Removing breathing sounds

[MrDollinger]

Hey all,

I'm kinda thinking ahead with this situation to try and both get the best out of the equipment that I have already, and trying to find the best solution. I'm aware I'm possibly dealing with a less-than-ideal setup, but it's what I have and I don't necessarily have huge amounts of money to sink into more equipment, plus I'm trying to keep this as compact and portable as possible.

The scenario: I've been doing some camera tests for a possible project I have coming up. It's just an Android device I have that I've got hooked up to a Boya BY-A7H. This seems to do a reasonable job of picking up audio from me, plus people within a reasonable distance, though of course there is a volume disparity between them and myself.

I can get around this to a reasonable level by some careful use of compression to try and make the volumes more even. There's only one thing that I have to contend with here - and that's my breathing. I've basically been one of those mouth breathers for most of my life and when I start applying make-up gain of course my breathing becomes very noticeable.

I'm planning a few different options to fix this. One is to try and move the whole camera setup a bit further away so the microphone picks up a bit less breath. There may be only so much I can do here though.

Option 2 is that I try and place an expander or noise gate before the compressor to try and cut down on the breathing. My concern here is that, with the whole point being to even out volumes between myself and somebody standing nearby who is quieter, this may start taking out audio I want to keep.

Option 3, and the one I'm thinking might have best results here, is a bit of noise reduction. With the video stuff, I'll be editing in Vegas Movie Studio, but I also have Adobe Audition. I'm thinking I could simply pass the audio over to Audition, use the tools there to try and reduce the noise of my breathing, and then bring the audio back in. Before I jump in and have a play with this, I'm wondering if someone has had to do something similar and whether they can offer any sage advice to help get the best results without introducing too much in the way of audio artifacts.

Obviously another solution would also be multiple lavalier mics or something similar going into an audio recorder. This would allow me to do the expander/gate trick on just my channel (which is what I do for my setup at home which solves it) but for reasons of cost and portability, this solution might not be achievable.

Edit: Well this is awkward .... just realised I've accidentally posted in the wrong forum and I can't find an easy way to delete/fix it. Hopefully someone notices and moves it to where it should be. Apologies .... this is what happens when you post after a long day

 

[Steenamaroo]

Hi,
As with many things there's the hard way and the easy way.
The easy way, as you point out, is to have multiple microphones so your microphone is only concerned with your voice.
That would, as you said, allow you to use a noise gate and not worry about anyone else.
You can make the noise gate's job a little easier by making a concerted effort to speak with clarity and consistency, avoiding mumbling or tailing off.
Basically talk as if there already is a noise gate, and the threshold is a little higher than you'd like.


In situations like this I'd avoid using compression to level out volumes.
Unless you're talking about hours and hours of recording, I'd go through manually and volume automate where needed.
Focus on bringing up the level of quieter passages (without the breaths) rather than bringing down the level of louder parts.

That way the breathing is made quieter relative to the spoken content, rather than being made louder.


Of course the fool-proof solution is to fix it at the source. That might be easier said than done but making a conscious effort to speak loudly and clearly and, if necessary turn from the mic a bit for large intakes of breath, or alter the shape of your mouth when you breath to reduce how audible it is is likely to make a huge difference.
It's not really unlike how professional singers 'work' the microphone, moving closer/farther for quieter or louder passages.
Ok, they're controlling a different thing but the similarity is in constant self-awareness.
Being calm and controlled can help a lot. The loudest breath sounds are usually the rushed snatched breaths.
If you allow yourself time for a nice controlled breath, it's likely to be much quieter.

I'd say
Plan A : Try to reduce the noise physically, if possible.
If not - Plan B : Try to get your own dedicated mic so you have more freedom to 'fix' the recording afterwards
If that's no good - Plan C : Try not to breath super loud while someone else is talking. At least that way reducing breaths in post wont affect someone else's dialog.

Hope that's helpful!

Edit : I moved this over to Vocal Technique forum for you. Hope that's OK!

 

[rob aylestone]

If you produce the breath noises, then the balance of wanted voice to unwanted noise will remain identical no matter the distance, because the noise comes from the same two sources - nose and mic. With this, I am not sure compressors work at all well, because they operate on volume, not the type of sound. The exception would be if you discover your noise is a very specific frequency and you could pop a sharp filter on that and use it as a side chain control of the compressor, so it ONLY worked on the noise. I suspect though, this won't work. What you could do is post a link to maybe 20 or 30 seconds of the voice, with the problem and people could try their usual solutions to see if any work? Sometimes these kinds of noises are unfixable, but other times, they can be worked on and cured or reduced. A good singing teacher or vocal coach would know the causes and be able to suggest solid solutions. I've seen it be things as simple as pulling your lower jaw 10mm back, which cchanges the shape of your throat which worked 100% for an old student of mine. Really simple and the pullback just opened up a passage that was wheezy and raspy in the relaxed state. She said it ached a lot until her muscles got used to it and her singing really changed.

 

[MrDollinger]

A good singing teacher or vocal coach would know the causes and be able to suggest solid solutions.

Just to be clear, this isn't so much from singing - this is more from a video production standpoint. (While I know this forum focuses more on music production, I would think there's at least some overlap in how to do deal with this from a technical standpoint).

It looks like @Steenamaroo has a few good tips as well, but I might look at seeing if I can extract some audio tonight from some of my test footage and try and post it to see if I can get something a little more specific to what I'm doing.

 

[keith.rogers]

Yes, use automation to surgically decrease the breath sounds, and then bring up the other voices just with gain or fader automation. That's not really a job for make-up gain (IMO).

There are tools, like iZotope's RX, that can process some kinds of breath noises fairly well. I don't know how the "basic" version of RX does with that, but it might be worth a try. (It's not always perfect, but boy can it do some amazing cleanups.)

If you plan to do more of this, you should really consider multiple mics and something like a Zoom F6/8n+, or similar.

 

[rob aylestone]

I know it’s not about singing, but they are very good at working out what causes sounds they need to reduce. Either that or an EN and T specialist. What I mean is that if you have a noise component, rather than try to remove it electronically, which will also impact your voice, maybe the cause can be helped. Processing is a paraceletmol, when you need to stop getting the headache.

 

[grimtraveller]

Processing is a paraceletmol, when you need to stop getting the headache.

On the other hand, if the processing achieves the required result, the headache isn't such a big deal.
Which is why I don't think it's an appropriate analogy.

 

[markmann]

I totally agree with everything that @Steenamaroo said. I am no expert at audio for video (I have done some in the past) but I have seen lots of videos that actually cut out the breathing between phrases. Actually chop the audio and video, especially commercials where I assume some of this is done to shorten the content.

 

[Ed Fones]

Should only be the slightest breathes in dialogue.

 

[Papanate]

Izotope RX9 Advanced will eliminate the breath noise cleanly without affecting the audio.

 

[Ed Fones]

Izotope RX9 Advanced will eliminate the breath noise cleanly without affecting the audio.

It's RX 10 now Papanate and the 'RX 10 standard' is the one for audio only.

 

[Papanate]

It's RX 10 now Papanate and the 'RX 10 standard' is the one for audio only.

And? RX9 Advanced is less expensive for the moment - what do you mean Standard is the one for Audio only?

 

[Ed Fones]

And? RX9 Advanced is less expensive for the moment - what do you mean Standard is the one for Audio only?

RX Elements and RX Standard are for audio recordings.

RX Advanced is for audio recordings film/outside audio. Removing wind, traffic etc. Repairing old sound etc.

I think. I use RX9 maybe 10 soon.

 

[Papanate]

RX Elements and RX Standard are for audio recordings.

RX Advanced is for audio recordings film/outside audio. Removing wind, traffic etc. Repairing old sound etc.

I think. I use RX9 maybe 10 soon.

RXX Advanced is for audio - it includes things you can also use on Film - but audio is audio no matter where you get it from - and the Advanced version doesn't edit video.

 

[Ed Fones]

RXX Advanced is for audio - it includes things you can also use on Film - but audio is audio no matter where you get it from - and the Advanced version doesn't edit video.

I didnt say it edits video. It is used to restore old film or even new film audio. It is used to remove traffic, wind and other noises from video audio.

RX9/10 Standard and Elements are used to remove clicks, and S'ss and other things like mouth noises and that from say vocal audio files recorded in a studio.

The Izotope website explains it better than me.

 

[Pinky]

My primary collaborator tends to have a lot of breathing between sections. If he's singing loudly I can use a gate to remove, since the vocals are louder than the breathing and easily removed by setting the threshold low. But if it's a softer song with quieter vocal passages, the gate is often chops into the leading edge of the lines being sung. For those songs I have to manually automate the volume between each sung section. I haven't found a good solution for the softer passages, and I tried a variety of plugins and settings.

Some trained (studio) vocalists will step away from the mic and do their breath intake before moving back to position to sing. Live performers pull the mic away and back to their mouth frequently, removing the breaths is a big reason why. If someone isn't trained to do this, like my collaborator, the onus for removing those breaths falls on the engineer/me.

 

[rob aylestone]

I just have to smile at the folk who put mics in strange places and spend ages treating the audio, or telling the singer the dreadful pitching was fine, we can sort it later - I've always believed in fixing faults at source, not repairing what they cause. Nothing will make me think that fixing faults with technology is better than fixing the problem. I know it works and I do it - but in this case, we have a nice voice in a noisy mouth - surely some rudimentary work to perhaps fix this is best? We can fix it for a recording, but what about if it has to be live? Remember Margaret Thatcher? She had a terrible weak voice when she first started out in politics, but by the time she got to be Prime Minister, it had been worked on really amazingly - she could fix the voice for recorded events, but not the live ones?

Pinky's comment on moving the mic requires amazing skill, and those flamboyant performers who do it really well make the engineers job much easier. Sadly, lots of singers copy their idols movements as if they were dance moves and engineers absolutely hate it - because they do this volume adjustment really, really badly.

I keep smiling because people all have these amazingly great repair tools nowadays, and some are mega expensive, too - I still maintain you should cure the problem at source if you can, using technology as the last resort. However, this, I appreciate is an old fashioned approach. There is no pressure put on the source to fix it any more, so we just get used to processing and tricks.

 

[grimtraveller]

people all have these amazingly great repair tools nowadays, and some are mega expensive, too - I still maintain you should cure the problem at source if you can, using technology as the last resort. However, this, I appreciate is an old fashioned approach. There is no pressure put on the source to fix it any more, so we just get used to processing and tricks.

Believe it or not, I wholeheartedly agree with this. It just so happens that there are also ways to fix it later and it's as well to know this. Personally, I aim for good clean tracks at a decent level so I don't have to perform any jiggery-pokery later. It happens sometimes, however, and one has nothing to lose by knowing how to deal with such a situation.

 

[Ed Fones]

Mic management or something like that it is called. Just putting a mic on a stand pointing at the vocalist and then walking straight up to it is going to create lots of editing after the recording. I think everybody is different as well and the mic is best placed in a position to suit the person. If the breaths, S'ss and mouth noises aren't there, then it is less editing work and far better for all. Placing the mic in a position where it doesnt pick them up so strongly and the vocalist practicing so not to make them is a blessing.

Editing can be over done as well. Removing breaths too much with the software causes problems to the vocals. The big breath at the beginning of the sentence can easily be cut out manually. The others in the speech can be lessened with the software, but gently does it.

 

[Papanate]

I didnt say it edits video. It is used to restore old film or even new film audio. It is used to remove traffic, wind and other noises from video audio.

RX9/10 Standard and Elements are used to remove clicks, and S'ss and other things like mouth noises and that from say vocal audio files recorded in a studio.

The Izotope website explains it better than me.

The Point is it's not designed just for Film audio - if you've ever done field recordings - the tools come in handy - at least for me.