Question about headroom when prepping for mastering.!!

Discussion in 'Mixing and Mastering' started by DmelloMarfi, Dec 10, 2019.

  1. DmelloMarfi

    DmelloMarfi Newbie

    Aug 25, 2019
    Likes Received:
    here seems to be a lot of mixed information when I try to find answers to this question, so I figured I’d ask you all.

    I’ve read that I should be mixing my tracks with the master bus peaking at around -14dB lufs so that the mastering engineer can bring it up to about -9dB lufs to distribute to streaming services. My question is: if I export my files at 32 bit depth (essentially infinite headroom), why do my output levels matter at all? Couldn’t I theoretically export my tracks peaking at 0? Even if the mastering engineer needs to bounce the track out at 16 bit (for CDBaby digital distribution for instance), the levels would already be baked into the track right?pcpartpicker

    What do you all recommend? What’s my best course of action for mix prep/mix down where headroom is concerned?nba reddit
    Last edited: Dec 12, 2019
  3. No Avenger

    No Avenger Audiosexual

    Jul 19, 2017
    Likes Received:
    Ok, several mistakes/misunderstandings.

    There are different LUFS values:
    • i, means integrated, means 'program'. Could be a track, an album, a commercial clip.
    • s, means short term, integration time of 3 secs.
    • m, means momentary, 400ms.
    • TP, means true peak. A calculated value for analog gear.
    Most, not all, platforms replay their content at -14LUFSi, at the moment. If you master your tracks at -9LUFSi and -1TP they will be lowered by 5dB (loudness penalty).

    If you render your tracks with 32bit FP, the peak doesn't matter for the mastering engineer, that's correct.
  4. Lieglein

    Lieglein Producer

    Nov 23, 2018
    Likes Received:
    Because most self-styled mastering "engineers" have no clue what they're talking about.
    Your level of the signal is not relevant for mastering and if he says so you should not work with him because he is obviously not competent enough to turn the level down by himself.

    As No Avenger said if used 32bit it's not relevant otherwise a level between -0.3 and -0.5db is recommended because of intersample peaks.
    Last edited: Dec 10, 2019
  5. Baxter

    Baxter Audiosexual

    Jul 20, 2011
    Likes Received:
    You should leave some headroom on your mix. Around 6dB headroom is a rule of thumb (-6dBTP @ 24bit).

    For the master (if you master yourself) you need some 0.4-2dB of headroom to avoid ISP during conversion. The coversion clipping is content dependant, so you can check the peaks for yourself (or the ME will do it). Spotify even suggests -2dBTP (regardless of ISP peaks). The peaks should be True Peak, not peak.

    Loudness (LUFS) has nothing to do with peak. You mix towards whatever loudness you think is sounding good. Then the ME will master it to whatever loudness he/she thinks is good/suitable, or from your request. If you are going for loudness, you should mix for loudness. Just don't go overboard, as it will leave the MEs hands to be tied to your loudness. Find a balance and communicate with the ME.

    24bit has 144dB of dynamics and is the required standard for many MEs. It's well beyond human dynamic hearing.

    The 24bit file (144dB dynamics) will be internally processed in 32bit FP (theoretically 1680dB of dynamics). If you want Redbook Audio (44.1/16bit, CD audio which has 96dB of dynamics) the ME will apply dithering to quantize the dynamics when downsampling.
    Last edited: Dec 10, 2019
  6. Qrchack

    Qrchack Platinum Record

    Dec 15, 2015
    Likes Received:
    Absolute levels don't matter at all - during mastering you can always turn the file up or down as needed. Your mix could be peaking at -24dB and there is no problem with turning it up +24dB. The idea to leave your track at -6dB comes from the fact that you'll probably have to use less dynamic processing and leave the job to the mastering guy. If you have your track already pinned at 0dB all the time, there's not much that can be done to add dynamics back in. The best would be to skip any master dynamics and leave the mix pretty quiet.
  7. Satai

    Satai Rock Star

    Feb 23, 2013
    Likes Received:
    It's mainly because mastering engineers are used to being anal about every little thing, that is their job in the production chain after all.

    So if you export at 32bit, sure he can bring up or lower the level to whatever he wants to work with, but he's going to be anal even about that and think to himself "that's an extra processing step. Could lower the quality in bit depth ranges that even NASA equipment cannot detect. That's unacceptable in my workflow, re-render the mix at such and such level for me."

    I'm joking but only half joking. The best course of action is to render at 32 or 24 bit, average levels at about -18db RMS and peaks hitting at around -6db fullscale. It's not super important, but he'll appreciate it and feel like he's working with a fellow professional. It's a ballpark range, if there are overs leave them as is - he probably has much better limiters to control those peaks for you.
    Last edited: Dec 10, 2019
  8. korte1975

    korte1975 Rock Star

    Sep 3, 2014
    Likes Received:
    if its a 32bit float , headroom doesnt matter, you can simply reduce the gain with a gain plugin before mastering thats all
  9. SafeandSound Mastering

    SafeandSound Mastering Noisemaker

    Dec 13, 2019
    Likes Received:
    London UK
    Hi all, what matters is gain structure which determines the mix level.. there are few factors.

    1) Not clipping your stereo master output accidentally when mixing (i.e. leaving sufficient headroom)
    2) Driving analogue emulations at the "correct" level.
    3) Gain structure digitally vs analogue input stages if/when recording vocals/instruments etc.

    A mastering engineer will be happy with a 24/32 bit depth file that peaks anywhere between -12 and -1dBFS the actual headroom below 0dBFS is not critical, what is critical is that you don't clip on the loudest peaks of your track.

    A sensible peak record level can be taken into acccount as well (at 24/32 bit depth) around -14 to -12dBFS (this is NOT an absolute) it will give ample headroom for any recordings and get your mic/preamps/desk optimized for low noise and least likelyhood of distortion or overload in the analogue domain.

    I would start a mix with peaky sources (let's say drums) peaking at -14dBFS on the master output then you are unlikely to run out of hearoom when building the mix from your sources.

    This will also be ballpark ok for hitting any analogue plug in emulations which likely have a reference of +4dBu = -18dBFS i.e. harmonic generation/saturation effects like emulation of valves and input transformers. (of course you can apply gain going in and attenuate post emulation processor if you are looking for "drive")
    • Winner Winner x 1
    • Interesting Interesting x 1
    • List
  10. 5teezo

    5teezo Rock Star

    Feb 2, 2012
    Likes Received:
    Q: Headroom?
    A: Leave some!
Similar Threads - Question headroom prepping Forum Date
AudioUTOPIA Pro Tools 12.3 First Time User Questions Pro Tools Friday at 10:08 PM
Waves VR question Software Friday at 7:10 PM
Question about content of Studio One 4.6 from sister site Studio One Friday at 6:05 PM
VERY SERIOUS HEADPHONE QUESTION!! Mixing and Mastering Friday at 4:46 AM
If you could to ask your favourite producer one question about their music, what would it be? Education Wednesday at 11:34 AM