiet parts. Theres a happy medium for a bit of squashing
100 percent
Right on!
With dance/club music, it's important to match the loudness of other producers because 1. It's not really comfortable for DJs to always match the volume for every song; 2. Your song, no matter how good it is, if it's not mastered properly, it will sound bad being played in the set after a louder and better-mastered one. I've been a producer for more than 10 years and also always been obsessed with the loudness issue. However, eventually it's very difficult for amateurs to mix/master tracks properly and make them sound good and loud, so in a way, sounding loud but good is like a threshold that separates pros and amateurs. I always had to match the loudness level of my productions to famous tracks, or they just wouldn't work when I play them, and it hasn't been easy. I've been a DJ much more than a producer. I get a track I love and it's quieter. I'm thinking, ok, just to remember to raise the volume up, however, it just doesn't worked after most tracks, even with the volume up. I know, you'll say, oh but the mix was not good, if the mix were good, it wouldn't have to be loud. But having to match the track's loudness to famous songs actually forces you to do a good mix and master because there's no choice. Also if I enjoyed the track, then the mix was not necessarily bad. I think, there is much more to commercial loudness than just standing out on the radio.
1. Noone listens to radio. 2. I hear a cheesy dance track that's too loud with not dynamic range I turn it down to match my vibe, then it feels lifeless because it has no dynamics and I skip it. 3. You'll only get love for your music if you have good ideas, melody, structure, philosophy etc. 4. Streaming is set to auto level then my ipad auto levels etc.
5. Get off the loudness train.
if you as a DJ are so concerned with loudness matching every track then it sounds like you should add a limiter to your signal chain and destroy every track’s dynamics on your end
Yes, it's true - but famous house tracks often come in around -9.5dB which is not too crazy loud
@@sqcaraudio He's not talking about radio. Not liking some music is fine, but the point of the OP remains the same. Some genres have loudness-competition built into them due to the nature of the industry and the venues we play at. Some genres are worse than others, mainly bass-oriented EDM, but loudness is important for most DJs that play at venues larger than 500 people. Clubs skimp on the systems and you have to work with the limits they have. Sadly that means some tracks won't work in context of others. I'd love higher end PAs and more dynamics in the tracks, but c'est la vie. Optimally you'd have masters for clubs and masters for streaming, but who can afford two masters of every track.
@@bananermat3798 can’t djs just use a limiter to loudness match on the fly instead ? i imagine that’s what most french DJs in the early 2000 did when incorporating 70s funk records into their sets
The ONLY reason streaming platforms request -1.0dBTP is due to intersample peaks caused by SRC algorithms in streaming codecs. In reality, these can exceed 6dB or more in some circumstances, but Apple pragmatically chose -1.0dBTP as part of Mastered For iTunes and streamers followed suit. Many argue that intersample peaks are inaudible, which is why some mastering engineers continue to use -0.3dBFS (i.e. some still rely on Full Scale and ignore True Peak, which is essentially an oversampling algorithm to predict intersample peaks).
The main reason for adopting -1.0dBTP for WAV masters is due to subsequent issues with lossy codecs (i.e. SRC algorithms when converting WAV files to other formats). This doesn't completely eliminate all intersample peaks, but it's sufficient for most.
In other words, adopting -1.0dBTP is a very useful, pragmatic metric for multiple reasons, but some diehards disagree.
Intersample peaks are only reproduced during analog conversion. This will happen at any digital level and it simply cannot be avoided. The DAC can handle a much higher voltage than the nominal corresponding voltage for 0 dBFS, and since the current is alternative, it will always oscillate to reproduce the true peaks.
Of course if you push the true peaks very high (artificially, because it’s technically impossible to do that with just music), then there will be a limit to the maximum voltage the DAC can output. AC cannot be a fixed voltage unlike DC, so it will not clip flat either, instead it will oscillate very fast and create unwanted frequencies.
Please leave the true peaks alone, this is what gives you that extra volume extension that none of the streaming platforms can detect. In any case, one sure thing is that if the loudness of your song is already squashed to -9 LUFS, it isn’t a true peak of +0.5 dBTP that will increase the max volume…
@@louisrmusic it's you missed the whole part about Apple music. In addition to streaming artifacts, they also apply processing/limiting at -1DB
@@JakeyWakey Limiting at -1dB Digital has nothing to do with the true peaks... True peaks will still happen when your speakers are reproducing the sound, this is inherent to the physical properties of an analog waveform...
@@louisrmusic yes and those have been happening since the dac and no one's been comparing or noticing audible differences, but that 1db apple chop is offensive.
@@JakeyWakey Wait, is Apple actually limiting the audio or just normalizing it ? That’s not the same thing.
The basic vibe I feel here is people's insecurity in wanting to be famous or signed.
Having been signed 20 odd years ago and now being retired I'll give you my take on this. All A&R peeps will use a volume knob when hearing a bunch of demos. They just will. They are not idiots. They won't hear a bullshit track that's -6db rms in old terms and go "wow that's the loudest track sign it nowww".. give me a break.
Personally mastering some later releases I chose -14 and -1tp and it gives me much more satisfaction in hearing crisp snares and punching kicks and in general more sparkle.
Very cool how you apply empirical data to a creative science. Results need to be measurable, repeatable, and most importantly-audible!
Very well said, I just downloaded the pdf and I'm gonna read it and seek further info because it's a very interesting and dividing subject. Keep up the amazing work, your channel is GOLD!
It's a bit weird that it is dividing. Standards are made to unite and make things easier for everybody 😅
@@Whiteseastudio You are right, but imagine being told you are doing it somehow "wrong" and get caught at being ignorant... Guess for some personalities this is hard to swallow. 🤷♂️
Florian Camerer was my teacher back then in 2007 (the guy who co-invented those new loudness norms and measuring) and mastering at decent- lower- levels was so normal for me that I used to watch that loudness war from a distance, scratching my head.
@@Whiteseastudio Yeah I totally agree, but people who have been doing it their own way for years and actually know what they're doing, a lot of times claim that those ''rules'' are ''crap'' because they could possibly ''kill'' individual preference and signature and stuff like that. Tbh i don't understand no of that, I think most of the times giving shit to these ''regulations'' is a way to boost your own ego and to feel like you ''know better than them..
It would be cool to see you compare and analyze some top 50 tracks regarding loudness etc
You mentioned A&R managers towards the end, and I truly believe they're the reason the loudness wars is still a thing. When they receive demos, the demos aren't played in end-user streaming services, but via some file-sharing or cloud storage services, which very much does NOT do loudness normalization. It should also be noted that soundcloud, which is where the loudness people mainly upload their music, doesn't offer loudness normalization to all its users
Na it’s the listener that matters
Interesting, I noticed there were huge differences between songs loudness on SC but didn't know it was due to them not offering normalisation to all of its users.
I now use Soundcloud as a place to test and compare my music to pro reference tracks.
Great idea
Soundcloud also compresses its tracks to shit. They sound awful.
I have tried several times to follow the recommendations of the streaming platforms and doing that...the tracks sound far below the references that I use (Artists that I admire). Of course you have to turn off Spotify's compression that evens out the volumes... to be able to perceive that.
I think that the benchmarks for ''BIG'' productions are guided towards what is a CD Mastering level... between -8 and -11 LUFS (approx.) and between -0.5 and -0.1 dbTP. When I started using these values everything sounded at a much more acceptable and satisfactory level to me! (it is also much more difficult to achieve without distorting). The platforms do not reject files with these values outside the recommendations...because the big industries that take most of the profits would make them undo that decision right away, they want the volume war to continue...and separate themselves from the ''amateur audio productions'' (ironically...there are missing dynamics in the ''PRO'' way)
Also...all these recommendations and streaming platforms could change or disappear one day, and I don't want my entire collection of music productions to be based on transitory things. Cheers!
I wouldn't call streaming services transitory. They pushed CDs out of use for good (check how selling CDs dropped). They became the vinyl of the industry now. People even stopped having any CD readers at home (especially on laptops and PCs - they're not even there anymore).
Everybody is accustomed to streaming, and that's a custom very difficult to turn back from now on. And it's been like that for years now. I wouldn't call it transitory. It's been well adopted.
@@---pp7tq I didn't mean that streaming as a way to consume music is transitory (but it could be) or that CDs will be imposed again. I meant that streaming services platforms and their guides values can change, even appear something newer than streaming. I mean that everything changes, and that's why I don't follow to the rules the ''moment''
Thanks for this video, it never hurts to hear this many times until people start getting it!
I follow the specifications to the letter and it’s funny because most of my tracks seem much quieter than most other comparable tracks, but my stuff won’t get penalized and still has dynamic flow. It’d rather keep the integrity of the music than JUST worry about loudness.
As a mastering engineer myself, I have been mastering to -1 true peak for a while, but 3 out of 5 times the client comes back to me and says it's too quiet. It's so frustrating, but you have to keep the client happy. I'm amazed how many people in the music and tv / film industry still don't understand the reason behind this, and still insist on having that extra dB.
In the TV industry too? they don't demand -23LUFS-i? Are you sure? Because here in Austria they obey that norm pretty strictly. I think same goes for other European countries.
You can educate the client using loudnesspenalty.com for instance
@@Mansardian tv mixes are -23, yes, but when I master music for tv (trailers mostly) they still want it maxed out at 0dBfs. Crazy, as they'll have to turn it down considerably.
I just master to -0.3. Apple has a plugin in tool on macs (and I’m sure there’s others) that can analyze your audio in various compressed formats for peaks and overs- the more compressed the codec the more overs. I find -0.3 true peak to be safe enough, it’s not like any streaming service is really going to use 96kbps these days.
Tell them to fix the mix next time.
One important point which missing the true peak get mostly in effect when music file will be encoded in a other format mp3,aac,oggvorbis etc. The encoding can also add little bit volume this is also why you not can see any peaking if you check your Orginal pcm wave file in DAW.
yeah, but why can i never hear the difference? If you listen to it in FLAC or WAV, it doesn't peak above 0, but when you export it in MP3, there it says for example on this song I just checked,
Sample peak 0.65. On some songs even higher but I never hear a difference?
Also True peak is always a little higher than the sample peak, i've never seen it the other way around.
@@asd2640 you can hear difference if you compare wav to mp3 if the peak is pushed to hard e.g. -0.3dB or -0.5dB. But this always can depending on kind of music you are listening e.g. Rock or Metall music where you often use tool like clipper.
I personally notice the effect on releases which I download from juno, of some of the release was more tired after listening. When I use the metering tools at wav and mp3 version it shows that they was pushed to hard for encoding.
I love this video, thank you 🙏
Excellent video and much needed, thank you! A discussion is also required regarding engineers recording too hot, trying to track close to zero dBFS for some reason, daring a loud hit or note to clip, when 24 bit recording gives us so much dynamic range (144dB theoretically) that recording with a peak level of -20dBFS and then compressing that signal 20dB still leaves us with way more dynamic range to play with than any studio's noise floor. You may say this is a hangover from the analog days when hitting a preamp and tape head hard was an optimal way to work, but some younger, digital-native engineers seem to have a mania for trying to get near zero (FS), as if it's some sort of race, like they will have a weak, feeble sound if the meter never hits orange. The loudness war starts at the preamp.
And starting a mix session with an iZotope RX De-clip party ain't no fun.
My masters aren’t loud even when I try lol but I do understand that perceived loudness really makes a difference when djing live. I’ve killed energy on the dance floor with bad masters or when i export a wip with bad limiting. But I agree, no need to fight the loudness war for streaming any longer.
Should we just export two masters?
Until every top 100 song on spotify really commit to loudness normalization nothing will truly change. I would love to follow the guidelines but I have to follow the popular songs that are the references and follow the client desires to match their favorite artist songs.
Who was first? The Chicken or the Egg?
The egg
@@Whiteseastudio heheh great analogy… the endless cycle….. I hear you, but really my hands are tied. Many of my clients do reggaeton, and those are played on clubs here in my country…. All of the clients want to match the loudness of Maluma, Camilo or J Balvin, if I don’t, they find someone else that do.
Exactly!!!!
@@oscarpatxot659 just export different masters (e.g. "streaming" "CD" "club")
In principle I agree with everything you’re saying BUT…. my competition at Sterling Sound and other Grammy Nominated mastering engineers are cranking out everything hitting-8 to -6 LUFS -8 hitting a limiter still sounds better than -14 on Spotify regardless if the normalization is on or not.
This is what I have found as well.
I work with Sterling Sound, and the masters I got from Ted are obviously louder than my mix, but also feel more dynamic and just sound better, and I can tell you, he's not shy on limiting!
Yup
Loudness is more than the number on your meters. Our brain recognizes compressed/limited signals as being loud because our ears do compress themselves if what we hearing gets too loud. So perceived loudness will always be a part of music production.
The thing nobody seems to want to talk about is lossy encoding and what the associated EQ changes do to peak levels. I've personally observed as much as 3dBFS additional peaks added by 192kbps mp3 encoding. Imagine what distortion that adds when your mp3 file is converted directly from a source that slams the 0dBFS mark with no gain reduction....
This conversation is tiring, and it's hard to explain to people that -3 or -4 LUFs is literally the sound of the arrangement and sound selection in these loud genres. It's not just a trick you do on your 2 bus... so it's a completely different conversation, one about style and tastes. You will never be able to make a Beatles track as loud as a tearout track because the arrangement and sounds won't allow for it. When we use synths or whatever to make our sounds and then process them to make them sausagey and fat... we are preparing to make a loud track from the very start with very full waveforms. If there's any "secret" to hitting those ridiculous loudness levels and still sounding clean and dynamic, it is to limit your subs and clip everything else as a general rule. You can do this with something like Kilohearts snap heap/multipass. If you arrange properly in heavy bass music, choose the right sounds (saturated and clipped/limited @ the track and bus level), you can hit -6 LUFs and still have dynamics and punch without much of a difference in quality with or without a mastering limiter/clipper. Getting to -4 LUFs from there isn't difficult with certain bass music genres/tempos. People that gripe about Loudness usually seem to mix everything but heavy bass music genres and aren't producing bass music themseleves. Done right, one can get most of the way there with arrangement and proper sound selection alone. There are many amazingly produced tracks out there hitting even -3 LUFs during the drops and it still sound clean, as proof.
Here the thing about loudness, I reference other tracks, and each genre has different loudness; I attach loudness to the style, which makes my track closer to the style I am looking for. Example Dubstep = -1 to -3 Luf, Pop = -8 to -12 Luf, Rock= -10 to -14 Luf. Once, I tried master dubstep to around -14, and It didn't sound the same; it sounded weak and hollow, which is why now I prefer to mix and master my song to a reference track, making life more straightforward rather than just wasting my time and money debating on something that clearly made to be broken because rule or recommendation is created to follow but also broken.
Ohh danm oldschool dubstep producers are crying now... is -1 to -3 LUFS really common in the newer stuff? I don't listen to that genre so idk but that seems very excessive, even for dubstep. Must sound horrible at that level. Up to around -6 LUFS I can still understand it, even if I prefer more dynamics. Goes the other way too though, beyond around -12 LUFS things become a bit too dynamic for most music to me and mix can fall apart with too little compression/saturation.
@@frogify_music yeah the dubstep scene is that loud now but just from my perspective it change to make it more dynamics now. I check most of the .flac version of most top dubstep producer the loudness is around -1 to -3 luf and nothing under that. I use to obsess with loudness nad master song around +2 luf which is dumb but good experiment.
@@frogify_music Also dubstep producer have a technique to work around the "over compress" sound which is the rule of play "one sound" at a time. Fun fact we only use compressor when need , most of the time we use a really good clipper " The Newfangle one". Also majority of the time they spend time sculpting space for the different sound. ✌🏾✌🏾
Thank you for this video! Very useful. I aim at -0.1 and NEVER read the instructions. Now, I will do what is expected!
Thanks for this video! We need more of these to help educate people!
when i would master to -14LUFS i was always told the music was quieter than others on a given streaming service. i've had that listed as the reason for being left off playlist etc. if you keep enough headroom you can have a lot more perceived loudness. ALSO, the playback is only normalized in certain situations. Also, listeners have the option to turn OFF normalization... bottom line, do what needs to be done
For bass ppl i always recommend Tipper - Mariscos it's a -10 LUFS track and great example to explain clients that by not getting -4 LUFS loud we can have more bass.
The end of the loudness wars is music to my ears. Welcome back to hi fidelity, timbre and dynamics.
A lot of interesting information. My only complaint is that making everything at -14 means the end user must use significantly more amplifier power on literally every playback device just in order to hear the music at a reasonable level for enjoyment. I certainly don’t appreciate “Loudness Normalization” toggles and I always turn them off. I want to hear the music at its intended listening levels, not have to crank it all the way up on my car or headphones from my phone. You can’t hear it loud enough in most situations. When done properly, you still have plenty of perceived dynamics. Don’t mistake me, I HATE overly compressed music, but when mixed properly, it can be very loud and not sound over compressed. -9 or -10 is still plenty dynamic. In fact, almost all my mixes before being mastered are sitting at -11 or -10 and barely need any final limiting to reach any sort of level and they still sound plenty punchy and dynamic. I think the loudness normalization is a scam.
Totally agreed. I don't think I can enjoy the dynamic and quiet songs in a noisy subway station. Also, even if applying the loudness normalization, songs may not sound even due to different energy distribution in frequency and those psychoacoustics stuff. So why bother turning the normalization on?
Mastering to a higher than -14 LUFS is not always about loudness itself. It has to do with dynamics perception. You want to listen to the dynamic classical record, or a medium-squashed pop song maybe a heavily compressed / limited metal or EDM record. Yes, at the end of the mastering chain you can technically turn down the file to -14 LUFS, but nobody does it bc then you'll need to use a separate master for almost every streaming platform, and loosing loudness when normalize is turned off by the end user. But they definitely don't want to hear a dynamic metal song or a squashed classical piece just because the recommendation is -14 LUFS so you'll give the snare that extra +3db peak when normalized. Also true peak -1dBFS is (as far as I know) only for the compression. When the platforms are streaming in low bitrate mp3 formats, the algorythm needs more headroom to not distort the audio.
And yes.... this message concerning loudness very worth repeating again and again... since even though the new recommended protocols came into existence as long as 8 years ago.. still too many fledgling producers know nothing of a) the importance of this and b) the meaning of it..
This whole issue is actually an absolute First and guiding issue.
No music producer should be without understanding LUFS... and why LUFS.
Hey Wytse, thanks so much for great and informative video's mate. One thing I would raise as a counter for discussion, is that in my experience a lot of client's still want physical production of their work and are not necessarily wanting different versions for streaming, once they approve the physical versions of the mixes.
Often these client's are not new to music creation and have previous releases at a given level of loudness. In this instance you don't want yours to sound quieter than the previous album. To client's it won't make sense. This is especially present in genres like rock and metal, where loudness IS still a thing.
Also I would argue that it is an art in itself to get a master to a very loud level (lets say -6 LUFS) whilst still retaining dynamic punch. At lower LUFS this is a little easier to achieve as you're not crushing the life out of it.
Anyway, thanks again for the great content!
As a DnB (Drum and Bass) artist, I usually master my stuff at -8 LUFS because it will translate in a club and I'd rather that streaming services turn my stuff down than up. I also use Pro L in True Peak mode to minimise overshoots.
what does "translate" exactly mean to you?
@@esteelvauce7288 None. It is a nonsens of course. Adding this "extra dB" cannot improve the sound in any ways.
For the loudness war to truly end, normalization needs to be forced on all playback mediums, and not just commercial streaming services etc. Because as you mention, the war is still in full swing when sending songs to A&Rs that they play from their mailbox or computer folder.
And since it's actually very difficult to take a dynamic mix of let's say -16 LUFS and master it to -7 LUFS without major shifts in especially the low end, many producers start pushing way early in the production/mix process, meaning the production itself is often designed around having a loud master. It's actually super common for producers to have a massive limiter on the master fader on their template, before even starting a track.
And because of that, the loudness war in the A&R room is also felt in the final releases since the A&R-approved production require a mostly intact loudness to not fall apart. So even with a mix engineer that knows his/her shit, they are still trapped.
The minus 1 dB peak point is there to give the system "wiggle room".
If somewhere in the chain the gain drifts above unity by .5 dB , it drifts up to -.5 dB, NOT .5 dB 'over'.
So in the past few years, I was a mixing & mastering engineer in my own studio, a mixing engineer for live band, an a system engineer for both audio company/live event/live band.
I used to question my self this : how an old song like Queen's song, Beatles, Etc, can have a nice & pretty enough loud master, while keeping the dynamics, the sound quality, the staging of the mixes ? And for me, I found the answers by learning & being an audio system engineer.
I finally understand how to mix & master at -14 LUFS, -1 True Peak, loud enough without losing the quality. I don't know if its the correct way, but when i compare my final master with those old songs references, i think i make it.
And i think that, yes, mixing is about art, taste. But we should also think about the logic, the science behind it. Without understanding the science, we could make a good art, but not a masterpieces. That's my opinion.
i always aim for -10 to -9 LUFS. -14 always seems so weak to me and whenever i check the popular songs in the style none of them have been that quiet so i haven't had a real reason to do it
This is the way. You can also still get decent dynamics at that level. I also think some level of squishing stuff sounds good as long as it's not totally hammered.
Yup. I look at -14 as “don’t go below that”. Sometimes with a ballad type song, since it’s lufs of the whole program, you’ll be lucky to get it to -12….with the loudest parts hitting -10 or -9, and that’s fine.
if it feels weak then loudness isn't the problem, the mix/master is; a strong mix/master is going to feel powerful regardless of volume.
@@nerds-nonsense A strong mix/master will be louder by default.
that's the range most pros go for. They respect the -1 dB, but they don't care about TP or being at -14 dB LUFS. They'd rather be turned down by spotify. They could turn it down themselves, but all that means is now they'd have to export an additional version of every mix.
Then Chris Gehringer shows up with it’s Dua Lipa masters at -0.05 dBTP / -6 LUFS clipping on all True Peak meters and it sounds fantstic. Same applies to other ME such Mancini, Hutch, Merrill, etc… I think music should not follow any recommendation/specification. Love ❤️
As long as you listen to those songs on devices that can handle inter sample peaks you should be fine. If the converters don't have enough headroom you will get distortion.
Loudness should be chosen by the listener, not the mixing engeneer. Having a standard to follow is a good thing.
@@johanneschristopherstahle3395 if you hear distortion, could be bad. If you only measure distortion, where’s the problem?
@@victorgarcia6912 the problem is consistency. Some people will get distortion, others won't depending on their devices. Usually you want a good sound for everyone.
Thank you for you in depth insight!
When you encode an mp3 from a minus 1db true peak wav file the generated mp3 will result in -5 db true peak depending on the encoding options like 320kbps o 128kps etc, and they do encode it in different formats because, for instance, when your internet connection is slow they will play the smallest file etc. Spotify, for example, has an option on the player that you can find in settings called Normalize volume that is ON for default and will normalize your track to -14LUFS but if this option if OFF you will hear the file exactly as it was delivered. Cheers
Loudness normalized playback is great for listening to a bunch of unrelated tracks, but it ruins albums so being able to disable it is still important.
I use transparent clipping, because it sounds better. Sometimes it even sounds better if i clip it more than transparent. I do it because the sound changes. If i compare my older tracks, when i didn't use clipping at all, i notice, that they have mixing problems that wouldn't be, if i had used clipping. Clipping is not bad per default. It depends heavily on the style. I even used clipping to some extend while mixing for vinyl and it came out perfect. This vinyl sounds exactly like i've intended it to be. I didn't clip the master for vinyl. But when it is not for vinyl the tracks sound so much more powerful on a club system, when i clip almost everything.
I had the chance to compare it. My fucking loud tracks weren't only louder than the quieter ones - they sounded better, even if i turned down the volume by a lot. I do it always how it sounds best. If it sounds better with -12 db lufs i leave it that way.
Serban mixes are -6 and true peaks usually above 0 and his mixes aren't bad. There is no hard and fast rule! Test it out
12:05 The answer is that in a lot of genres, the djing has become more basic, so the dj will play with the gains in the same place. So for instance in drum and bass, if you gave them a track at -14 instead of -6, you would find that on the beatport preview, it would be 8db quieter and the dj playing that out would need to turn the gain up. Now when you do that, there will be more lights showing on that channel in the mixer as the less limited track is more dynamic.. or if you "matched the lights" the more dynamic track would be quieter as the lights dont show LUFS. The interesting strange thing is that -14lufs is great for people who have lost control, such as listeners on youtube. The loud ones stand out, with the users who actually do have control, such as the djs, who you would think would turn up and down the tracks. Correct, louder masters sacrifice quality. Loudness war ends if you took away the control of djs, clip preview volumes and Label/ A and R people. Hence, Ed Sheeran Celestial distorting at the start of the track. That team of control listens to volume rahter than quality (see a and R room).
As a DJ/Producer who is trying to break through in my genre which is asking for -8 minimum LUFS, it comes down to being able to mix my own originals with “industry standard” (in my genre, not AES standards.) releases. Yes I can turn them up but the difference in compression changes the dynamics so drastically on these huge P.A. Systems that we run, that my tracks end up clipping before they can punch as hard. Perhaps this is due to something else I am missing in the mixing/mastering process but I’m not sure what that something would be.
Word up
I bought all the dB on my DAW’s meter and I’m damn well gonna USE all the dB.
The red lights at the top of the meter are the main cause of climate change. Save the planet, master at -14.
Excellent speech
Seriously
One thing I noticed when I first started mastering to -1dBTP was the stuff I uploaded to SoundCloud simply sounded better, especially in the higher frequency range
Also, if someone wants to listen to loud music, just turn the damn volume up
But don't complain about your tinnitus problem...
The big reason to stop with the loudness gimmick is because it's making millions of people deaf, that includes us engineers.
Taken from that same AES document :
"If Portable Media Players (PMPs) are played too loudly for too long, they will cause hearing loss. The
Scientific Committee on Emerging and Newly Identified Health Risks estimates that between 2.5 and 10
million people in the EU are at risk of developing early hearing loss as a result of listening to PMPs [8].
Europe was the first region to implement regulations to protect the hearing of PMP users: any PMP and
headphone sold in Europe must comply with CENELEC EN 50332. In the original 2013 version,
compliance was achieved by limiting the maximum gain of the device. This successfully reduced the
maximum SPL from PMPs sold across Europe, but prevented quieter audio content from being played
loudly enough to be heard under demanding listening conditions. A second adverse effect was that
content louder than the EN 50332 Reference Test Signal still played at dangerously loud levels."
I don’t think i would like spotify to remove the normalize toggle - It’s really useful to be able to reference between your daw and spotify, if you’re like me and mix into a limiter ofc. In any case it’s interesting to hear what the mix/master for a certain track actually sounds like before the loudness penalty :)
Great video - I struggle with this - I always end up around -.5db - I end up having to re mix in order to get to -1 AND get -14 lufs
@@att6915 No. Just no. Limiter is NOT a fix for bad mixing. Limiter is a LIMITER for the individual small peaks that good mixing cannot get rid of without introducing other problems. It is the last resort. Never do the easiest thing. Always strife for doing the correct thing.
And, of course my comment is context sensitive. I'm not talking here about ALL usage of the limiters. I'm talking about it ONLY in the context of getting the correct true peak value which is just a very small segment of use cases.
Its no problem as long as you are above -14LUFS
-9 or 10 LUFS always sounds loud clean and competitive with any track out there.
I think people put too much stock in streaming services. The loudness war is not being fought on spotify. It is being fought in live venues on huge PA systems. If you want your edm track to compete in a live set back to back with other tracks mastered up at -8 to -6, your track better be hitting that target as well.
I love this subject as I had been discussing about it on forums as many loudly mastered EDM tracks are placed around -4.5dB to -5.3dB LUFs and are thusly exceeding that -1 true peak. I always refer this back to the medium as the tracks we are making are intended for live DJ performances, but are sacrificing this standard on streaming platforms, but I'd personally prefer my tracks to sound linear to the tracks I play than having to level match using MASTER level faders on a DJ mixer. Again, one needs to be aware that the sacrifice it gives when not following the guidelines.
Having loudness normalization was about the best benefit when switching from DJing with CDs to using my laptop. Though I still think using CDs it was far easier to navigate through my library.
"Many audio engineers do not know this…"
1 second in: "Loudness and-..."
Me: So true, so true.
All I can say is: download some successful tracks from wherever. Open them up in Audacity (it's free) and take a look at the red that appears. Red is where the reconstructed waveform goes above 0bB. You might be surprised. Note, however, that if you open the wav file vs the mp3 file, you'll likely get slightly different results. Some wav files will not clip when the equivalent mp3 will. In any case, you'll learn a few things. Like, for example, that you probably can't hear the distortion of the true peak overs. Or, you might even *like* that distortion (because we're used to it at this point).
===> means streaming/converting below 256kbps encoder may overshoot at levels above -1.0 dBTP
"Codecs
Safety limiting should take into account the peak signal level anticipated at the decoder’s output, which
may be higher than the peak level at the encoder’s input. High-rate (e.g., 256 kbps) coders may work
satisfactorily with as little as -0.5 dBTP for the limiting threshold. Typically, peak overshoot increases as
the bit rate decreases, so the limiting threshold may need to be reduced below the
recommended -1.0 dBTP."
If you can get it competitively loud but still make people want to turn it up then you’re winning. Loud doesn’t have to be uncomfortable
The problem is now is not so much just conditioning but the systems we play on are not all build for dynamic range. Cheap amps and speakers (ie phones, laptops, TV speakers, little Bluetooth speakers) can’t handle dynamics well. I don’t like mixing/mastering for specifics but if the majority are playing on those systems what choice do we have! Get everyone on Dolby Atmos and we have a solution, however I can tell that is an even more confusing subject for most punters who still don’t really understand that it’s not just another surround sound medium.
Great topic covered in detail. Thanks.
This is super helpful information, great video.
The problem for me and why I tend to master very loud is because I produce and mix/master drum and bass music. Sometimes I get a reference and even tho I am at - 6db lufs, it's still not enough with some tracks.
They'll all be the same on Spotify...
-6dbfs? say hi to deafness.
Either you have deafness, a bad mix, or you have terrible monitoring.
@@Whiteseastudio the problem comes with Beatport and dj mixes, a lot of djs (and I know personally some) won't play tracks that are not loud enough compared with other tracks in the same genre. I know they have a gain knob on the mixer but some don't know that.
And then DJ's get upset when audio engineers call them stupid...
@@saricubra2867 listen to something new from Teddy Killerz for example and maybe you can explain to me why people are doing this.
Sonible True, says my mix has +1.5 db truepeak. I have a 2x FF-L2 as last on my chain. How is it possible the limiter passes audio above the threshold lvl?
Interesting to see a mainstream discussion about this, I came to this epiphany about 3 months ago and realized the respect for dynamics over volume. I hate limiters. In essence leveling is the most important fundamental of mixing.
limiters are great, but often used incorrectly or abused
I know that I ask for a lot of my rock tracks to be not as loud, because I want to subconsciously inspire the listener to turn it up, a quintessential part of rock n roll to me. Better that than to have them reach to turn it down, anyway, imo.
I used to turn the toggle off on Spotify in regards to loudness but the amount of engineers who ignore the recommendations drove me crazy. It's on now and I am not looking back. My next step is to send a complaint to the regional TV station because well, it's way to loud in comparison to any other commercial TV station. Maybe I'll forward this video to them.
In any genre that uses alot of sub frequencies like EDM, Trap, Hip Hop, etc -9 LUFS or louder is standard. End of discussion. These genres are played in clubs and in car stereo systems that are about loudness and impact. You can still have good dynamics and not overly squashed waveforms.
Facts I like mine -9 LUFS anything less it ain’t hitting right
You can be louder with the dynamic transients in the mix.
Will spotify go to -16lufs?
Be nice to have a couple of dB extra
Noob question. How do we deliver a track at -1 dB True Peak? Setting the Output Gain on the master limiter to -1 dB? The last plugin in the chain on the masterbus
Yes, also check using a loudness meter.
By using a true-peak (oversampling) limiter, and by using a true-peak meter. You can set the output gain, but you have to check the true-peak meter for -1dBTP
@@Whiteseastudio thanks guys, using Sonible Smartlimit as my limiter and Process Audio Decibel for the metering 👍🏻🤘🏻
What do you think of HoRNet Harmonics Pro? I can't seem to find any videos on it. Snake Oil?
This was very helpful, thank you!
If you’re using a DAW like Logic, its 32 bit float internally. You cannot clip a 32 bit float signal. But you might clip a analog style plugin. I also doubt you can clip a 24 bit signal.
I think we’ve all mastered files so loud that the tops are flat. You don’t hear any clipping. But that’s not a good thing to do.
the interesting thing with TPM is that every measurement method produces slightly different results. even if you look at the loudness meter in Cubase vs their own insight tool...
I just checked a track that's been bothering me due to its loudness (Tides - Ed Sheeran) and as expected it is normalised to 0dB and true peak shows as 0.7dB on the meter I'm using. This was ripped from CD as 320kbps mp3 so PCM may be slightly better. A quieter track on the album is normalised to -1dB and keeps the true peaks to 0dB. It seems that they've used brickwall limiting and clipping as a creative decision.
It can be mastered differently on streaming services though. And what you'll hear with their normalization enabled is also slightly different story.
I don’t get the -1dB true peak request. Couldn’t the streaming services just normalize to -1dB before converting the wav file?
But what does the streaming service do with bit-clipping?
@@Whiteseastudio I don’t know what that is. Maybe a topic for a video. 😉
But let’s say you have a file at 0dB true peak. Can’t you just turn that down 1dB? And if so, can’t the streamers do that?
You can use Sonic Visualiser for rendering waveforms, it is completely free.
I see a lot of pro’s talk about -8db lufs. According to them a lot of popular songs are mastered like this. So why are people still sticking to -14db lufs??
Cause it's the most compliant for a good variety of dynamic ranges among most genres, that's the main reason films are around -23 and such. And still as you can see in the linked paper, recommendations are at -16...
Because times are changing, the only need for -8LUFS (in most situations) was because of the loudness war, not because of audio quality reasons.
Music at -8db LUFS is not good audio, it is a distorted mess without clarity or making deafness as marketing.
"We have the loudness song ever because we have bad monitoring equipment, musicians playing live with lackluster audio quality and producers that have deafness" .
I do here some people say that. Worst though is some who would do this and say one should reference at -8 LUFS and when I look at their meter, they are using momentary. Like really??? why? I thought its already been cleared that when it comes to this topics and this tech specs and requirements are always LUFS Integrated. they might as well just use peak metering if they just keep using short term as the basis for their LUFS.
@@gregrodrigueziii8075 Short term and momentary are better indicators of if your pushing too loud. You should check Intergrated to see what normalisation will do though.
i loved what u said about (considering to look inside of urself conserning the obsession over loudnes), cz really...it aint what can make ur track STAND OUT.
Re: 12:20 “Why”? Colt Capperrune says that he gets complaints from clients if he DOES’NT master at -8 LUFS.
Major 🔑:: the true secret that no on talks about is dynamic range. If your your dynamic range is between 6-8db you can make the music as loud as you want. Once you start to fall outside of that range your music will sonically sound like shit. So. If you can get it to - whatever db and your dynamic range stay in between the range I just mentioned it will be okay.
Thanks, this is very informative :)
As someone who doesn't have the money to pay someone *else* to do this stuff for me, I appreciate your clear explanations on how to keep things in spec so they don't sound worse online. Thanks!
I've seen several respected sources say to keep your peaks when recording no louder than -12 dB for the best results. I'm just getting back to recording, and up until now I would always set just under clip, and it never has sounded that great
Great video, profesional and didactic. Thanks
Nicely said Wytse! ps: Did you use Darude Peakstorm for the track on the T-Shirt? 🤭
I Master louder than -14 , like about -10 because drum and bass just needs a bit of squashing, but I always check that the true peak stays between -1 and -0.5
I loved this video, thank you 🙏
I studied for live audio engineering (among other event technician stuff) and had just one small course about studio audio engineering. This thing about true peak was one of the first things we had to learn. I'm genuinely surprised if these are not a well known facts. Granted, probably most audio engineers are self taught so there can easily be some specific gaps in their knowledge, but still...this is the basic stuff. Nothing advanced knowledge.
true and I self educated myself, but since I started, one of the first topics I stumbled upon was loudness, how it's acquired, why there is loudness, what loudness means and I also learned about the loudness war going on since years. So if I will recieve some audio from customers one day, I'll know what to do, to grant them a well balanced result and not an overboosted mix, that peaks, clashes and completely destroys the endusers ears...
I found that -1 dbtp is plenty to keep your files under 0 when converted to other formats… most of the times I lost headroom between 0.4 and 0.8 dBs when I did some backtesting on converting files to aac mp3 and ogg vorbis at different rates…
Interesting info helpful!