music and dialogue - punching a hole

[Josh Bass]

Ok, so you're watching TV/a movie, and there's crazy SFX, music, the whole bit, but somehow, even whispered dialogue comes through the mix so that you don't have to strain to hear or decipher it. What's the secret?

A friend of mine gave me some tips on creating a hole in the. . .uh. . frequency spectrum, where you basically cut out certain bands from the sound track, that are normally the bands that human speech occupies, and then the dialogue cuts through because you don't have the same bands in the soundtrack competing for your attention.

Sounds good in theory, but in practice, it significantly alters the "tone", if you will, of the soundtrack to an undesirable degree, making music sound very odd and tinny.

Is good old Vegas just not good enough to pull this kind of thing off, or am I doing something wrong?

[Douglas Spotted Eagle]

Vegas is just as good as anything else for this.
Creating a hole is done with automation (which Vegas 4 and newer has) and compression. Using a mix of EQ on the soundtrack, using a compressor and EQ on the whisper, you should be able to make this fly. Keep in mind it's also placement. If the soundtrack is wide, there is lots of room in the center. If it's not wide, find a place in the spatial stage to locate the whisper.

[Josh Bass]

I'm sorry Doug, all that went over my head. Can you elaborate?

[Douglas Spotted Eagle]

When you're punching a hole in the soundtrack, you'll need to automate the EQ on the soundtrack, so that the frequencies roll out during the whisper, rather than knocking them out for the entire playback. You might consider using a widener like the S1 from WAVES or the Ozone from iZotope to widen the background/soundtrack a bit in the stereo spectrum. (right to left space)
In the whisper, pop the hi mids so that the whisper is cutting, use a fair amount of compression, something like 4:1, and coupled with the automated EQ in the soundtrack, this will allow the whisper to cut through hard.
If your sound track has percussives in it, this will be harder, so you'll want to use panning to find the place that the whisper best sits in the spectrum.

[Josh Bass]

Ok, starting to get it. Any of those. . .wideners. . .free?

The whisper was just an example of the issue I was having problems with, but I assume the principles still apply to regular volume speech?

[Douglas Spotted Eagle]

No, none of the wideners are free. WAVES is about 400.00, iZotope is 300.00. Both come with additional tools.
the same principles apply to any kind of audio mixing.

24p is overrated, my mom is underrated.

[Josh Bass]

Durn. Thanks.

[Ty Ford]

Now you see why audio people get paid a lot to design sound and do mixes. Cutting a hole with EQ is a last resort, btw.

Part of the magic is........um, I forgot...:)

Regards,

Ty Ford

[Dennis Vogel]

<<<-- Originally posted by Ty Ford :
Part of the magic is........um, I forgot...:) -->>>

This is why I like this place!

Good luck.

Dennis

[Josh Bass]

So what's the secret to getting the dialogue to "cut", when the volume of the music is fairly loud, I mean, I know it's not just a matter of making dialogue louder.

[Ty Ford]

Ac tually turning down the music to the right level is the correct approach. Having a dialog track that successfully cuts is a brain salad of having the right voices, mic, and knowing 23.6 things about EQ, compression and limiting in order to allow the voices to prevail. Also essential is having good monitors in the right environment to make the best decisions.

Great mixes are seldom simple. they just sound that way.

Regards,

Ty Ford

[Josh Bass]

Ha! Sounds like I'm screwed til I get budgets and stuff.

[Jeremy Davidson]

It seems a lot of stuff on TV (especially commercials) makes very liberal use of compression and limiting. They've squeezed the range between loud and soft so that it basically ranges from loud to... um... loud.

For videos I've produced with dialog over music, track automation (aka envelopes aka "rubber bands") seems to work well. During the parts with speech I reduce the music track so that it doesn't swallow the dialog.

EQing is another way (I've heard it's typical to cut around 2k-4k), but it doesn't necessarily need to be so drastic that it noticeably affects the music quality. If you can play/loop a track while tweaking the EQ you may be able to find the right frequency to make the speech "pop out."

In regard to widening, a mono signal is as narrow as you can get (with a typical stereo speaker setup, it sounds like it is coming from directly in front of you). Good place for the dialog itself, but most music will sound better in stereo (I'm still learning how to produce music that works like this).

Anyway, that's my take on it (again, I'm still learning, so don't just take my word for it).

Spot: EQ automation? 'Sounds interesting. I'll have to try that one sometime.

Ty: There are only 23.6 things to know? I thought it was more like 31.2. j/k

[Josh Bass]

Thanks, I'll keep all this in mind.