Introduction
I love listening to dynamic music. There's just no replacement for those clear sounds and clean, punchy drums that make the mix sound powerful. With many of the songs I critique, loudness is one of the biggest issues, and when the producer or engineer just backs off of the limiter by 3-6 dB, everything sounds cleaner and more professional.
In a previous post, we explored the context for measuring loudness in decibels, how to read LUFS meters, and how to arrive at the perfect loudness for your master.
However, there are a couple of really neat tools available now that didn't exist when I wrote that article, and finding the sweet spot for your music's loudness is easier than ever. Let's take a look at what we need to know:
We're Still Measuring in LUFS
Just a refresher, Loudness Units relative to Full Scale (LUFS) is still our preferred scale. We'll be using LUFS to measure average loudness. LUFS-Integrated is a scale that gives one value for the loudness of your entire program, no matter if it's a two minute song or a two hour movie. And if this still seems confusing, be sure to check out the more in-depth explanation in my previous article on loudness.
We're Still Targeting the Same Loudness
Ian Sheperd's loudness observations per streaming platform are still in effect. At the time of this writing (August, 2018), this is how each of the major streaming platforms reacts to loudness:
- YouTube's algorithms target approximately -13 LUFS-Integrated as a loudness target, and YouTube does not turn up songs that are quiet.
- Spotify's algorithms target approximately -14 LUFS-Integrated, and Spotify does turn up songs that are too quiet and can use limiting to do so.
- Tidal's algorithms target approximately -14 LUFS-Integrated, and Tidal does not turn up songs that are quiet.
- Pandora's algorithms target approximately -14 LUFS-Integrated, and Pandora does turn up quiet songs, but does not use limiting to do so.
- Apple's iTunes algorithms target approximately -16 LUFS-Integrated, and iTunes does not turn up songs that are quiet.
All credit for these numbers goes to Ian Shepherd and his research.
You can read more of his writing on his website: productionadvice.co.uk
An Exciting New Tool: DROffline MkII
MAAT Digital may be a young company, but it's already making a big splash in the mastering community. And one of the neatest tools it makes is called DROffline MkII, short for the second version of Dynamic Range Offline (abbreviated as DRO2 for the rest of the article).
In short, it's a stand-alone program that crunches the numbers on a wav file to tell you exactly how loud it is according to a number of different metrics. I find this approach really useful because it generally takes less time to export a song and scan it with DRO2 than it does to play out the song in my DAW to see its integrated LUFS measurement using a metering plugin. And, DRO2 is a more precise measurement tool, and it also shows other relevant metrics at the same time. I prefer to use it with the Modern Mastering preset, which cuts out a lot of data that's relevant for film and broadcast work, but not for music guys like me.
How I use it:
When I'm finishing up an audio project, I drop the exported wav file into DRO2 and let it quickly scan the file. DRO2 acts as a neat double-checker as it displays the sample rate of the file, the bit depth, and also the bits used. But more importantly, it tells me everything I need to know about my loudness: primarily, how loud my loudest inter-sample peak is, how loud the entire program is in LUFS-integrated, and as a bonus, how much dynamic range my audio has.
I always aim to have my inter-sample peaks hit -1.0 dBFS, no higher. If this reading shows anything else, I adjust the ceiling in my limiter to fix it. Streaming services don't play nicely with songs that peak over -1 dBFS.
I aim to have my LUFS-integrated read -14.0, since I prefer dynamic music. If this reading shows anything else, I adjust the threshold in my limiter.
And the dynamic range figure just gives me a fuzzy feeling when I look at it, knowing that I'm not crushing my music with excessive limiting or compression. It's one of the small pleasures in life.
Why I recommend DRO2:
I don't know of a quicker, more accurate way to get to the heart of my music's loudness, which allows me to quickly make specific changes in my limiter to set things right. And at $49, it's not only convenient, but affordable too.
Another Exciting New Tool: LoudnessPenalty.com
LoudnessPenalty.com is Ian Shepherd's latest project, and it works really simply. Drop your mp3 or wav into the website, and your browser crunches the numbers for how loud your music will seem to all of the major streaming services: YouTube, Spotify, Tidal, Pandora, and iTunes. DRO2 tells you the truth of your loudness, whereas LoudnessPenalty tells you the reality of what will happen to your volume. And, best of all, this website is free to use.
We already covered above how each streaming service targets loudness. But what we don't have a handle on is how the math behind each service reacts under different circumstances. My hunch is that Tidal and YouTube have the most accurate measurements in response to true volume. The trouble is that both services can turn down loud songs, but they don't turn up quiet songs: so you may not know from reading those numbers if your volume is perfect or way too quiet. Spotify and Pandora, on the other hand, seem to respond a bit more harshly to dynamic music than the others, but both are capable of turning up quiet songs, meaning both are capable of giving you some perspective on how quiet a quiet song really is. Lastly, iTunes seems a bit sluggish. It will be the first service to turn down your volume since it is geared the most conservatively, but how much it turns the volume down doesn't seem to track consistently with a song's loudness.
How I use it:
I pop my song into the website, wait for it to analyze the loudness, and look for problems. Although if I'm using DRO2 properly, there won't be any.
For me, problem-free means that YouTube and Tidal aren't touching my volume, Spotify and Pandora are are turning me down just a decibel or two, and the iTunes number seems reasonable (anything between -4 and 0). As you can see, LoudnessPenality is easier to understand, but harder to use.
Why I recommend LoudnessPenalty.com:
It's great to get a picture of how each streaming service will react to the loudness of your song, it's visually easy to read, and it's free!
Choosing Your Target loudness
So you now have tools you can use. But what loudness should you actually aim for? This may seem like a cop-out, but it's up to you.
I like dynamic music. With this in mind, I aim to have my integrated LUFS measurement in DRO2 read as close to -14.00 as it can. If it's not reading -14, I tweak the limiter's threshold, export, and measure again.
For a slightly more competitive master, I'd aim for DRO2 to read as -13 LUFS-Integrated. Look for YouTube to leave the volume as it is, and for Tidal to turn the music down by 1 dB.
If you want a more competitive loudness still, aim for DRO2 to read as -12 LUFS-Integrated. Look for YouTube to turn the volume down by 1 dB, and for Tidal to turn it down by 2 dB.
I wouldn't push your loudness any more than this. You can do it, but it really won't sound pretty. I've heard arguments for making EDM painfully, punishingly loud in order to be competitive, and that the loss of quality isn't such an issue. Maybe it's even expected as part of the sound. I personally disagree, though I expect most EDM masters to push a lot louder than this. But for everyone else? Save your music!
The rule of thumb:
Consider -12 LUFS-Integrated as "Good", -13 as "Better", and -14 as "Best". And let's throw in a -16 as "Ultimate" if you're targeting an audiophile audience. Just for fun.
Loudness Consistency Across Multiple Tracks
So we found the perfect loudness measured in LUFS-integrated according to your preference. But if we apply that to every song on your album, the perceived loudness of each song will vary. Unfortunately, there's no true way to measure loudness as accurately as the ear does.
The best approach is to import all unmastered songs as 2-track wav files into your DAW on separate tracks. Pick just one track to measure. It should be loud and busy and sound typical of the whole album. (In other words, don't pick the slowest, quietest song.) Use the process above to bring that track to your target loudness, be it -14 LUFS-integrated, -13, or -12. You now have one song at the appropriate loudness for your entire album. For all of the other tracks, one by one, match the volume of each by ear to the first song. Remember: set your limiter's ceiling to -1.0 dBFS, and adjust your limiter's threshold to bring the volume up or down to match the loudness of the first song.
TL;DR - The Quick Guide to Achieving Perfect Loudness
To find the perfect loudness for each song on your album:
- Start with a loud, full song that you feel represents the peak of your album. Use this song for every step below.
- Decide if your loudness target is ultra-conservative (-16 LUFS), conservative (-14 LUFS), moderate (-13 LUFS), or aggressive (-12 LUFS). I personally target -14.
- Use a quality limiter, like Ozone Maximizer, Pro-L2, Invisible Limiter, etc. Set your limiter's ceiling to -1.0 dBFS, and remember to set your limiter to True Peak mode, for it to detect inter-sample peaks. Adjust your limiter's threshold until it is just responding a little to the music, but not too much.
- Use DRO2 or a similar tool to discover the integrated LUFS value for your song. If your loudness isn't at your target, adjust the threshold of your limiter and measure again. Repeat this process until you nail your target loudness for this one song.
- Double check that things are looking right with LoudnessPenalty.com, according to your loudness goals.
- Use your ear to match the loudness of all the other songs to the loudness of your first song. To make another song louder or quieter, adjust the threshold of the limiter on that song's track.
And you're done!
Wrapping Up
Once again, I feel like I used a million words to describe a somewhat simple process. It's just that this process seems to be incredibly murky even after lots of research, and it took me a long time to get my head around when I was learning it. Yet it's so important!
I hope this article is useful to you, and will help you target your preferred loudness with confidence and courage.
If you have any questions or additions, please mention them in the comments below. I love hearing from you guys.