ratio between audio and background music?

[Phil Gadd]

Hi

I recently put together a few business product videos, where there is some background music continually through the film.

Out of curiosity, is there an industry standard ratio i.e. when someone is talking, how low do you put the background music?

Currently I just do it by ear but its not ideal really if I want to do things more professionally.

Any ideas anyone?

[Paul R Johnson]

Music is fluid and has a constantly changing dynamic range - so any formula or maths based approach won't work, so your by ear approach is best. Lots of techniques - some style of music need to be very low and almost inaudible, while others need to be a feature. The thing to watch for when doing it is to always try the balance on both a system with extended bass and 'hi-fi' quality sound, AND a rotten old flat screen with two tinny (and tiny) loudspeakers. Often if the music has lots of low end energy, it's easy to get the balance wrong.

[Gary Nattrass]

It can also depend on how the dubbing mixer has set out the compression on the voice and music to create the correct dynamic range, that can be from gentle to the extreme and is related to the feel and genre of the piece.

So if I am mixing some nice strings with a gentle narrative voice I would have a totally different dynamic range and mix to a slap bang in your face commercial or rock music show.

As a further example careful setting up of my compressors on sports coverage will mean that no matter what happens I can make sure that the commentator is always heard but the effects are just as loud and add to the excitement of the coverage.

I have also done drama scenes in a nightclub where the music can be absolutely banging away but you can still hear what is happening with the dialogue, use of stereo or surround soundstage can help here as you can make the music big and roomy with lots of stereo small room reverb but make the dialogue narrow band so it sits in the middle and can still be heard OK.

P.S I do it professionally and tend to do it by ear rather than technical levels although there may be delivery specifications to be met for maximum levels and overall loudness.

It's all subjective though and it was always amusing on programmes like Heartbeat to mix on the cautious side but still get the letters of complaint from Mrs Cannybody in Byker that the music was too loud!

[Phil Gadd]

Thanks Guys

I know when i used to mix my backing tracks for stage, I have a nice set of JBL studio speakers and I was given a tip to turn the music low and this would give me a feeling for certain elements being too high i.e. bass etc.

Maybe I should re-invest in some quality speakers (only using desktop JBL's at present JBL Spot Black&White 2.1 Active Speakers | av home cinema systems)

[Gary Nattrass]

Std practice for TV and film is to balance on your main speakers but also do a check on a TV or small speaker set-up with a more narrow bandwidth.

[Garrett Low]

Gary, do you ever use EQ to create a whole in the frequency range of the voice you want to come through, something like you would do when mixing vocal music?

[Gary Nattrass]

Originally Posted by Garrett Low

Gary, do you ever use EQ to create a whole in the frequency range of the voice you want to come through, something like you would do when mixing vocal music?

Yes but more so in film or surround mixing where the dynamic range is far greater and you have a dedicated centre speaker.

Dialogue can be quite narrow band anyway with a high pass filter of up to 150 hz and an EQ lift at around 3.6k, some people roll everything off at 8khz too but it all depends on what the delivery is for.

On sport productions the commentary is narrow band anyway with sennheiser HMD25 headset mic's or lip ribbons being used and if I am doing live presentation I use a high pass of 160 hz to get rid of any wind noise or PA system bass nasties.

[Garrett Low]

Thank Gary, I've never tried it in corporate videos but maybe on my next one I will. A bit more work than just dropping the bg music volume and applying more compression but it may end up sounding better.

[Gary Nattrass]

What can also be very effective in proper audio dubbing is splitting dialogue, effects and music off into separate sub groups or stems each with their own compression or limiting inserts!

You can then balance things easier and set the gain structures to offset dynamics, this is what I do as standard for live sport events and use 1khz line up tone to ensure that the crowd effects are always compressed more than the commentator, if I recall correctly an offset of about 4db's in dynamic range tends to be the norm but it all depends on the audio desk being used!

[Jon Fairhurst]

I apply an EQ template to projects where I want the music and voice to both come through loud and clear, like in an action scene where the music is pulsing heavily, there is a lot of Foley or Sfx like car engines in a chase, and then the characters yell something at one another. I don't want to duck the music and Sfx so much that it sounds like compression pumping, but I don't want the dialog buried either and I'm out of headroom.

Note that I'm not mixing for the big screen, so I'm doing more of a "loudness war" mix than one would for cinema.

In my template, I do a low cut on the voice up to about 100 Hz. I might boost the music a bit here. The 200-300 Hz range is the voice fundamental and the precise frequency depends on the speaker. I cut the music here. 600Hz is not critical for voice, so it gets a cut. 1.2 kHz is critical for consonants, so I'll add a (possibly sharp) peak here and cut the music a bit. 2.4 kHz is an interesting space. It's the "nasal" area. It helps us differentiate similar voices as well as an oboe from a flute. Boost a dull voice here, but cut an overly nasal voice or a harsh viola/oboe/bassoon in this range. So, it depends. Above 5kHz, one can intermix the sources without headroom or clarity issues. Just balance "air" and noise on your stem tracks.

I apply my basic template and then tune it, depending on speakers and music. I set and forget across the track, unless I'm fixing a specific problem, and I'll do that with an additional EQ plugin. With the set-and-forget approach, nothing pops out or changes during the scene.

Of course, well composed music will make some space for the dialog by lowering complexity and intensity to avoid stealing attention from the script. Like in a good rock song, you don't play your guitar solo in the middle of the chorus. But that's not necessarily taken into account when using library music.

One other note: EQ dips can be sharp, especially when cutting annoying frequencies. EQ bumps should be smooth to avoid adding resonances. The exception I find is around 1.2 kHz on dialog. One can add an extreme sharp boost here to add clarity to consonants. However, listen for resonance issues and tune to avoid those problems.

This lets me keep everything loud and clear without compression pumping. And when you're mixing for the web, keeping it loud is generally important.