Looking for solutions to avoid Mouth noise ?

Mouth noise - tongue clicks, lip smacking, swallowing, teeth clacking … are as I understand it, the bane of the industry. Now, how do I avoid it.
Apart from employing some of the arcane solutions that abound out there… Although probably the best one is eating a Granny Smith apple! Could improve my health no end.

I am on a mission to solve this to my own satisfaction, and have read lots of interesting things from very experienced people. The bottom line seems to be.
You can’t avoid it. You just have to edit it in post recording. Pre recording, apply the usual, make sure the narrator is well hydrated up to two hours before the session - with water… eat a Granny Smith, don’t under any circumstances drink coffee or tea. Natural Apple Juice is fine…

As I said … You just have to edit it in post recording
Difficult and time consuming. Here’s a good example.

This is the waveform view of a tongue ‘click’. I can hear it, but I can’t see it. Not in this view. Not even on great magnification.

This is the spectral view. And presto bumpo - there it is. So I should also ask, how do I tell what frequency this actually is?

I can go through - slowly, and edit out those that stand alone - much like this. Some are even more clearly defined, making it relatively easy. If still time consuming, which I don’t mind really because it’s al lpart of the process. If I want someone to be happy with my recording, I need to take the time and patience with it.

I can’t apply overall filters, like click and pop removers - because on an audio - voice - track, that removes the “sound” of the voice as well in some areas. In other words, it can trash the recording.
I don’t think I can apply a filter for exactly that part of the spectrum, because each noise is a little different, and still within range of the human ear.

So, from one of the great licking and slurping masters of the art, I’m also all ears…

Right. The characteristics of your mouth and voice that make you sound like you, also produce the clicks and pops.

It may not seem like it but this is the exact same problem from posters who want us to split two voices in the same track into individual tracks. Probably not. Also see: leaving the TV on behind your voice by accident. That’s usually enough to kill your show.

This is Chris Pratt presenting the voice of Emmet in The Lego Movie.

They could get anybody to do this work. It’s a cartoon character, but they wanted someone who can act the script…and whose mouth doesn’t click and pop.

Koz

… as I said.

so meantime, continue with editing in Spectrum view.

I seem to recall that Paul Licameli wrote an effect for removing mouth smacks - but I can’t find it right now - anybody?

Peter

I posted some links in this post: Something else that confounds me ... - #111 by steve

So what technique do you use, other than filters? Do you simply cut or do you use “silence audio?” Which sounds better? Which is faster? A method that takes four mouse clicks is better than one that takes eight mouse clicks. Or at least it seems that way after an hour of hunting and killing breathing noises.

As a rule, you can’t Silence Audio. That will leave tiny Blackness of Space sounds in the performance.

It is recommended you paste a short snippet of room tone (normal background noise) into the hole. That will generally be a more theatrically graceful and pleasant solution.

Whatever you do will cause the normal five-to-one editing rule to greatly expand. This accounts for the intense search for a global de-clicker filter.

Koz

Thanks to Koz for the brutal truth. I’m new but the 5 to 1 rule is starting to seem optimistic. I guess I’ll have to learn to breathe quietly.
I just blew it by asking this question in another post. Sorry about that. Thanks for all the help.

the 5 to 1 rule is starting to seem optimistic.

One is scanning the work to see what needs to be done and Five is playing the finished work to make sure you didn’t screw up. So there’s 2/5 right there and we’re not even editing yet.

People do what you’re doing all the time and you just need to know unless something changes, you’re not going to be Quick Turnaround Charlie. Like I posted up the thread, lots of nights and weekends.

There was a horror story of a newbie Producer who walked into a large video house in Hollywood with an hour television show in pieces and told everybody the show was going to air in about four hours. After they found he wasn’t joking, they marshaled virtually the entire facility and all the people in it, split up the work into chunks and smashed it together in time to meet the satellite uplink slot.

“Hello, New York, this is Hollywood Center. Do you see our bars and tone.”
“Boooooooooooooo”
“Roger that Hollywood. Break your tone. Restore. Break. Restore.”
“Transmission is recording. Start Playback in Five… Four… Three…”

No word on what happened to the Producer, but assuming he didn’t get trampled and squashed, he’s probably in a nice home somewhere.

Koz

You mention the “intense search for a global declicker filter.” I would settle for a quicker way to paste in room tone. As it is, I’m deleting a noisy bit, then finding some room tone to copy and paste, then doing it. Then I listen for the next problem but when I find it and cut it, the noise goes to the clipboard, so I have to find and copy some room tone once again before I can paste and the amount of room tone I copy might not be the same length as the part I cut. I’m not a programmer but would it be possible to store ten seconds of room tone that wouldn’t get erased every time you cut something? Then the program would measure the length of the cut part and replace it with a section of room tone that was the same length? This would fall short of a global clicker filter, but would save a lot of mouse clicks and therefore time. And the timing of the unaffected words and phrases wouldn’t be changed.
This seems doable to me but as I said, I’m no programmer. I will be dreaming of such a routine as I re-do 30 minutes of narration…

Have a look at this post: Punch Copy/Paste

I’m going to try that the next time I work on my own edits for mouth clicks and noise. That part is the most time-consuming process for me. Sometimes I can successfully just isolate the click and delete it and no one would know otherwise. But many times, deleting it causes a noticeable glitch in the sound, which is unacceptable. I’d read about using the generate ‘silence’ effect on the mouth noise/click area but when I tried that it sounded even worse.

Punch copy/paste is an awesome feature. I just found out about it a week or something ago. I assigned a custom keystroke shortcut when I paste so I don’t have to mouse click on the menu all the time. It’s saved a lot of time. I just do basic copy and paste right now. I don’t understand yet how to use the other parts of the tool.

These are things I’ve also learned over time. Something that has helped for me are the granny smith apples - but I’ve also found that just regularly swishing my mouth out with water before I swallow it helps about as well. Neither the apples or the water have a long-lasting effect, especially when I have to do multiple takes of a line or paragraph. If I didn’t have to do multiple takes, the apple/water treatment lasts for maybe half a page. I haven’t found that hydration makes much of a difference in my case. I already drink about a gallon of water a day and still make noises, so … :confused: At any rate, I have gotten good at noticing when it happens and stopping right there to eat an apple slice or swish with water; I’ve also gotten better at being able to enunciate words in such a way that the clicks/noises are reduced.

I’ve also gotten better at being able to enunciate words in such a way that the clicks/noises are reduced.

So you’re turning into a voice performer.

Koz

Yes, it’s true isn’t it, that one’s diction improves over the course of a recording session. Such that sometimes, you listen to the first track you did, and think maybe you should start the whole thing again. :slight_smile:

As for cut and paste of room tone, that’s easy. I use Spectrum View now for editing the clicks and slurps, and so I find a nice piece of room tone, say 0.5 seconds long, highlight it - , and on the Mac, use cmd+c to copy the segment. Windows is Ctrl-C, and when I find a ‘click’ that needs editing out, highlight the click and then cmd+v to paste the room tone over it. The bit of room tone stays in memory, so you can use it over and over… I know it has the potential to change track timing a little, but that’s easily fixed anyway.
So as part of the editing process, it’s not too onerous.

you listen to the first track you did, and think maybe you should start the whole thing again

It’s the curse of the first audiobook.

ACX tells us consistency is important but possibly even more important, a do-over may be a more enjoyable experience for the listener.

If you’re a serial audiobook performer, this only happens on the first one.

Koz

These are all good ideas that will help me. Thanks to all who answered.

I know, right? My problem is that I don’t have time to start the whole thing again at this point. I spent so much time fighting with settings and levels and figuring out how to pass ACX Check that I lost about a week or so of recording. I’d already done the first 15 minutes, which got approved by the author, but ended up re-doing all that once I figured out what worked best.

As for cut and paste of room tone, that’s easy. I use Spectrum View now for editing the clicks and slurps, and so I find a nice piece of room tone, say 0.5 seconds long, highlight it - , and on the Mac, use cmd+c to copy the segment. Windows is Ctrl-C, and when I find a ‘click’ that needs editing out, highlight the click and then cmd+v to paste the room tone over it. The bit of room tone stays in memory, so you can use it over and over… I know it has the potential to change track timing a little, but that’s easily fixed anyway.

It shouldn’t change track timing if you use the punch copy/paste feature. It’ll insert the exact amount of time that the click takes up when you use punch paste. That’s the beauty of this feature. I’ve also been making sure to hit the “z” key for find zero crossings before I punch paste. I’m not sure if that really makes much of a difference in my case but I’ve read that it’s supposed to help avoid clicks at edit points.

It’s not that I’m worried about track timings with audiobook chapters. That’s why the simple cut and paste is ok at the moment. Mostly I just delete the click/noise whatever. If it shortens the gap - changes the speech pattern - too much, then I past in a bit of room tone.
I’ll have a look at the punch/copy/paste thing, but am thinking it’s just one more thing to remember. There are a few things I’d like to see Audacity be able to do before I worry about that. For one, be able to drag a highlighted section of a track right or left past the window edge, so that it scrolls as you drag. Get that moveable vertical bar in so that if I am at a position in track, it’s reflected by the vertical bar position in the time scale bar at the top of the screen. Which is a feature request now anyway. And, be a little more stable ? I know that sounds odd, but it’s hovering right on the edge of being very unstable. So I’m not pushing it. It’s a great program though, and admirable for what it does. I love it.
Anyway, right off track here :slight_smile:

I now know why narrators charge so much !!!