Hey guys. What decibel level do you guys usually mix to? I've always tried to keep my dialogue between -6 and -12, and try to record it the same way. Are you supposed to turn it up louder though when mixing? Reason I ask is because I did a commercial for a client and it hasn't aired yet, but it turned out being a lot quieter than other stuff on TV.

Do you guys usually try to get your levels closer to 0 db?

Scott

You need to get hold of the delivery specs for the broadcaster(s) that are airing the spot (or from the ad agency that is placing the airtime buy). Every station and network has engineering standards for HD, digital, and SD materials and you must mix according to those requirements. You also need to be mixing in a calibrated environment where -20dB or -12dB on the meters corresponds to an exactly determined and measured sound pressure level coming from the monitors at the mix position. Luckily that's easy to set up and only requires an inexpensive $50 sound pressure meter from Radio Shack.

Wnen you are RECORDING, keeping speech peaks around -6 to -12dBFS is good. Maybe even a bit agressive for unpredictabe live events.

MIXING is a more relative thing, getting the proportions of the dialog, music, SFX, etc. all at their proper proportions. I generally mix with peaks around -10.

RELEASING is yet a different matter. If this is a local spot then it probabaly needs to be compressed and processed to have the same *loudness* as the competing commercials. If the other commercials are THAT loud, they may even be normalized all the way up to 0dBFS.

Reason I ask is because I did a commercial for a client and it hasn't aired yet, but it turned out being a lot quieter than other stuff on TV.


Yeah, commercials are taking a cue from the audio world and are becoming more obnoxious. "Smashed" is the term to use for heavily maximized audio with zero dynamics. However, it seems that commercials mixed in 5.1 are preserving more of the dynamics, which is a good thing IMO.

Essentially, you have two choices Scott - deliver a final mix to a mastering eng and let that person find your mix's volume sweetspot or buy some plugs from TC Electronic, Waves (L1 has been a goto plug for almost 15 years now), iZotope, and/or UAD and try it yourself. The outboard gear that is typically used is kinda pricey if you're not already deep into the pro audio world.

Steve - I will definitely get the delivery specs from the broadcaster... something I should have thought of already.

Richard - I follow you up until you say "releasing". If you mix to -10 at your peak, what are you doing to get it louder? Do you do this in mastering? I do my sound mixing in Final Cut Express, and don't really have a master fader.

Aaron - I'm definitely not deep into the Pro Audio world, but I've had an Mbox, ProTools LE, and a decent set of studio monitors for about 5 years now, (was into home recording before videography). I'd rather not send this to a mastering engineer, so what are my options with what I have? Keep the audio in Final Cut, or try doing something with it in ProTools?

Here is the commercial, by the way, in case you guys would like to check it out. I'd love some feedback. The sound was recorded with the Rode VideoMic, but I've since upgraded to an NTG2 and Zoom H4n.

Sunpoint Tanning Commercial on Vimeo

"If you mix to -10 at your peak, what are you doing to get it louder?"

Look up the function "Normalize" in your editing workflow. I would wager that both Final Cut and ProTools feature a way to normalize the audio.

Normalize means to adjust the volume of the entire clip so that the maximum peak is at the specified level. If the highest peak in the final mix is at -10dB, then if you normalize it to 0dB, you would boost the entire clip by 10dB.

From your description, I'd bet that the other commercials are normalized all the way up to 0dBFS (the absolute maximum in digital audio.) And probably significantly compressed also.

Richard, thanks man! Sorry for my ignorance on the subject. You guys have helped a lot.

Sorry am I being stupid here, but do any of you guys know what that little switch on the front of the mixer that says line-up tone is or what it does? Calibrate this correctly with what ever you are recording onto and you should never have a problem with incorrect levels. Buying cheep products with Vu meters doesn't help either as these are a complete wast of space. Needled PPM meters using either the BBC or Nordic scale are a much better system, they have clarity, visually are far simpler and also tell you exactly what you need to know in terms of sound peaks.
Listening back to something on tv after it has gone through an edit process is no indication of the level at which it was recorded (unless it was distorted in the first place) it is merely an indication of how much compression, or lack of, that had been added in the final Tx mix.

Sorry am I being stupid here, but do any of you guys know what that little switch on the front of the mixer that says line-up tone is or what it does? Calibrate this correctly with what ever you are recording onto and you should never have a problem with incorrect levels.

Calibrating to lineup tone doesn't guarantee recording at proper levels, though it helps. But even if I've adjusted the camera's recording levels to align 0VU tone from the mixer to -20dBFS, or -18dBFS, or -12dBFS on the camera's meters, depending on what standards I'm following, it doesn't mean I'll have the right levels if I don't ride gain on the mixer's mic channel faders during shooting to keep the average dialog levels hovering between 0VU and +4 or so or the mixer. Which alignment level you use depends on the material being recorded - with normal speech, an average recording level of 0VU results in a peak level of about +8 DB. Aligning to -20 means you'll have 20dB of headroom before you risk clipping, -18 gives 18dB of headroom, and -12 gives about 12dB before the camera levels hit 0dBFS and you clip (or preferably trip limiters a couple of dB lower).

The considerations when setting levels are different for original source recordings and for the final mix coming out of postproduction. The line-up tone you're describing is for aligning the mixer's meters to the recorder, be it camera or audio recorder in order to be able to look at the mixer';s meter and tell what's going on in the recording device. But that original recording level is not ncessesarily going to be the level at which the track will appear in the final mix that is going to be delivered to the broadcast station or other destination way down the line after it goes through all the post production processes of editing, mixing of the various stems, and mastering. The same master track as recorded in camera at, say, 0VU=-18dBFS (EBU standard) could end up going to DVD or theatrical release with peaks kissing -1 dBFS, HD broadcast with a peaks limited t0-3dBFS, or SD broadcast with peaks at -10dBFS. If the final production was a drama for broadcast average dialog levels might be adjusted to -22dBFS while if it was a used car commercial, it might be compressed and boosted to much higher actual and subjective levels. Take a look at some the batch of online articles on the whole issue of "dialnorm."

FYI, while your criticism of LED meters on consumer or prosumer level equipment is somewhat deserved (too few steps so inadequate resolution, unknown ballistics and calibration), the LED meters on the Sound Devices mixers and similar pro level equipment are both more accurate and faster responding than the old fashioned needle meters. I think SD's combination view with simultaneous average and peak displays, coupled with peak hold, in particular gives you far better information than the needle meters ever could. And for post and broadcast, the current bar is set by Dorrough loudness meters and guess what, they use LED displays.

I'd rather not send this to a mastering engineer, so what are my options with what I have? Keep the audio in Final Cut, or try doing something with it in ProTools?

Well, if I had my post studio fully up and running, I'd say send it me :), but I'm not quite ready. In that case, I'd highly recommend downloading a trial ver of Sound Forge from the old SonicFoundry crew (now Sony). I've been using this audio editor since ~1996 and you're able to do some pretty good stuff ITB after you figure out the way it works.

So mix the audio on whichever platform you're most comfortable with and then do the final tweaking in Sound Forge. I'm pretty sure the iZotope plugs have been included for awhile too.

And don't rely on normalizing to smash your audio. When I track a live show, I'll typically let my peaks for most tracks top out ~ -1db because I've put in the time learning the program material (either musical instrumentation, speech, vocals, etc.) before pressing the little red button so I'm comfortable engaging some lite compression/limiting on my mic strips. If I would merely normalize, I'd gain less than a db until I peaked at 0db.

So don't be afraid to use compression on tracks that require it during mixdown. It will help those tracks "sit" well in the mix without drawing undue attention to transient peaks. But save the smashing for mastering the final mix.

I'll edit this to add (after watching your example) that that would be a difficult shoot for audio because of the varying levels of the main (only) vocal track. Compression would help, but it would also bring up the noise floor level at the same time. Ideally, if a consistent vocal level was desired, you'd have to have the boom op follow the lead with the mic both in and out of the shot. That is tricky because that person would have to stand in the other aisle and do it blind! Wow, tough conditions! But, the result would crystal clear vocal during approach and during exit. Then you could smash the mix and not be too worried about all of the extraneous "noise" - fluor lights/ballasts, "room sound", etc. No wonder why ADR/foley is so popular...