Audio EQ Settings For Voice And Music Mix?

[Kevin Wayne Jones]

I'm trying to get a good audio mix on a FCP project.
A voice-over and music playing at the same time.
I want the music to be nice and loud,
but still be able to hear and understand the VO.

I seem to remember someone suggesting using the equalizer effect and setting the voice and music in such a way that they were complimentary to each other without competing.

If someone could make a reccomendation or suggestion I would appreciate the help.

Thanks.

Kevin Jones
Jones Media Productions

[Mike Demmers]

This has to be tuned to the specific voice and music bed, but roughly, you want to do something like this:

Look for two frequencies in the voice (temporarily use eq boost to find these, they should be obvious when you find them). One frequency is the where the main 'body' of the voice is. Typically somewhere around 500-1000 Hz. The other is where the 'presence' of the voice lives, typically in the 2000-4000 Hz range.

Set the voice (with no EQ again) to your reference level. Now bring the music up around it to about where you would like it.

Now adjust the EQ on the music to the 'body' frequency found earlier, and begin to cut the eq on the music. You should hear the voice becom more clear as you do this, more intelligible.

If that is not enough, you can them also drop the 'presence' frequency in the music found earlier to make the voice sound even more clear.

Sometimes it helps to drop some of the low end of the voice (just a little). Sometimes it helps to add a slight presence boost to the voice (just a little). But it is usually best to alter the voice much less than, or not at all, compared to the music.

Some mild compression on both may help.

If the voice is not even in level, you need to level it first for consistent results.

If you want it to sound like a radio ad, do all the above to great excess...

That should get you started.

-Mike

[Shaun Roemich]

Also, throw in AGGRESSIVE compression on the voice over track if you want that "radio ad" presence. By lowering the overall dynamic range of the voice, you make it possible to raise the overall level (since compression by itself ALWAYS lowers volume) of the voice. The quiet sections therefore become louder. And then coupled with the EQ "trick", you should get your "pop".

[Robert Lane]

This thread should be moved to the "All Things Audio" as a sticky; this is one of the most timely and to-the-point explanations of how to get good VO tracks that I've seen explained and is spot-on on how to correctly setup voice and musicbed for proper balance.

Also, getting good compressed voice-overs is easily done with the "Voice Channel" from ART. It's a channel strip that can act as a standalone device and has 5 different outputs: XLR, TRS, SPDIF, USB and TOSLINK. It also has a WordClock connection if you use it. The output is amazingly clean, bright and the built-in compellor (compressor/limiter) along with built-in EQ allows you to make the vocals perfect first-time out without having to tweak them later in post. It also won't break the bank.

[Shaun Roemich]

Robert: that thing is a BEAUTY! When I get relocated and settled in, I may need to invest in one...

[David W. Jones]

While there are quite a few low-end channel strips out there similar to the ART Voice Channel, it's still not a replacement for learning how to properly mix a VO.

[Robert Lane]

Originally Posted by David W. Jones

While there are quite a few low-end channel strips out there similar to the ART Voice Channel, it's still not a replacement for learning how to properly mix a VO.

Which is why I suggested this thread become a sticky; Mike's "how to" post is spot-on.

[Mike Demmers]

I think this is too minor for a stickie. But I have seen this asked quite a few times, so perhaps this could be added to the already existing audio FAQ stickie?

Not sure how to do that...

[Mitchell Lewis]

I mix voice with music almost on a daily basis producing television and radio commercials. Here's my setup for good voice over with music in FCP: (NOTE: I normally do this with our audio software as the effects have a far better GUI than FCP)

VOICE
1) Add the Effects>Audio Filters>Final Cut Pro>Compressor/Limiter effect to the voice track
2) Set the ratio to 3
3) Check the box that says Preserve Volume
4) Adjust the Threshold to taste (listen for pumping sound)

MUSIC
1) Add the Effects>Audio Filters>Final Cut Pro>3 Band Equalizer
2) Set the Mid Frequency to 1550
3) Set the Mid Gain(dB) to -5 to -10 (make this adjustment while listening to both tracks playing together)
4) If necessary, adjust the level of the music track (command+option+L)

Ideally, you would have a second Compressor/Limiter set to cover both tracks. This would bring the Music track up in between the pauses in the narrators voice. But FCP doesn't allow you to add Effects to the Audio Mixer. I sure wish it did, it would save me a lot of time going to our secondary audio program (MOTU Digital Performer) and back to FCP.

[Robert Lane]

Originally Posted by Mike Demmers

I think this is too minor for a stickie. But I have seen this asked quite a few times, so perhaps this could be added to the already existing audio FAQ stickie?

Not sure how to do that...

"This thread should be moved to the "All Things Audio" as a sticky..."

...was my original request; it's up to the moderators to decide to move it and/or sticky it.

[Chris Rackauckas]

To tell you the truth, I think you are looking a little too simplistic. Mike's post will get you close, but you need to tailor it to the voice and to the music. Send it through a frequency analyzer and look for where it's resting. From there play with the eqs a bit and find what you can dip without cutting any of the body out of the vocal, and boost somewhere in that mid range (depending on the voice. In general, males are lower than females, but you never know. Analyze it! Don't generalize here!). Cut some mids out of the music respective to the music itself. Over compress the vocals a bit always helps, but if you do so MAKE SURE YOUR SOURCE IS CLEAN! Otherwise you will hear a lot of noise since the compression will dramatically raise the noise floor. Lastly, high pass the vocals so that way they aren't taking up any of the low range that they don't need to. That will keep the bass and the kick drum pumping so you can hear the beat loud and clear, meaning you can feel the music driving even if you killed the mid range (and that movement is really all you need from the music in the background).

However, I can't stress anymore that you can't just say that there is a magical frequency for all, it changes by the voice actor, how he feels that day, the mic, what music you are blending it with, etc. Tailor it to whatever you are doing and don't just go by a generalization found on the web.

[Mitchell Lewis]

I think Chris is mostly talking to me with his post.

I've been using the frequency of 1550 hz for years. I've tried to adjust it many times, but to my ears (degree in audio engineering and 20 years of mixing audio and video), 1550 always seems to work well. Yes, 1600 or 1650 or 1400 or 1450 might work a tiny bit better in some instances. But this isn't rocket science, you don't need to over think it. If you want, try it for yourself. Here's how...

Get the vocal set up using steps 1-4 (note these are designed to work when a voice is mixed with music for television or radio commercials. The setting are too much for a solo voice. In that case, try a Ratio of 2 and tweak the Threshold setting) Once you've got the voice set up, turn on the music track and add the EQ effect. Drop the mid-range eq level down a large amount so you can really hear what it's doing (-10 or -12) and then sweep the frequency from 500hz to 3000hz and pick which setting you like the best. Then bring the eq level back up to something around -3 to -5. Honestly, these aren't precise settings. Some days I might use a bit more of a dip in the eq. Other days I might bring the volume level of the music down. Try it and you'll see what I mean.

All your doing with this technique is "cutting a hole" in the music to make the voice lay in better.

Also, sometimes the music is too dynamic and it too needs a bit of compression. You can try adding the Compressor/Limiter to the music, but don't go with an extreme a setting as the voice. Try a ratio of 2 (or maybe 1.5) and play with the Threshold so it's just effecting the music a little bit. Doing this will make the music less dynamic and thus makes it easier to mix with the narration.

Hope this helps! :)

[Mike Demmers]

You do really have to use your ears to do this, there is no paint by numbers method that will get it right. EQ will change according to the specific voice and music, and amounts of compression differ as well - compressors have different effects according to type, and different amounts sound good on different kinds of voices and types of music.

There is one other thing that no one has mentioned: How do you know when you have it right?

To really get this right, you need to have at least two speaker systems. A small, single driver system (like Auratones - a single driver works best for this purpose) and a full range system. I actually use use three: small, medium, large. But two is usually good enough.

Set the balance on the small system. When you think it is right, throw it up on the full range system. The relative balances should not change. Yes, it should sound bigger and fuller on the large system. But the basic relationship between the voice and music should remain the same.

Further, on each system, change the volume from pretty soft to pretty loud. Still, the relative balances should remain pretty close, the voice perfectly intelligible. At very low levels of course the music will lose a bit of high and low end due to the ears own losses at low levels (Fletcher-Munson curves, Google it). At higher levels make sure the voice is still intelligible, though the music will seem to have more depth.

Now, what if it doesn't sound the same? Well, this probably means you have an EQ problem on one or the other. For example, a very boomy voice might not sound so boomy on the small system, so you set it to a certain level. Then, on the big system it sounds too loud. So maybe reduce the low end of the voice a little, so they match again between the two systems. Or the same type of problem could be in the music.

Again, there is no rule here, just a hint that if it sounds too different on diferent sized speakers, look to the EQ balances for the problem.

-Mike

[Chris Rackauckas]

Originally Posted by Mitchell Lewis

I think Chris is mostly talking to me with his post.

I've been using the frequency of 1550 hz for years. I've tried to adjust it many times, but to my ears (degree in audio engineering and 20 years of mixing audio and video), 1550 always seems to work well. Yes, 1600 or 1650 or 1400 or 1450 might work a tiny bit better in some instances. But this isn't rocket science, you don't need to over think it. If you want, try it for yourself. Here's how...

Get the vocal set up using steps 1-4 (note these are designed to work when a voice is mixed with music for television or radio commercials. The setting are too much for a solo voice. In that case, try a Ratio of 2 and tweak the Threshold setting) Once you've got the voice set up, turn on the music track and add the EQ effect. Drop the mid-range eq level down a large amount so you can really hear what it's doing (-10 or -12) and then sweep the frequency from 500hz to 3000hz and pick which setting you like the best. Then bring the eq level back up to something around -3 to -5. Honestly, these aren't precise settings. Some days I might use a bit more of a dip in the eq. Other days I might bring the volume level of the music down. Try it and you'll see what I mean.

All your doing with this technique is "cutting a hole" in the music to make the voice lay in better.

Also, sometimes the music is too dynamic and it too needs a bit of compression. You can try adding the Compressor/Limiter to the music, but don't go with an extreme a setting as the voice. Try a ratio of 2 (or maybe 1.5) and play with the Threshold so it's just effecting the music a little bit. Doing this will make the music less dynamic and thus makes it easier to mix with the narration.

Hope this helps! :)

Yep, two monitors help. I just have my N12 set so that way setup B is my KRKs and setup A is just a mono pair of computer speakers. It helps a lot.




And Mitchell, I wasn't directing that at you, actually, I skipped over that part your post where you mentioned the 1500 hz. I was just posting what I thought without pointing the finger. Everytime I try to set a preset way to do something it just gets killed by some random change, so I stopped trying. Maybe you are working with better and more consistent talent in a better room and that's the reason, but for me I have to take it case by case to get the job done.



And about music being too dynamic, just hint:

If it's modern, it has no dynamics. Search loudness wars and Bob Katz's vendetta against it, modern music has killed dynamics, so you really don't have to watch out for it. Older music, on the other hand, may need to get squashed a bit. Again, that's a generalization and listen to it first, but for this there is a pretty stark difference.






Oh yeah, and as you can see from the fact that people argue about this: everyone has their own way. If it works for you, do it.