Using normalise.

How to do this, that and the other. Share, learn, teach. How did X do that? How can I sound like Y?
RELATED
PRODUCTS

Post

Hi. When I render file into wav or anything,should I use normalise function or not? What is it for?

Post

hi, yes you should normalize all samples and stems,
that will give a nice level for mixing
when you mixdown, normalize to start with
-then use a compressor and limiter to master,
and normalize the mix when needed

normalizing brings the volume up to the loudest spike in the .wav or audio file

Google it

Post

normalizing (maximizing volume) gives you more room to adjust the main volume slider controls with
The only site for experimental amp sim freeware & MIDI FX: http://runbeerrun.blogspot.com
https://m.youtube.com/channel/UCprNcvVH6aPTehLv8J5xokA -Youtube jams

Post

Traditionally the "normalise" operation in audio editors, samplers etc normalises the peak loudness. This is good for maintaining highest possible quality when downsampling to another format.

But it may not give a very even "loudness" if both drums and pads are normalised by their peaks! Some programs like acoustica can normalise by LUFS which will get closer to convenient "pre-mixed" levels... but you have to watch out for clipping on transient-heavy samples!

It's a compromise.

Post

OMG, never normalize stems, terrible idea, because in the future if you need to re-sum them yourself, or if you need to send them to external mixing, mastering or re-mixing, the balance will be all wrong!

I also don't understand the logic of normalizing before mixdown, again, an exceedingly bad idea. The reasons are too numerous to even bother mentioning, just don't do it and be careful who you take advice from on the interwebs.

I actually don't think there's really even any reason to normalize, other than a raw sample before you bring it into a project. But even then, you may find, for certain sounds, all it means is that you'll need to bring it back down by 15dB again to sit in your project. Shakers at full scale? No thanks!
Last edited by Mr D on Wed Feb 21, 2024 1:07 am, edited 1 time in total.

Post

You may want to normalize tracks that go in a project so that the levels are similar when you start. Unless you've already decided the balances during recording. You also want to normalize your master to -0.1 dbFS.

Post

I used to normalise way back when, but haven't touched it for years now. Not really any need to. Not sure I see the point of it nowadays...

Post

Normalising makes your audio the loudest it can be without altering the dynamics or introducing overt distortion. It just turns up the volume so the loudest peak hits 0dB.

You have to watch out for ISPs (inter-sample peaks) though, as these are not detected by standard normalising algorithms. This is where the actual samples are still technically below 0dB, but there is actually a hidden peak between them that will clip the reconstruction filter. This is because samples are actually "connected" by curves rather than straight lines. Because everything is sine waves.

0dB will also cause clipping on some old CD players. But honestly, those are probably not still in service.

So instead of normalising, use an ISP-aware peak limiter and set it to max output of -0.1. I use IK Stealth Limiter.
THIS MUSIC HAS BEEN MIXED TO BE PLAYED LOUD SO TURN IT UP

Post

jamcat wrote: Fri Feb 23, 2024 12:13 am Normalising makes your audio the loudest it can be without altering the dynamics or introducing overt distortion. It just turns up the volume so the loudest peak hits 0dB.

You have to watch out for ISPs (inter-sample peaks) though, as these are not detected by standard normalising algorithms. This is where the actual samples are still technically below 0dB, but there is actually a hidden peak between them that will clip the reconstruction filter. This is because samples are actually "connected" by curves rather than straight lines. Because everything is sine waves.

0dB will also cause clipping on some old CD players. But honestly, those are probably not still in service.

So instead of normalising, use an ISP-aware peak limiter and set it to max output of -0.1. I use IK Stealth Limiter.
I normalize to -3 db (you can choose the value in Cubase) to avoid exaxtly this.
A limiter is not the same as it has a knee, and even a clipper will yield different results.

I don’t normalize everything, but prefer to have audio levels so that plugins are hit at appropriate levels. I don’t see the point if I am anyway going to pull down the fader at the end and I’m not using plugins that cares about levels.

Post

Normalize wherever you want IMO,
Stems can be normalized before mixing

Normalizing allows high fidelity by maintaining bit-depth

...and of course, the last thing that happens to a master is a normalize

Post

Don’t normalise your source audio. You don’t gain any resolution out of it, because you can’t add what was never there. There’s no resolution lost from fader moves or processing, either, because modern mix engines are 32-bit float or 64-bit double precision, so there is near infinite headroom.

Normalising is a destructive edit, which should be avoided whenever possible on source audio.
THIS MUSIC HAS BEEN MIXED TO BE PLAYED LOUD SO TURN IT UP

Post

I conjecture that it's not destructive,
it does not take any fidelity away
(...what information is lost?)
-I prefer to mix hot samples/stems,
and deliver a hot master

Post

I don't normalize much when producing music, but I've done it a lot editing audio for video/podcasts, especially if someone else recorded it. Most video software doesn't have the sonic sophistication of your typical DAW/audio editor. Once I trimmed files I would normalize to a high enough level that they'd trigger compressors/gates without having to add gain plugins, or tweak clip levels later in the video editor. And as @Mr D sez, normalizing stems will destroy the mix.

Post

normalizing pre-mix does not screw up the relative amplitudes,
because when we mix,
we set the levels
post mix, it does undo any mixing

Post

nix808 wrote: Fri Feb 23, 2024 6:19 am I conjecture that it's not destructive,
it does not take any fidelity away
(...what information is lost?)
The operation is a simple multiplication with floating points. Good enough for Rock'nRoll, but not perfect.

Image

So there is a tiny little rounding error introduced in the very least significant digit. That won't matter in practice. I know of no AD or DA converter that practically goes beyond 20 bits and these errors are in the 24th bit. Some people might care, most won't.
We are the KVR collective. Resistance is futile. You will be assimilated. Image
My MusicCalc is served over https!!

Post Reply

Return to “Production Techniques”