so, what is your reference mixing level?

And if you choose ‘other’ please leave a short comment about it!


  1. -18 dBFS that is. How can you have 0dBFS as a reference, who mix without headroom..? Want-to-be-pop-master-engineers is what I can think of, knew a guy like that once. No headroom at all.

  2. David Köhler says:

    i mix at about -6-10dB to leave some headroom…an don’t need to limit,to keep the sound unchanged…

  3. 0x45455844 says:

    Just what is comfortable which depends on mood
    And then I forget to check and keep never knowing
    Maybe here’s good place for some advice?

  4. as long as it ain’t clipping…

  5. You cannot use numbers as a reference for mixing, as long as its below the clipping threshold and its sounding warm and sweet coming out of the monitors its all good.If one could really use numbers as a reference we would all be mixing with headphones and thats a no no 🙂

  6. Trust your ears 🙂

  7. Im mix in between -18/-12 dBfs, give my self that 2-8 bit of head room

  8. in general,it’s between -3 to -6 db

  9. A great many even so-called “professional” converters can sound very bad as you approach the top 6 dB. of their range due to poorly implemented analog stages and minimal power supplies. This can lead people to making choices they wouldn’t otherwise make that optimize their mix to one monitor chain’s peculiarities while making it sound worse most everywhere else. On the other hand I’ve gotten 24 bit ITB mixes to master that had a maximum peak of -20 and sounded spectacularly good, far better than average mixes I get that were done on SSLs and Neves.

    The lesson to me has been to always aim at erring on the low side rather than the high side in the digital domain. This also helps a great deal in applying any kind of non-linear signal processing such as an emulation of analog equipment.

  10. Mastering at 0LU = 83dB SPL (a-weighted, slow). EBU R128 calibrated reference loudness, yeah boyeee. Peak control for loudness after the fact (the private distortion-masking clipper i use is that good) which is obviously where I have to toss references out the window. Voted -24dB since that’s only 1dB away from 0LU, and it doesn’t mention what flavor of dB it is. (RMS is a horrible indicator of subjective loudness, in case you hadn’t noticed yet)

  11. I use -3 to -6 depending on how loud the song should be.

  12. know your room, know your DA and know your ears…then it depends on whatever you want to do. after all, let’s be practical: there’s any need to waste bits once your beloved 24/96 has been shredded into 16 bits? do whatever you want as long as it makes you confortale, when ITBing, and then adjust to what you need. vinyl? mp3? cd? broadcast? different uses, different needs, different settings..(for not to talk about different genres.. i can’t use the same detection systems for electro and blues…) so I would say an internal calibrated meter with preferred settings as well as continuous control and a CLEAR peak led can be really useful. but there’s a lot of them out there. (or is it just for the metering section of your next plugins?)

  13. -18dBFS = 0VU for me. Most converters are calibrated to that, or -15, or -20.

    Most converters sound bad in the upper few dB as the analog portion of the converter saturates (or worse, clips) and for icing on the cake the converter loses linearity. Additionally, if you record too loud, your gain staging is entirely messed up (even if you aren’t using any outboard gear). I mix about 100 records a year, some of which have done very very well. The people who track really hot – those records don’t sound as good as the people who track at more reasonable professional levels (not hitting over -12dBFS, max). Improper recording levels are the BIGGEST problem I face when getting records to mix from ameteurs and home recordists. Analog tape decks back in the day were calibrated for a reason, and so are digital converters. Ignore it at your own peril….

  14. Seems to be that -18dB is the consensus. Does this come from the recent mix desk plug-in trend?

    • No. It comes from most converters being calibrated that way such that -18dBFS = 0VU. That way you have the same headroom working with digital or analog and can freely go back and forth. Recording at -3 is like hitting your analog console at +15!!!!

      Another way to think of it is miles/hour vs. kilometers/hour. You just convert. If you drive 100km/hr you can just do drive 100m/hr on the same road – you’d crash and burn. Same thing with audio.

    • For the record, -20dB RMS is the industry standard. It doesn’t matter what the vote here is, as for as what an industry consensus is. But that’s starting to change over to EBU’s R128 standard, which comparatively has even more headroom, not to mention is much more accurate to actual subjective loudness. 🙂 The payoff is that you’re mixing at the same SPL levels. Not the same signal levels. And that’s much more important.

      • According EBU, -18dB RMS is standard …

        • Before R128, you’d be right per EBU. I should have qualified it as… AES and film industry standards. The film & broadcast industry are also the first to be jumping into R128 so that won’t last for long. It would be cool to see another poll that actually has FS or RMS put on it. And of course an R128, and “other” option. If I was confused into assuming that it meant RMS, when most people voted FS, then… the results are probably not trustworthy.

      • Actually -20 VU is the same as -18 RMS on pink noise due to the differences in the meter dynamics. This causes a lot of confusion.

  15. Somewhere north of -18db most of the time (approx -12 I guess).

  16. I’m using the M-Audio FastTrack Pro. According to the manual, the green LED lights up at -24dBFS and the clip LED lights up at -1dBFS.

    I chose -24dB

  17. @ALL
    Do you talk about a OLU-calibration for 83dB SPL (c-weight,slow) = -18dB FS RMS for mixing and mastering?
    I want to recalibrate my optimized room soon.

    Thx and greetz

    • I’m using 0LU = 83dB right now, with hand-calibrated linear-phase low-ripple parametric EQs. And loving it. But my Apogee gear’s distortion rating (0.0001% or some such) is rated at +24dB on a +4dBu level line input. So… it’s unusually clean.

      Still though, the EBU P-LOUD group (which I’m in) that came up with R128, put a lot of thought into how it would work with the ADC/DAC available out there, even the bottom-shelf stuff like M-Audio. It’s designed so that you’ll never have to worry about peaks, and it’s recommended that any meters don’t even have peak meters. Only clipping indicators (-1dB FS TruePeak), which will hopefully not encourage people to limit the transients on their recordings/productions, unlike peak metering.

      As far as actual record levels though, I’ll agree that no one standard will really be optimal, no matter how great your gear is. How about Schoeps mics into Millenium pre-amps into Burl interfaces, it *still* makes a difference. Been there, done that. Eventually for a particular input source and source content, you’ll find a point where you balance the amount of noise you’re getting, and the amount of gain (power) you’re applying from the rails to your audio path. It’s important to not supply too little power from the rails, or in many pre-amps (especially of lesser robustness) you can lose definition from that too, especially in the bass.

      Suffice to say though, that’s much less of a problem these days than providing too much gain. Cheers. 🙂

  18. It depends on what you are mixing, the size of the room and, interestingly, the size of the screen if you are actually mixing a movie where the listener will have no volume control. For audio production where the listener has a volume control, such as pop music, it’s critical that the mix sound good at a wide variety of listening levels because of how our hearing is affected by different levels.

    As you advance the volume, a good mix will really hold together while bad mixes will only sound convincing at one level in one room on one speaker. You can learn to listen for this as two frequency bands with one above 500 Hz. being compressed and everything below 500 being uncompressed. I find that’s a more useful way to think about it than the equal loudness curves from studies during the 1930s.

    • “As you advance the volume, a good mix will really hold together while bad mixes will only sound convincing at one level in one room on one speaker. You can learn to listen for this as two frequency bands with one above 500 Hz. being compressed and everything below 500 being uncompressed. I find that’s a more useful way to think about it than the equal loudness curves from studies during the 1930s.”
      Bob, I really enjoy your contributions here – can you elaborate just a little bit on that statement? What does it mean exactly to split it that way at 500Hz? Why and what for?

  19. It has to do with the physiology of how our ears work. Everything below around 500 (it’s a slightly different frequency for different people) will get louder much faster. We amplify and compress everything above 500. “Flat” response, such as it is, lies around 85dB. SPL but it’s obviously a moving target. You can hear this very clearly with full range speakers or headphones by slowly advancing a volume control from zero.

  20. I’ve found this two band conceptual model based on inner ear physiology more useful as a means of understanding. The problem with the equal loudness curves is that no two people have the exact same physiology so no two people can have the same curve. As a result people get caught up in which curve is more “accurate” or “up to date” and miss the fact that the upper band is dynamically adapting to the sound level while the lower band is not.

    • Fancy seeing you here Bob. Respect. =) Good insight for new comers to investigate. How about these plugins eh?

      p.s. for those that don’t know, Bob was in the trenches at Motown during their hay days. he’s a living legend, so listen to what he says. i’ve never heard a word of bad advice come out of his mouth, although he might say that’s not true, hehe.

  21. Mixes are at about -20 db RMS here.


  22. We write, mix and master all at once. So we are always hearing the final product. I know this is unconventional but we found that it works great for us. Speaking of, Bootsy, we thanked you in a our new CD liner notes, if you’d like album as a thank you, just pm me for link. Style is industrial/ebm.

  23. -23dB = 0LU

  24. FWIW LucasFilm uses -22 rather than -20 because of their consoles’ lack of headroom! You really need to get out your test gear and find the sweet spot of analog gear. A couple decades ago the guys in the shop took care of a lot of this so we mixers and mastering engineers never needed to think about it.

What do you think?

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

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

Google photo

You are commenting using your Google 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: