Dealing with vocalist's breaths

[1ace1]

Originally Posted by DonaldM

Option two would require buying another plug-in...not in the budget right now.

You can make your own de-esser using a duplicate track of what you want de-essed as a control source for the key input on the side chain of a compressor.

EQ the frequencies you want removed on the duplicate track so they are more audible. Send it out to bus 1.

Place the compressor on the original track and use the key input (set to bus 1) to control the compressor.

Homemade de-esser!

A:)

[DigiTechSupt]

Originally Posted by albee1952

I agree and rarely remove them, but some singers just get too loud. 2 simple routines solve DTS's concerns with my method. Duplicate the playlist first(never a bad idea) and consolidate the track when you finish the edits.

But that's many more steps overall than my method (just sayin'!).

With my method you don't have to duplicate the playlist, nor do any consolidation afterward.

If there are a lot of breaths, you can save considerable time as you don't have to select and process each one (select, open audio suite, gain it down, check it, repeat as necessary...), but just select and pull down the automation.

But hey - everyone has their preference...

[daeron80]

Two counterintuitive things I've discovered that can work brilliantly on noisy breaths:

1. The Digi Dyn3 compressor, Knee 15, Ratio 100, Attack 10.0 uS (that's micro, not milli!), Release 5 ms, Threshold variable around -40 or -50. Enable the LF filter in the Side-Chain section, high pass, around 3 kHz. That will cause the compressor to work only dependent on that high frequency racket in the breaths. Just smack the livin' daylights out of each breath. It looks crazy, but it sounds smooth as silk. It's basically an analog brick wall limiter emulation except it's operation is only high frequency driven.
2. If breaths have a lot of ugly saliva sounds or mouth clicks in them, the compressor may pump. In that case, use a click and crackle noise reduction plug-in with extreme settings. BIAS Soundsoap works great for that. I even like the cheaper Soundsoap better than Pro for this one application (Pro is superior for other things) because there's just one slider for click and crackle. Jam it to the top and process.

As with the other Audio Suite methods, short crossfades at each end. There's always plenty of room. Just zoom in. You can make them as short as 4 samples if necessary, but 10-30 ms is ideal.

[DonaldM]

Originally Posted by daeron80

Two counterintuitive things I've discovered that can work brilliantly on noisy breaths:

1. The Digi Dyn3 compressor, Knee 15, Ratio 100, Attack 10.0 uS (that's micro, not milli!), Release 5 ms, Threshold variable around -40 or -50. Enable the LF filter in the Side-Chain section, high pass, around 3 kHz. That will cause the compressor to work only dependent on that high frequency racket in the breaths. Just smack the livin' daylights out of each breath. It looks crazy, but it sounds smooth as silk. It's basically an analog brick wall limiter emulation except it's operation is only high frequency driven.
2. If breaths have a lot of ugly saliva sounds or mouth clicks in them, the compressor may pump. In that case, use a click and crackle noise reduction plug-in with extreme settings. BIAS Soundsoap works great for that. I even like the cheaper Soundsoap better than Pro for this one application (Pro is superior for other things) because there's just one slider for click and crackle. Jam it to the top and process.

As with the other Audio Suite methods, short crossfades at each end. There's always plenty of room. Just zoom in. You can make them as short as 4 samples if necessary, but 10-30 ms is ideal.

That's an interesting approach. For this session, though, Albee's suggestion seemed to work just fine. By just reducing the gain everything sounds a lot more natural, and I removed some altogether just because they didn't sound right in the phrasing.

[albee1952]

Originally Posted by DonaldM

That's an interesting approach. For this session, though, Albee's suggestion seemed to work just fine. By just reducing the gain everything sounds a lot more natural, and I removed some altogether just because they didn't sound right in the phrasing.

Interesting for sure. FWIW, I have never needed to add a crossfade when doing my routine. Maybe I'm just lucky or have really good aim(yeah, its luck......)

[daeron80]

I developed that approach editing old spoken word messages that were monstrously over-compressed with a fast release so all the breaths were at the same volume as the voice. Had several series of 20 to 40 minute messages mistreated that way. The breaths are completely exposed, of course - there's no music to mask anything. A missing crossfade is easily heard. Any roughness becomes a distraction. I would go through the whole thing squashing every breath with the compressor, then do batch 30 ms crossfades, then start the actual editing process. Works brilliantly. Sounds great on radio.

[Kryst]

Make the singer do it again and introduce electro therapy.

[daeron80]

Electro shock could work. Or just give her an alternate mix where each gasp is raised rather than lowered.

[Kryst]

Originally Posted by daeron80

Electro shock could work. Or just give her an alternate mix where each gasp is raised rather than lowered.

Nice one.

I'd also like to suggest the Electro Shock treatment may also give you some cool effects! LOL Like live electricity currents, a shrieking after each gasp...

[chrisdee]

The most brilliant way to deal with breath is to get the vocalist to learn to sing both on in and exhalation