That's always a good idea.

 

I just said "iTunes will turn you down". If you have a tapped-out master that has zero dynamics and looks like a solid black waveform, YouTube, iTunes, Spotify, etc. turns you down so that your LUFS is consistent with their max. It doesn't matter how loud your master is -- in fact, my whole point was, if you have a super-loud master, it will sound like over-compressed shit when it gets turned down to -13/-14 LUFS. Especially when it goes up against a master that was only -19 LUFS with little or no limiting.

 

Sorry, man, I know I'm explaining this badly, so here's an article on the subject. And here. And a short Wiki entry.

Some services (like iTunes) may adjust the gain on your -16 LUFS track so that it comes in at -14 LUFS ... something you can actually do on your own, if you've ever used the "Sound Check" function with your own library. You can also push the gain all the way up on an individual track without going over 0dB, so there is some internal brickwall limiting involved to prevent you from clipping your audio outputs. There's even a Sound Enhancer function, which is basically a (multiband?) compressor. I'm sure how they treat music for streaming or download is akin to batch normalizing for RMS, rather than limiting or compression. If you deliver a quiet album, it's probably based around the loudest track (in other words, if your album has a track that tops out at -16 LUFS and one that tops out at -14 LUFS, nothing changes).

Most mastering engineers these days will master your stuff in the -14 LUFS range so that none of this is a problem ... unless, of course, you demand crushing your mix into a black bar. It's also problematic for vinyl, since all these over-compressed dope club mixes and whatnot will cause a record lathe to flip completely out. They'll end up just turning it down, which once again will amount to your black bar master sounding both quiet and overcompressed.

 

Last edited: Sep 21, 2018

They don't turn you up. Your mix stays at -19 LUFS where too loud mixes get turned down to -14 LUFS. You're still quieter. Definitely on YouTube, would have to test for other streaming. Don't know about CDBaby, but I bet not everyone does this.

This is almost true, but keep in mind the -19 LUFS mix stays at -19 LUFS. You should actually keep it at -14 LUFS thereabouts. It's also a decent level for CDs and other physical media.

Also, keep in mind the resulting LUFS should be judged by how it sounds at all times. For electronic stuff (neurofunk, drum&bass, dubstep etc) you naturally end up with high values because of the inherent lack of dynamics coming from the synthesizers and programmed drums. When I'm mastering I use a limiter with automatic gain compensation to have a fair comparison. I also try to limit just the drum peaks and stuff that just sounds over the top, not all the rest inbetween, which usually sits way lower. This is my starting point and then I wiggle the threshold around to see what sounds best for the song.

 

I was using a "-19 LUFS mix" as an example -- most people don't master their stuff at -19 LUFS. The average master is probably closer or above the -14 LUFS mark, except for the "make it loud" shit that gets mastered to around -4 LUFS. And whether YouTube turns your stuff down to -14 LUFS or well below that, the principle is exactly the same: once your brick-walled master is turned down from -4 LUFS, it will sound like shit. The only reason an overly-loud mix sounds "better" to you than a quieter one with dynamics is because it's loud. Take the loudness factor away and it sounds terrible.

Except that's not what YouTube is doing -- they're doing exactly what I described above in regards to iTunes, Spotify, etc. They're using an LUFS normalizing algorithm. See for yourself.

 

No, no, that wasn't explained badly, maybe I was a bit too picky.
I can remember when I first tinkered with this matter it took some time to get the hang of it. Although meanwhile I think that's all pretty easy.

And I think we agree that every song should, of course, first be mastered for audio purposes and then modified for the various streaming services to get the max result.

Btw, that -4LUFS example is still making my day.

 

I do! The whole reason I ever brought up LUFS at all was because I think that was the root cause of the OP's problem -- one of his meters in the pictures he posted (limiter, I think) was set to LUFS instead of dBFS and he rightfully got confused. LUFS is still a highly confusing issue, period.

As for mastering audio, especially if you're mixing and mastering on digital, yeah, stick to dBFS. I think most people post-Loudness Wars are mixing and mastering properly anyway, and what happens on streaming or digital delivery won't affect the sound too much. Actually, nothing's really changed at all, except that people with crappy mixes and masters are now going to be that much easier to spot.

 

You're going to need to lower the volume of all those tracks so that the redness stops happening at the top of your meter (top picture)

• In Ableton you can setup all of those tracks by clicking on any track to highlight, then hit command A to highlight them all.

• Then hit command G to group all of the tracks into a single group.
  
• Then in the track that says "group" at the top you can adjust the volume of the entire project with one fader.
  
• Lower that fader until the needles stop hitting the red area (top picture).
  
• Then put your limiter on the master fader and lower the "threshold" bar until the signal starts tapping the top. You should notice the louder sound you're looking for now without the distortion like when the signal is hitting the red area (aka "clipping") You always want to avoid hitting that red area to keep the sound clean and distortion free. If the sound is too big and things aren't sounding distinguised from each other anymore try decreasing the "release" on the limiter until it sounds better, or just raise the threshold a little bit until it sounds better.

 

Last edited: Sep 22, 2018

FYI: The virtual mixer on your modern DAW is not your granddaddy's analog desk. I understand Ableton Live for example runs internally at 32bit floating point. This means your gain staging can be absolutely fucking bananas and you are still not losing any dynamics or bringing up the noise floor to audible levels (unless third party plugins fuck you up). There are no technical reasons why you can't be going way past 0db at every stage on your way to the master channel and then just turn the signal down at the master channel.

 

Last edited: Sep 22, 2018

Now that is what I call an entrance!

Actually in DAW mixers, since they are digital the level isn't as important as the amount of signal running into an analogue board which would have more of an impact. You don't want to surpass -6db to give enough headroom to accommodate everything and then allow for mastering to make things louder, if you've already mixed everything at high levels it will actually make the final render quieter as its crushing everything down. However OP's issue has already surpassed that point and they are clipping. They also have some automation parameters setup on the individual tracks so the only way to lower all without affecting those would be to group them and lower the group volume. Then with enough headroom to master limiter the sound back up to the big sound they were looking for.

 

In your 32 bit floating point DAW mixer, you have almost unlimited "headroom" at all times to the point the concept kinda loses it's meaning. The meters and the zero db line on your channels are practically arbitrary. If you are hitting your master limiter too hard, you can just trim the volume down at the master channel stage or adjust the plugins input gain. Nevertheless it obviously makes sense to stay within reasonable limits for various workflow purposes, but there are not technical reasons to do that. User relexted was right.

In most practical applications, floating point is still used within a 32-bit framework, but with 24 bits allocated to the mantissa and eight bits allocated to the exponent. If you do the maths you'll find such an approach provides the utterly ludicrous theoretical dynamic range of 1500dB, which means you will never run out of headroom inside the processing, and never lose signal in noise floor.

So, as you found, you can increase or decrease the level to the most ridiculous extremes inside a floating-point system, and as long as you restore the gain to something more appropriate to feed the output converter's dynamic range, you will not suffer from the noise or clipping that a more conventional fixed-point or analogue system would. Which is very impressive.

Sound On Sound Magazine

 

This is absolutely correct. With 32bit FP you can even export a wave with +20dB without getting any clipping. But I have the impression we tried to keep that topic away from this thread so as not to confuse the OP even more. Therefore we told him just to lower all the channels until the master in-level drops below 0dB. Then he can apply his limiter and all is good.

 

But you're still dithering to 24 bit solid right?

 

Eh, nope. I don't do DVDs. 32bit FP or 16bit is sufficient for me, as well as a noise floor of around 90dB.

 

Please use more true-to-life examples next time then. We're talking target levels in LUFS, -19 and -14 is a whole world of difference.

True

The post says about how Youtube does scaling - if you have a master that's too loud, the example given:

means when you set your volume slider in Youtube to 100%, this track will actually play at just 54% volume. Also if you turn the volume to 50%, you get just 27% volume.

Here's an important bit from the post that you're repeatedly failing to acknowledge @Iggy :

It's important. You don't want to deliver a master that's too quiet. -19 LUFS doesn't cut it, "just don't worry and let the mix be quiet" without mastering doesn't cut it either. It won't be turned up. Your loudness will be stuck at 100% and you won't be able to have any more volume - which is a big deal for laptops, phones and pretty much everywhere. If your clients don't have any more volume left in their device's adjustments (which again is common on phones and laptops - you often have the volume all the way up), you're in trouble - they won't be able to turn your song up to be able to hear the song properly. It is a big deal that should not be overlooked.

 

This is almost true, but the problem is your processing might become bananas. All plugins that depend on level: meters, spectrum analyzers (including the ones in EQs), compressors/limiters/de-essers, distortion/waveshapers/amp sims/bitcrushers/exciters/saturators... they no longer work properly. Your meters don't show you anything because the scale doesn't go above 0dB so all you see is the graph right at 0dB clipped off. Your compressors and limiters rarely allow you to set a threshold above 0dB. Distortion of all kinds goes without question - its principle of operation is a input volume to output volume graph. More volume - more distortion. They are often not defined for above 0dB and for that reason you'll often find they're clipping everything above 0dB on the input.

Not to mention hardware emulation plugins, but you probably know that.

 

No mixes have ever been done at -19 LUFS before? Interesting.

I think you need to get away from this whole "-19 LUFS" thing before you burst a blood vessel or something. As I explained in my last post, it was cited as an example, and even in the original post where I mentioned it, it's obvious that I threw it out there as an example, not an absolute. At no point did I say "You have to master your tracks to conform to -19 LUFS" ... although, there are plenty of -19 LUFS (and even lower) mixes out there, especially before the Loudness Wars became a thing. Truthfully, most mastering engineers are now mastering tracks between -11 LUFS and -16 LUFS (unless you tell the ME otherwise), which is wise, as the standard for digital streaming is -13/-14 LUFS.

Finally, the concept I think you're failing to acknowledge is "normalizing". Let's go back to the -19 LUFS example, as you seem to be fond of it: You have a mix that tops out at -19 LUFS. YouTube's LUFS normalizing algorithm increases the volume to -14 LUFS. You have a mix that you crunched to -4 LUFS. YouTube's normalizing algorithm decreases the volume to -14 LUFS. Still with me? The two mixes are now the same perceived loudness on YouTube (or Spotify or iTunes or etc.). in other words, you can make it as loud as you want and it will still only be the same volume as that -19 LUFS mix on YouTube. They'll both be at -14 LUFS. That's what normalizing does. The difference at that point between the -19 LUFS mix and the -4 LUFS mix would be that one would have great dynamic range (the -19 LUFS mix) and one would sound like squashed shit (the -4 LUFS). But again, just so you don't give yourself a heart attack, the mix doesn't have to be -19 LUFS. It can be -16 LUFS, -11 LUFS, a straight -14 LUFS, whatever. But making it overly loud will not make it sound any louder on YouTube or any other streaming platform.

 

Last edited: Sep 24, 2018

good gainstaging...and there is no need to limit for the first , am i wrong ?
when i dont use analog emu i just look for at leat -3db headroom on master (fader on -0,1db) and good gain staging in the first..faders are only for arrangement ...you must use utility gain tool....

 

Last edited: Sep 24, 2018

They have been. They no longer make much sense.

By common definition it's setting the loudest peak in a piece of audio to a set level, often 0dB. Not all streaming services do this, as Youtube doesn't turn up, it only turns down, as I've mentioned several times already.

It doesn't increase at all. This is the problem. It will only turn down to -14 LUFS if it's louder than that. It won't turn -19 LUFS up to -14 LUFS.

Yes, this is accurate.

Again. In some services they are, in Youtube they are not. The -19 LUFS mix stays at -19 LUFS and the -4 LUFS mix is turned down to -14 LUFS. So you have -19 LUFS and -14 LUFS which is still a difference.

Again, they will not

But they don't do normalizing. They do something of their different, as explained in the post you linked to earlier.

The difference would be this if both were playing at -14 LUFS, but yeah, the idea is exactly that.

Of course it doesn't. It's an artistic choice. The point I'm making is to go for around -14 LUFS so people listening on their phones and laptops aren't angry because the song is quiet for them and they can't go above 100% volume.

Definitely. It's all about loud enough to be convenient, not as loud as possible.

 

So, again, what do I do if I still want my track to be as loud as hell but still upload it on You Tube?
Do I make two versions, one with a huge ass brickwall and one with more space?

Let's say I want my track to be at -8 LUFS, can I upload that to You Tube or Sound Cloud (not even sure what SC does to audios)? Or I'd make another version that has less loudness?