Spotify says negative two decibels. One mastering engineer will say negative point three. Another another one point two. Another another one point one. Another one another one negative one. That's about true Peak. There is so much noise but what exactly is the correct ceiling to set your limiter to, and the answer is it depends but not on what you might think. See, limiters aim to attenuate a signal before it reaches a nominated Peak threshold, and by preventing that ceiling going above the threshold, a signal can be raised to utilize all the additional Headroom, at which point a ceiling is set to dictate what the nominal peak level of the limiter's output will be. Now why did I say nominal? Because your nominal peak level or ceiling doesn't always correlate to the output peak level. Due to the ballistics of some limiters, they can overshoot several tenths of a decibel, and due to the nature of sample-based digital audio, there also may be some inter-sample Peaks present between two different sample points. In both of these instances

it will cause the output to go over the ceiling or the nominated peak level. Now, this is why you get various recommendations like negative 0.3 or 0.2 or 0.5. They're all guesstimates aiming to negate the side effects of limit and Sample-based audio. But I don't do guesstimates. I like assertive and definitive decision-making like you smashing the like button, clicking subscribe, and ticking the notification Bell. Without a limit of going over zero decibels would be a nightmare of clipping, distortion, and unusable audio. However, overshoots from a limiter and into sample Peaks are typically tenths of a decibel and span in the timeframe of single-digit milliseconds, sometimes longer. But the harmonic distortion introduced by digital clipping at this magnitude is around negative 50 decibels full scale below the nominal program material, which means it's often inaudible. When it is audible is in sustained passages of clipping or when

overs registers as whole integer values above zero decibels, where Distortion becomes really clear. So to audition this, what we're going to do is we're going to first add a hard Clipper after our metering. See, modern DAWs use 32-bit audio, so your signal never plays back clipping even when the meters are in the red. A hard Clipper at 0 Isabel's full scale effectively acts like the digital ceiling your audio will have when it's exported. Next, you want to add an AAC codec at 96 kilobits per second, and this is the normal audio quality Spotify streams to mobile, desktop, and tablet devices. When your master is transcoded to lossy formats, this is the process it uses, and it chews up Headroom and increases the Signal's peak level, so it's important to audition how these changes will sound on your master when a consumer hears it. So now we've got that, we're going to make three markers. We're going to find the loudest true Peak, the

louder short-term LUFS, and also make any markers for passages of sustained bass. I've already gone ahead and done this, and the max allier Fest is also the same passage with sustained bass. Here we're going to audition and identify both orally and visually how much of an impact the overs into sample peaks and signal above a ceiling is actually affecting what we hear. All right, so let's look at doing this. I'm going to start by sealing at negative 0.2 because that's usually where I have it at default. We have Ozone 10's codec enabled, okay, so that's enabled to 96 kilobits per second. We have a level meter so we can monitor our levels even though we're going to be using our ears to listen right now, and then after that, we've got a zero decibel hard clipping. So any signal going over zero decibel, we're going to assume it's going to be hard clipped, and that's why we've got that there for auditioning purposes. What we

can do here to sort of hear the effect of that clipping. We can bypass this on and off as well as visually look at what's being clipped down here. So let's monitor this. Firstly, this true Peak. I'm curious 0.1. What will happen there? Look at this peak level. Oh, that squashes a little bit. That kick there. Let me put that on the loop. They're very subtle. It's just a bit too hard for my life. I'm bringing that back down to point two. I don't think the Peaks are negatively affected by that clipping. I'm I know for a fact the base is. I know for a fact this low end here gets a bit funny. So let's audition this here. I wish now I know there's two sections of this audio. First is this sustain section and then this kick hit here for. No, I think that's not too bad. Let's have a listen here. Oh, it chokes it. Chokes the low end. That clipping when it hard clips that lowering gets really choked in. You can

hear that it's almost like that the energy in the low frequencies sort of like shift up a bit so that's ceiling at point two isn't doing me any satisfaction. I'm going to bring it down a bit at point four. At point four, that kick just sustains a bit bit better. I'll play it three times at point two and then three times at point four. Much better, much better. The clipping isn't as bad, like it's there but it's not bad. It's sort of works for the music. It's not negatively artifacts like crunching away. Okay, so we've monitored the true Peak and the Mac short term Alia fists typically. Now what I would do is I would take the limiter here, okay, and audition start to end. This is just your Peak and your Mac short term RMS section. So your loudest Point by average value, your loudest Point by Peak value, it just markers to set this ceiling. You still have to audition everything before you export it, but this at least for this particular

material. You know how I said it depends. 0.4 decibels is the ceiling based on what I've auditioned here. I might listen to the song, and there might be other sections where it's crunching up a bit, and I might back that ceiling off a little bit more, but this process will help you set your ceiling every single time consistently so you don't destroy your Masters. Because remember when I said at the start of the video, what exactly is the correct ceiling to set your limiter to, and the answer is it depends. This is the process. It depends on auditioning the overs and how the clipping sounds itself. Now I've left three elephants in the room. Firstly, Spotify's recommendations. Negative one or negative two decibel true Peaks. These are just recommendations because not everybody uploading to Spotify is going to be uploading masters from a dedicated mastering engineer. A lot of artists uploading to Spotify will Master it themselves. And to avoid emails from disappointed artists that their music is distorted during playback, Spotify gives

them a recommendation that is by far truly safe for them to adhere to. However, the process I've laid out for you negates this because you've already had the opportunity to hear the distortions present from the transcoding process and adjust your limiter accordingly. Next, Apple Digital Masters negative one decibel true Peak. Now, as a certified Apple Digital Master provider, my process for producing Apple Digital Masters is very different from what we've discussed, as it adheres to a very detailed and stringent guideline Apple has laid out for this certification. And to do this, I have a separate bus which takes the audio in from my A to D converter, and that goes through a true Peak limiter set to negative one decibel true Peak. And before delivering that file to a client, I convert the master using an Apple codec droplet and check for overs in the terminal to make sure there are no overs on the encoded format present. You're probably asking, Why use negative one decibel true Peak for Apple Digital Masters? Use a true Peak limiter.

everything you've said you don't want to do on these Masters why don't you just do that for all your Masters and call it a day. So, majority of my clients focus on genres like pop, hip-hop, rap, EDM, dance, and want to sound the biggest and the best, which means using every last bit of bandwidth available. And they're not necessarily concerned with pursuing Apple digital Masters as a distribution option. So, having a method to achieve big loud Masters in a careful and calculated manner is why I've presented to you the method I choose to go with. Finally, why all this, why not just use a true Peak limiter and avoid the kerfuffle. Well, I don't like the sound of true Peak limiters, and to be honest, I don't know why. Just when I use them, they don't work, and at least to my ear, they don't sound as good. But that's a topic for another video.

ARE YOUR Masters Being DESTROYED When You Limit Like THIS?

Hello. I'm Nicholas Di Lorenzo. I'm a mixing and mastering engineer and owner of Panorama Mastering Studio in Melbourne, Australia. I’m an Italian-Australian. born and raised in Melbourne.

This channel is for anyone involved in the music making process, artists, engineers, producers or musicians.

To stay up to date, hit the "bell" icon for alerts as I release videos weekly on all things audio, mixing, mastering and industry.

-

Want to learn more about what I do?
https://panoramamastering.com.au 
@Panorama_Mastering