Bristol_Jonesey
The downside to using any sort of mute, or cutting/splitting clips up and removing the breaths entirely is that the transition from an inherently noisy signal to total digital silence could and often does, sound quite unnatural and distracting.
 

Well, as I mentioned above, I think the client's requirements are kind of odd. But there's no noise. They want it all NR'd within an inch of its life, even before you get to the breaths. So the muting sounds okay.
 
I think once I've done this one according to current requirements, I'm going to try to sell them on updating the standards. To my ears, it sounds way better if we greatly reduce the breath sounds, but leave them in there. It kind of ruins the cadence of speech to completely eliminate them.

Personally I don't think this is an odd requirement at all. I do it all the time for:

• vocals (I record metal, so it's not meant to sound natural. sometimes there's bleed from the vocalist's headphones too)
• drums (mute drum tracks between hits to reduce bleed, keep kicks out of the bottom snare mic, etc)
• guitars (remove fret or muting noise before and after a passage)

etc. Chopping up one long track into many shorter audible parts is a common and important procedure.
 
Yes, a gain envelope would work, but you have to manually draw in each part. Tedious. (I don't buy the argument about it affecting dynamics processors though. If there is no audio on the track at that point then you have no dynamics you want processing.)
No, Split at Zero Crossings doesn't work. That eliminates the high frequency pop at the end, but there is still a discontinuity in the waveform that is abrupt and audible. The same would apply to any automated mute system that didn't come with a fade setting.
 
Having a setting that could generate fade-ins and fade-outs whenever a clip is split would solve this problem, and also my drum editing problem. I believe that Cubase, Reaper, Pro Tools, and Ableton all have this capability.

Yeah, I just think it's odd for the context, which is spoken word with no backing.

Ah... in that case then, agreed! :)

Might be a case of someone getting freaked out by the sound of their own voice on tape and focusing in on stuff no one else would notice and/or care about. We tend not to hear things like our breath and lungs rattling through our skulls so it seems weird hearing that stuff after performing a piece over and over and over again and all of a sudden there is this new element. The instinct is to try to make it sound like we hear it in our head which can end up being detrimental to the final product.
 
People used to pay good money to see me sing so I figured I was doing something right and it sounded good in my skull. Once I started recording it? Drove me freaking crazy to the point of embarrassment. Play it back for others who wer familiar with my stuff and they said it was what they heard and liked live except more polished so better. I still couldn't stand it.
 
Playing guitar is so much simpler. :-/

No, the people commissioning it have hired in a bunch of different readers, who are all pretty good.
 
My impression is that since the guy is using all freelancers, he's had to come up with a set of rules just to get usable work out of the less experienced people. Which is fine. I'm going to do this one according to the rules, and then send him a sample of another approach. I reckon I can do it in a quarter the time, but that'll be with breaths reduced rather than eliminated. Will have to see if he finds it acceptable.

Kylotan
Having a setting that could generate fade-ins and fade-outs whenever a clip is split would solve this problem, and also my drum editing problem. I believe that Cubase, Reaper, Pro Tools, and Ableton all have this capability.

Please see this post. It's unorthodox, but it's fast and I use this technique all the time to specify fade times when splitting clips. I use this not only for narration, but also when creating sample libraries and need to isolate hits.
 
The main limitation is that the fade in/out times are the same. I usually do 7 ms (the same amount ADATs did, LOL). Fade times are limited to 100 ms but I've never found that to be an issue.
 
If you're using Take Lanes, you need to do this with the parent track. It won't work on individual Take Lanes.

John I know they don't want this but I have found from experience that dropping breaths down by 10 dB works really well. (no clip splitting required hence no fades required either. You could drop them down by 30 db as well) Once some compression is added to tighten up and level out the VO the breaths don't come up in volume at all but they are barely there but still keep the VO natural sounding.
 
You can over complicate the issue in terms of adding fades to the start and end of clips. The way FreeFlyBertl has suggested is the way to go.  You only have to make the fade times very small eg 5 ms or less and any clicks will go.
 
I still feel an editor program is the way to go as well.  eg Adobe Audition etc.  It is just WAY faster.  No DAW can compete doing those sort of jobs.
 
And there is NO quick or automatic way to do it either.  You have to edit the VO all the way through manually to get the perfect result. 

Jeff Evans
John I know they don't want this but I have found from experience that dropping breaths down by 10 dB works really well. (no clip splitting required hence no fades required either. You could drop them down by 30 db as well) Once some compression is added to tighten up and level out the VO the breaths don't come up in volume at all but they are barely there but still keep the VO natural sounding.

 
I also prefer dropping the level, it sounds more natural. Once on a session an engineer removed all the inhales from Barbra Streisand's vocal, and she said they needed to put the breaths back in. IMHO she was right.
 

I still feel an editor program is the way to go as well.  eg Adobe Audition etc.  It is just WAY faster.  No DAW can compete doing those sort of jobs.

 
I'm not totally sure of that, given the approach I mentioned using. The one thing I do miss from Sound Forge, though, is the keyboard shortcut that auditions pre-splice followed by post-splice so you can hear the results of cutting out a section. Think I'll do a feature request 
 
That's why I like the approach I recommended. You just deposit transients where you want the splits, and split then all at once.

Craig's approach sounds great but still involves clips being split.  And the moment you split clips you really have to check how things sound over the boundaries.  Longhand, in real time.  There is still no quick way.  You will have to reconsolidate the file too.
 
What I like about an editor is once the VO has been recorded I open up the whole full length file no matter how long it is.  It is real fast to recognise the breaths and just swipe the area with one action.  Then I have got a key command setup so one button press drops the breath by 10 db or removes it.  Then you just rewind a little and have a listen and move on.
 
Occasionally you might get a click right on the end of the previous word or the start of the next word in the swiped area but mostly not.  Often there will no sounds there at all just the breath gone or dropped well down.
 
Sometimes too you may want to remove an area of silence because the reader just hesitated or took a slight pause for some reason.  In the editor you can also swipe the area of extended silence and in one key command it gets chopped out and the whole file to the right moves back in an instant.  That is what I mean by fast.
 
It is easy to undo anything too in an editor.
 
Then you just save the newly edited file with a new name and you are away.  You can always open up the original file too if you need to and go back.
 
I have always believed anyway that a decent editing program is a must alongside your DAW.  Especially if you are doing lots of serious VO work.
 
I open up every track of a multi track mix before mixing in the editor and do some prep work there. There is ton of stuff you can do there that will save buckets of time later.  Checking levels with a VU meter being one of them too.  Any quiet bits can be easily lifted or loud bits pulled down a tad.  Stray noises removed.  Noisy sections noise reduced out etc...