• Wondering which camera, gear, computer, or software to buy? Ask in our Gear Guide.

Audio levels

Mixing audio. What are the standard max audio levels for dialogue, and for bigger stuff, like music?

Of course I tried googling, but mostly all I found was a bunch of technical jargon that's way over my head. All I understand is that little audio meter that ranges from -infinity, through -30, to 0.

What I've gathered, so far, is that normal peaking, depending on who you're getting advice from, should never top -6 to -3. Really? I should never creep towards 0?

Actually, I kind of already knew that, or at least sort of. My primary question is this --

What about a scene that is just dialogue? How loud should it be? How quiet is too quiet?

It wouldn't really make sense for me to have a dialogue-only scene peaking at -3, would it? Cuz that's the peak I save for the really loud scenes, when there are sound FX, yelling, and bombastic music (oh yes, my soundtrack is bombastic). Intuition tells me that, obviously, the dialogue-only scene should be more quiet.

I just can't figure out how quiet.

1. If it's a "regular" conversation, what should it peak at?
2. How low can I go, while still being heard? What if somebody's whispering?

Thanks, in advance, for the help!
 
I'll bet you will get sound advice ( pun intended ) from Alcove.. I have been endeavoring to keep most "normal dialog" at -12db, whispers lower and only have the loudest sounds peak at -3. I would love to know the best settings myself other than using my ears. I figure the best thing is to aim for numbers and go w/ what your ears say when most levels are close, adjusting where required.
 
Dialog riding around -6, peaking around -3, over zero will clip, so we avoid it like the plague! Run a compressor to keep it within that range as well.

I always mix everything individually to fit these standards, then mix them into the whole thing using the dialog as the primary level that everything else revolves around. I try not to change those levels if possible. Everything gets separate compressor/limiters and reverb/echo.
 
Dialog riding around -6, peaking around -3, over zero will clip, so we avoid it like the plague! Run a compressor to keep it within that range as well.

I always mix everything individually to fit these standards, then mix them into the whole thing using the dialog as the primary level that everything else revolves around. I try not to change those levels if possible. Everything gets separate compressor/limiters and reverb/echo.

Thanks. What's your process? Do you have a compressor plugin for your editing software? Or, are you exporting the raw audio, then compressing externally, and re-importing? Do you have a recommended compressor?

Also, I've never really understood the difference between compressor and limiter. How do you know when to use one or the other?
 
A Big Can Of Worms

can-of-worms.jpg

"There are no guidelines in terms of average loudness, peak or any other level measurement. You achieve proper levels by properly calibrating your listening environment, so that it resembles the environment of the theater...

Then mix by ear... If it sounds good, then it is good." - Danijel Milosevic

This is how it is for theatrical release films. It's a guideline that I've followed almost since I began doing audio post. Different rules apply to DVD, TV, Radio, Games, etc. There are no rules - as of yet - for the internet. For DVD you are basically adjusting the mix to compensate for the X-curve and the smaller listening environment. For TV each network/broadcaster there are very specific delivery specs which can vary greatly from one network to the other. Radio works the same way; prerecorded material must be delivered according to network/station specs (music and commercials are a whole different can of worms). Different gaming companies have their own delivery specs, but most of them have to work within multiple gaming environments such as Wii, X-Box, GameCube, Play Station, Nintendo, etc. which are becoming more standardized.

If you really want to dive into the nitty gritty you can start here:

http://www2.grammy.com/PDFs/Recording_Academy/Producers_And_Engineers/SurroundRecommendations.pdf

http://www.hsmpte.org/2010 Meeting Reports/SMPTE Loudness Meeting.pdf

http://www.dolby.com/DocLibTechLanding.aspx?taxid=216

http://www.aes.org/technical/documents/AESTD1001.pdf

http://www.sdds.com/pdfs/SDDSPrintMaster.PDF

http://www.hometheaterhifi.com/volume_7_2/feature-article-dialog-normalization-6-2000.html
 
Sweet. My issue with that first quote is that I'm mixing on cheap monitors (KRK Rokit 5's), so how am I supposed to know what it'll be like in the theater?

Anyway, thanks for those links; I'll check 'em out.
 
If it's a local television commercial then take all the advice here and add about 12db. After all, nobody will shop at your mattress outlet unless your commercial is AT LEAST twice as loud as the program it's scheduled for.
 
If it's a local television commercial then take all the advice here and add about 12db. After all, nobody will shop at your mattress outlet unless your commercial is AT LEAST twice as loud as the program it's scheduled for.

Fortunately this will no longer be true in the near future. The CALM Act has been passed by Congress. Just think, no more diving for the remote to turn down commercials that rip the top of your skull off!

Official Summary

H.R.6209 - Commercial Advertisement Loudness Mitigation Act - Directs the Federal Communications Commission (FCC) to prescribe a regulation prohibiting advertisements accompanying video programming from:

(1) being excessively noisy or strident;
(2) having modulation levels substantially higher than the accompanying program; and
(3) having an average maximum loudness substantially higher than that of the accompanying program.
 
Sweet. My issue with that first quote is that I'm mixing on cheap monitors (KRK Rokit 5's)

And here's the important part of the quote for you...

"There are no guidelines in terms of average loudness, peak or any other level measurement... Mix by ear... If it sounds good, then it is good."

The hard part is "translate-ability" or how the mix translates to other listening/watching venues. You're probably used to your own living room (or den or whatever) sound system, so burn a DVD and check out your mix there. You should also listen to/watch lots of films on your KRKs and try to make your project sound similar. That is exactly what I do when I'm not mixing in a certified facility.
 
Last edited:
Fortunately this will no longer be true in the near future. The CALM Act has been passed by Congress. Just think, no more diving for the remote to turn down commercials that rip the top of your skull off!

Official Summary

H.R.6209 - Commercial Advertisement Loudness Mitigation Act - Directs the Federal Communications Commission (FCC) to prescribe a regulation prohibiting advertisements accompanying video programming from:

(1) being excessively noisy or strident;
(2) having modulation levels substantially higher than the accompanying program; and
(3) having an average maximum loudness substantially higher than that of the accompanying program.

Best piece of news ive heard in a while :yes:
 
The environment I'm in has other users who turn the speakers' volumes up and down at will, so I go by the meters... then by ear. The meters are how I calibrate my environment before I'm able to wing it. This ensures that I'm always hitting consistently good levels with my mix... because I have a solid starting point for normal to loud dialog.

For sound work, I'm using Sound Track Pro within the Final Cut Studio package, but it's just as good to use Audacity as a multitrack studio (with frequent saves)... or: Ardour ( Mac/Linux only - http://ardour.org/ FREE!). Almost every audio package will have some sort of compressor/limiter... the difference between the two is that the limiter deals with the loud bits... the compressor will do that plus amplifying the quiet bits a little as well to normalize the soundscape of the track. Make sure to compress/limit each track separately to provide as much control over the sound as possible... that's a track for each character's dialog, a track each for hand noise and foot steps for each character, ambient, tone, FX and Music (which usually doesn't need much more than riding the volume slider throughout to cover and dodge the rest of the tracks to taste.
 
I've been using Audacity for a while.

The Spitfish, Blockfish suite of freeware plugins are decent compressors and de-essers...

Audio mastering is a bitch. Spent 6 hours doing it today and still got it wrong. My problem is I like driving a signal until it clips, even digital.

My earlier uploads were way too quiet...

Now my latest mix is too loud.
 
Last edited:
Rerecording - mixing - is probably the hardest part of audio post.

Just as a rule of thumb, dialog should be at about -20, screams and yells should peak at about -12. That gives headroom for explosions, crashes, etc. and the score to have more sonic impact.

I almost never use compression on dialog, but do use a limiter on the dialog buss. I will compress sound FX to increase their impact.

As a rule my master levels are always on the quiet side; I probably peak at about -4. Once the mix is completed I increase the gain of the entire mix to -.2.
 
I've chosen to the go the route of a Juicedlink adapter for my audio needs, and that means that the audio will be written into the video file in my camcorder. All this talk of audio mixing is making me think that I should separate the audio and video after I complete a 24p pulldown (PF24 </3). Should I edit the audio separately from the video? Or would Sony Vegas's sound features suffice?
 
Back
Top