The narration track he gave me was pretty much unusable. A little noisy, there’s all kinds of mouth sounds, odd breaths, sloppy pronunciations . . . How can I salvage this?
Well, this is why he’s in a booth and we’re in the control room – you gotta crank that shit up to hear the imperfections. It’s also why we are the producers instead of letting people self-record.
Many years ago, an old boss insisted that I sometimes use this voice-over guy who lived in the middle of nowhere and was cheap. You sent him the copy, he read it in his home studio and uploaded a file for you. Every single time, I had to tell the guy to stop editing the tracks; that’s my job. I need options, variations, etc. The raw material. Nowadays, of course, you can do the whole session virtually and “produce” him online.
My advice is to replace him. That’s the best thing for the project. You didn’t say if this was for a radio spot or a video narration, or what, but the requisite audio surgery may or may not be worth your time. Professional VO artists make phat coin for a good reason. Look at it this way: for $250, they can bang out a session in ten minutes, sound like velvet doing it, and get you both home early. Or you can find someone for free that will take 45 minutes to read and two more hours of editing and processing just to achieve a mediocre result. Is it worth it? Depends who you work for.
In 14 years I’ve only had to send “professionals” home four times. They just weren’t cutting it and we were spinning our wheels (two were VO, and the others were on-camera). However, I’ve replaced dozens, maybe hundreds, of amateur/free voices. Usually I don’t even tell them; I just wrap it up quickly and cut our losses. The pros deserve to hear about their performance, though, because there are lots of folks billing themselves as VO artists who have no business doing so. It’s an extremely competitive business, and we tend to use use the same proven handful of golden-throats over and over.
I used to run a DAT backup of the entire session in case there was a glitch in the workstation, and there are a few gems in the collection that make nice holiday reels. Some are actually from my talkback, like the time I lost my patience after the 37th take on the same phrase, and told the guy I was gonna come back there and beat him. He was laughing, I wasn’t (we didn’t have visual contact at that studio). Many are off-color jokes (something about a bulletproof booth brings out the bravado in some). And one other time two people were doing a book-on-tape scene together (about financial planning) that ended up hooking up after the session, and all their thoroughly-unprofessional-yet-hilariously-steamy banter between takes was forever preserved.
Most of the problems I hear are not due to vocal technique. These days, clients tend to like a little personality instead of just the Voice of God, so some minor quirks are almost welcomed. Generally, the lousy VO people are lousy readers. Good readers have a “buffer” in their brain that allows them to read one phrase ahead while their mouth is performing the previous one. Very little takes them by surprise, and they understand the context so they can inflict the proper inflections. These are the people you want for long-form narrations. Everything else is coachable on the spot.
As for removing room tone or hum, there are lots of tools available, and they all work pretty much the same way. The one I use is called X-Noise, and you feed it a short sample of the noise with no voice under it. It analyzes that, then applies it to the rest of the track by either phase reversal or notch filtering it out, I’m not sure. Magical. But you can’t apply it to excess or you compromise the vocal tone, as the voice is often in the same general frequency range as the noise.
The rest is clever editing. Don’t just cut out all the breaths, or it’ll have this subliminally-claustrophobic feeling. Always put your cut just before the first waveform of the sentence but after the inhale, with the inhale for the following sentence ending this region. This way each region begins clean when you move it around or use it elsewhere. You can fly in syllables from other words. Learn to recognize the various sounds that make up language; there aren’t as many as you think, and they can band-aid a slur sometimes. Not much I can do about mouth noise in the middle of a word. Of course the producer should’ve caught all of these things during the session.