Voice artists, although usually an eccentric breed, have to breathe just like the rest of us! We work with hundreds of different artists at Voiceovers-Uk and I can honestly say that there is a wide variation in the level of their breaths that end up on the raw recording. Some artists are very loud and gasp before attacking a paragraph, some are more muted and at times their breath is barely audible. The truth is that a Voice artist will perform the way that is natural and most effective to them, and asking them to turn their breathing down in a session isn't really an option!
So we have done the initial recording, the client is happy with the takes and it is now down to us to produce a finished, polished voiceover.
Why are recorded breaths so audible when we barely notice them in normal conversation? That is partly because we place a mic so close to the artist (this can be as little as 2-3 inches away) that we capture the breaths at an unnaturally loud level ( If you think about it – we very rarely place our heads 3 inches from someone's mouth to listen to them !!). Also, the use of any compression (which basically evens out levels) will bring up quieter sounds in a recording, so therefore increases the levels of breaths. And of course, voiceovers have to take large breaths to allow them to get through the paragraphs in front of them with authority and impact!
So, a very important aspect of post production is de-breathing - Do we leave them in or take them all out?
Well, the answer is that it depends on the situation. Most narrations benefit from having breaths removed to give a polished feel and to allow the listener to become absorbed in what the artist is saying, rather than the fact it was a recording. The removal of breaths removes the evidence of the physical and technical nature of the recording and allows it to exist wherever the listener wants it to. In the case of Voiceovers on videos, the de-breathing allows the voice and picture to merge into one product. For example, in the case of a narration over an atmospheric video of oceans, dolphins, sea life etc. the creators will want the listener to feel that the voice is there observing and commenting on the events on the film. They won't want to feel that the voice is coming from a 6ft x 6ft Voice booth with simply a mic, a screen, headphones, a glass of water and a script in it
There are exceptions, a natural conversation type voiceover may be better with breaths left in, or simply minimised. Our experience though, on the whole, is that most clients prefer their recordings de breathed.
Traditionally we would manually edit out breaths using fades to smooth over any blips. Where the VO leaves a long enough gap between the breath and the next word, it is sometimes possible to remove the breath completely. At times room tone has to be put back into the recording to maintain continuity of the sound. All of this has only really been possible since the birth of non linear editing.
A lot of time is involved in editing breaths out and we recently decided to look at a possible software route. We demoed the Waves plug-in "De-Breath". I admit that I was very sceptical indeed that a piece of software would be consistently effective and produce a natural result. I was initially totally amazed at the effectiveness of the plug-in – which, when set up carefully, sensitively removes or minimises breaths and saves literally hours of manual editing. The graphic interface allows the user to see and analyse the breath level and to set the plug in to remove or minimise the breath- It also allows you to put room tone back in if necessary.
Any drawbacks? I'm afraid yes, and it's quite a big one. Every now and again the software will mistake a word for a breath and duck it out, and the problem with this is that it's really difficult to trust that all the files are done correctly. The last thing we want is a voiceover to dip out at some point. The answer I guess is to listen back to the whole project to check? But then if you adjust the De Breather for one errant word it may affect all the rest..Hhhhmmmm. Suddenly the hours of saved time disappear into a corridor of uncertainty.
It's great to use on short sessions (radio / TV Ads etc) mainly because it sounds very polished and even. But when you have 650 VO files to batch process..For the moment its back to manual..At least for now!