Hello folks (and folkerinos),
Thanks for reading this guide! It will help you to decide how to master your tracks for streaming platforms in 2020. There is a lot of information going around regarding that so I decided to combine the most important findings into this short guide.
First of all, why does mastering for streaming matter at all? Modern streaming services now have something that is called volume normalization - an effort to make all the songs played roughly the same volume so the listener doesn’t have to fiddle with the volume knob every time they switch a track.
What is volume normalization exactly? It is a set of algorithms to decide how much your song will be turned down (or up). Most of them work roughly the same - apply a psychoacoustic filter (More on that later), scan the music file in small increments, and find the average/peak volume.
After that all the service has to do is to increase/decrease the volume - and viola, every track sounds as loud as the next one! Except if you ever used those services, you know that’s not exactly true. There are still variations of volumes? Who is to blame, and what can we do to avoid sounding quieter than our competition?
You’ve got to understand what those algos do. They either turn the volume down or use a limiter to turn the volume up. Or allow stuff to clip (Hi Pandora).
Unfortunately, for various reasons, there isn’t a single answer.
The first reason is that different streaming services use different algorithms to normalize volume. Individual targets are in the table below. Let’s look at them!
At this point, you‘re hopefully getting a bit confused and uneasy. So what is going on, what is the optimal level?!
Here’s the thing I’m trying to convey: Don’t chase those targets! They are all over the place, WILL change over time (Just like what Spotify did in 2017), and they use different algos to calculate the “true” loudness.
And they are not even on by default sometimes! Such is the case with Apple and Spotify when integrated with third-party devices. Users have a choice to turn it off on the majority of the services. So that means there will be situations when your tracks will sound incredibly quiet compared to the competition - definitely not good!
Some other reasons why not to master to -14 (15,16, whatever) LUFS:
- Being overcompressed is a sound in and of itself - In many genres of music, such as Pop, Metal, EDM overcompression is a sound that’s been around for at least 20 years now. Listeners are used to it. Your super dynamic master might be “louder” than mainstream stuff - but it might also be alienating to some listeners. Not to mention, good luck getting that ripping sound of guitars in your industrial metal at -14 LUFs. Mick Gordon’s dynamic range on Doom Soundtrack is 6, just for example.
- Algos like Replaygain use the 95% loudest peak to determine how loud a track is. This is the reason why a lot of classical music sounds quieter on Spotify - because they have one or two big crescendos throughout the track, with the rest being relatively quiet. It judges the volume by those crescendos - meaning the rest of the track gets downturned more severely. So if you have music with a lot of dynamic parts you might do a more even compression/expansion to avoid having chunks of your music being too quiet.
- Some music mastered to -14 LUFS will still be quieter than mainstream overcompressed stuff. This has to do with a million factors I can’t cover here, but a lot of it is psychoacoustical effects which no algorithm can fully cover.
- Undershooting the target loudness will result in either your tracks being limited by the service or not increased in volume at all. Definitely not something that you want.
- Services like Bandcamp do not have volume normalization. Your CD/Vinyl release will have different goals in regards to target loudness. So you’ll need multiple ISRC codes, as well as multiple masters - if you release a lot of music it can quickly become confusing and overwhelming managing all of that. Extra points if you are enough of a Mad Lad to create different masters for different streaming services.
All those facts have only 1 important conclusion to you as a creator:
Master your music to the level where it sounds good. Doing heavy metal or EDM? Smash that loudness. You WILL be competing against other people doing the same. Doing classical? Yea chill out on the limiter/compression. Use the modern references and your ears to determine where you need to be - not some arbitrary goal posts.
Also, here is a question of more philosophical variety: If volume normalization makes more dynamic tracks sound louder in comparison, does it actually do its job? Or is the volume measuring we have now is flawed? Feel free to discuss on Twitter, @tangerine_beams or in the comment section below!
In the next blog post I'll go over payola - the types, is it worth spending money on and how to identify people who just want to take your money.