Lets talk about mixing levels (again)

Some years ago we had lots of discussions about proper mixing levels in the digital domain – with mixed (sic!) results, IIRC. Meanwhile, more and more influencers are claiming that targeting -18dBFS with a VU meter readout is the “digital audio sweet spot” and the way forward in terms of plugin gain staging. In practise that would imply mixing digital peak levels at around 0dBFS again but maybe I’ve missed something during my absence in recent years. So, to what mixing levels are you up to in your DAW today?


  1. Hi, I’m using the K-system (intergrated in my DAW which is Mixbus 32C from Harrisson Consoles, based on Ardour). I mix to K-14 for my music (modern pop-electronica).
    Explanation of the K-system here: https://www.aes.org/technical/documentDownloads.cfm?docID=65

  2. Its been a while since I mixed Bootsy (I have been spending most of my time Mastering for quite a while now). I used to be around -19 / -18db when i did rock.
    Hope that helped mate. Cheers..

    • My Soundcraft Digital mixer puts out +4 dBu at -18dB FS, This is indeed standard behaviour. But not very relevant when we quit the digital domain only at the consumer end

  3. heavymetalmixer says:

    Given that I always set my interface level to the same value, I don’t need to worry too much about clipping, but that’s it, most of the time I don’t worry about levels, except maybe if I feel any of the plugins could clip internally.

  4. Lately i’ve taken to starting with everything dropped by 26dB in the DAW and mixing there (not paying attention to exact VU readings since i use different types of analyzers in each track type template). I can’t remember exactly how i came up with that number and can’t find the notes, but it was supposedly based on some sort of logical reasoning. Sometimes that’s too low for certain plugins (e.g. too low for a threshold to grab, or missing the sweet spot range) in which case i’ll compensate the in-out levels as appropriate. Since adopting that system i feel i’ve been getting better results overall, but have to be careful about letting certain elements stay too dynamic which can then cause issues later during mastering (i.e. a peaky drum beat causing an underlying guitar to duck every time a limiter hits).

  5. RMS around -18 dBFS, peaks go to -12. It sounds best and is most pleasing to work with. Interface is at max (Line 6, I suppose it is around 70dB).Unfortunately I have to pump up the volume during mastering but I never go above -9 dBFS RMS. Everything above that sound horrible to my ears.

    • Tryggvasson says:

      What Viper said. Good numbers. Essential for working with analog emulations – many don’t know, but the in-volume can make or break your signal through those. It can make it sound like Santa Claus just came to town, if you set it right, or add mud and muffle your transients, if you overdo it, or not do anything much at all, if your levels are too low. And many times a few fractions of a db can really change your tone. That’s why you have input on compressors, and not only threshold. That’s what you set up first, before you touch the compression – and that’s your overall timbre/transient/saturation setting. It’s only after that you start compressing, with the tone made.

  6. I usually normalize my source media items to around -21/-23 dBfs RMS and do not care so much afterwards, knowing that my DAW fader is post fx anyway.

    In terms of DSP I don’t see the point of a digital sweet spot (apart from the lower the volume, the lower the resolution). Where I cound agree is in these more modern plugins who utilize more meta controls and tune some hidden thresholds to usual target loudness values (like “compression” in “%” rather than “threshold” in dBfs). But this applies more to the dynamic/waveshaping realm, since filters or time based DSP should be independent from loudness.

    Don’t really know what these component level vintage modeling girls/guys would say about this topic.

  7. Tryggvasson says:

    Hi, I know it’s off topic, but I’ve seen you’ve closed the Beta topic, so I’m writing here, because I think it’s useful and kind of important.

    A bug in NastyDLA:

    that has been there since forever, starting with the 32 bit version and it persists in the 64 bit one, too.

    The “Wet Only” toggle state doesn’t get stored with the project. Every time you reopen a project saved with it ON, it loads it in the OFF state, so you have to remember to switch it on. It messed up a few exports, like that.

    Thought it was worth mentioning.

    Looking forward to the big news.

    Thanks for all you’re doing!

  8. The -18dBFS does not really affect a plug-ins behavior (maybe some Acustica Audio stuff, but some say it doesn’t even matter for them). Most plug-ins and DAWs run on 64- or 32-bit FP so you can literally have the tracks at +100dB as long as you have another plug-in/volume utility before, or on, the master output that reduces the level again so you don’t clip the master output.

    I think it’s mostly a habit that has stuck around since the analog days. It’s not bad in any way, and it can help some beginners to get a cleaner mix from the get-go, but it’s important for people to understand that it’s not something that is required to do in the digital domain. It can also be something that is good to learn early on if you plan on incorporating a lot of analog eqiupment later on.

    • You might have missed the point here: If a plugin implements some sort of non linearity, e.g. some saturation behaviour, then you can’t push it as you want. The discussion is about where that sweet spot has to be expected, actually.

      • Absolutely. If you work with analog emulations, or saturators with a sweet spot (or without an input gain), the input volume makes a HUGE difference. Combined across the tracks, I’d say it’s one of the critical decisions in can make in your mix, in terms of punch, clarity, brilliance, glue, weight, dimension, taming the transients, if you work with real instruments, etc. Which is not surprising, seeing that it’s the same thing mix engineers that have worked on the actual stuff will tell you. Yeah, if you don’t use any of those, it probably won’t matter. But why wouldn’t you use those, when the best ones can achieve with a knob what you can’t achieve on a different one turning five?

  9. Rodrigo says:

    I use IVGI2 by klanhelm for this task. I really like the idea of ​​having a plugin to take all my signals to nominal level before I start mixing.
    Anyway, most of the modern manuals no longer clarify anything about it.

    • To understand your approach fully: you use IVGI for gain staging, for saturation or for both? I really like IVGI, though rumour has it that it has no oversampling (unlike it’s bigger brother) and can cause hearable aliasing. Never tested this myself and I had no issues with it’s sound quality tbh.

  10. uncajesse says:

    Not a single mention of R128 ( https://tech.ebu.ch/loudness/ ) or 1770 ( https://www.itu.int/rec/R-REC-BS.1770-4-201510-I/en ) for loudness normalization? That’s what we’re using in broadcast land, and we’d like to see more of that in music-related software for mix engineers. It’s used heavily in audio mastering although.

  11. Adriano Michael Petrosillo says:

    Personally, I am not very disciplined in this. I track peaking at -12/-6dBFS (I aim at -12dBFS but then you get the occasional hit peaking higher than that, but I try to never go over -6dBFS). I’m still using an Echo Audiofire 12 though, which has its 0VU at -12dBFS (it’s an older interface after all), what I do is set the external preamps at shy of clipping so the level is actually set by the preamps. Since I like to multi-mic my sources (I have lots of channels after all) once I mix the mics I often get hotter levels. When I use analog emulation plugins (eg. Decapitator) I find that I have to keep the input level low otherwise it crushes everything, but it’s manageable. With better gain staging (which I plan to implement in my next mixes) it will probably be easier to find the sweet spot. Anyway I presume that you should set level according to converters, that is, if converters are rated at 0VU=-18dBFS, you should expect them to be cleanest there. If you track hotter some converter distortion may arise (just like analog gear has headroom but it doesn’t sound “best” when you stay always within the headroom, unless you are pushing your gear on purpose), even though it would be beneficial to actually analyse how the ADCs behave. On the other hand, we master music aiming at much higher than -18dBFS, which does make you wonder whether having such high levels is stressing the DACs.

What do you think?

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: