Remove noise from audio files with Audition
Learn how to use Audition to keep the sounds you want and remove the ones you don’t from your audio and video projects.
Published
Find the noise in your audio files
In this video, you’ll review the typical noise scenarios you will encounter when cleaning up your clips.
Hi, my name is Jason Levine.
And I'm the worldwide product evangelist for Adobe's Creative Cloud and our audio and video tools.
And in this tutorial I'm going to talk to you about how to use Adobe Audition CC to remove noise from a variety of different files and also take you through exactly which noise reduction effects you should use, because we have several of them.
So you've got lots of different variations, lots of flexibility there.
I'm going to try and take you through some of the basics.
So before we get started,
Add audio files
we're going to open up some of our files.
You can see that I'm using the Media Browser inside of Audition CC.
And located on my desktop I have a folder called AACC_NoiseReduction.
And this will contain the 4 files that we'll be working with.
So I can simply drag-select across these and drag them in to my Files panel.
We're going to double click on the first file here:
Review noise types
Interview_Horn_BGNoise.wav.
So if we take a quick listen to this, you're going to hear a horn for one thing, a car horn, and you'll also hear some background noise.
So take a quick listen.
Ok.
So you can hear that there was a horn in there.
Now again for this particular file, you can also hear that there's some background noise, but typically before you start changing or trying to reduce background noise, if there are additional sounds like that horn that you just heard - and by the way, you can see I've also created a marker for you, so you can find it a bit more easily.
We want to tackle this first.
We want to get rid of that first.
And to do that we're going to use a new effect in Adobe Audition CC known as the Sound Remover.
Now if I go to the second file here: Interview_Bell_BackgroundNoise.wav, this is sourced from the same interview, but there's again another noise.
This time it's actually over top of the actual speaking.
It sounds like a... like a bell either from a cable car or something like that in San Francisco.
Take a quick listen.
So you can hear that chime in there.
So again, a little more difficult now.
And standard noise reduction techniques probably wouldn't do it, but the Sound Remover in Audition CC is going to give us that much more flexibility to try and tackle that and really kind of preserve that voiceover without destroying the overall content.
Now in this next file interview 3, here again different interview, background noise, but the background noise changes slightly.
And while you may not hear the change, you will actually be able to better see the change in something called our Spectral Frequency Display.
Spectral frequency display
That's this button up at the top here.
We're going to focus a lot on this for our noise reduction.
What you're basically seeing in this view here is frequency over time.
You have frequency on your vertical axis, time along the horizontal and color is amplitude.
So the closer the color is to a brightish yellow, the louder the amplitude at that particular frequency.
The darker it is, the closer it is to purple or black and all the subsequent gradients leading to black, the quieter it is at that particular frequency.
So you can see that we have some talking and some things here.
And then we have some sections that are just noise.
And here.
But again even as you listen and look at these, you can see that there's a slight variation in the color.
And for this we're probably going to use our adaptive noise reduction.
This would probably be a better effect to really tackle this, because the background noise changes thus allowing Audition CC to adapt to those changes without the need to take a noise profile which I'll talk about in the next episode.
And then the very last example here.
This SqueakyDoor_HealPaintbrush.wav file is an example of how to use some of our Photoshop CC style tools in Audition CC to literally paint out sounds that you don't want.
So this is the sound of a metallic door opening and you'll hear a nice little squeak in there.
Take a listen.
And if you look, the squeak is right in this section here.
So once again I'm going to talk about how to use now the Spectral Paintbrush to actually identify that and paint it out and use a couple of different healing methods to repair that sound as well.
So now that we've identified the noises, we're ready to actually start removing them.
So this was Part 1 'Getting started; identifying different types of noise'.
We'll see you in Part 2, where we actually start to tackle removing sounds and removing some background noise.
Learn to remove rings and beeps
Sometimes you will want to remove specific, isolated sounds.
In this video, you will learn how to use the Paintbrush selection tool to select the unwanted noise, create a noise profile of the unwanted sound, and apply the Sound Remover effect to remove it.
Hello and welcome to Part 2 of 'Removing noise with Adobe Audition CC'.
For this first file we are going to tackle using the Sound Remover to remove a specific sound, in this case a car horn, that occurs right before the talent begins speaking.
Now as I mentioned in the first episode, you notice that I've even created a little marker section here to show you visually where the horn is in time.
So if I just click my cursor a few seconds before and hit Play on this, you can hear if I make a selection inside the marker.
That's where the horn exists.
Now the problem is: Yes, we could edit this out, we could simply cut this.
It extends a little bit beyond into his voice there, but that might sound a little strange.
And again if this is tied to video we don't want to do that, because that would throw off syncs.
So we don't want to change the duration here, we simply want to remove that sound.
Well, you wouldn't normally do that here in the Waveform or the Editor view.
For that we need the Spectral Frequency Display - which is mentioned in the first episode - shows you frequency over time, where color is amplitude.
And now that we're looking at the colors you can actually see - and if I just choose my Marquee tool here -, here is basically where that horn lives.
In those horizontal lines that you see there, that's the actual horn.
So take a listen.
Ok.
You can hear that much better now.
Now we basically want to make selections of each of those specific harmonics or frequencies of the horn.
So to do that we're going to use the Paintbrush Selection Tool.
Paintbrush selection tool
Now if you want to adjust your paintbrush size, you can use the bracket keys on your keyboard.
So I'm going to make an adjustment on my paintbrush size here.
And I'm simply going to start to paint over the frequencies of that horn.
There's the first one.
And if I hold down my Shift key and keep it held down, now I can select all of the harmonics of that horn.
So I'm going to do that very quickly just like this.
Ok.
Looks like we've got a few more here.
There's another.
Maybe one more here.
And you can see that as we go up into frequencies as the frequencies get higher, we're seeing less and less of that horn.
Ok.
So let's take a listen to this.
Sounds very good.
Ok.
So at this point,
Learn sound model
all we have to do is right-click and choose to Learn Sound Model.
And as you'll see here this tells you that the audio selection is going to be captured and loaded as a sound model when we open up the Sound Removal Effect.
So we're going to click OK on that.
Let's go ahead and deselect - if you're on the Mac that's Command + D - and I'm going to go back to my time selection tool here.
I'm just going to select a few seconds before and just a few seconds after.
So this is the selection that I'm effectively going to process.
Now I'll go up to Effects,
Sound Remover effect
Noise Reduction / Restoration, Sound Remover (process)... and just using the default settings, we'll take a listen.
Now what you're seeing here this is known as the Preview Editor, where we basically have the before on the top.
And you can see all those horizontal lines of the horn.
And here's the after down below.
And you'll see that you have actually independent zoom control here which is why the views aren't exact.
We can actually turn these to Mirrored; so you're seeing the exact same thing.
And already you can see that the noise is gone.
But let's just take a listen again using the default setting here.
Now here was the before.
Let's turn it off.
And here's the after.
Pretty awesome.
Now again you've got lots of presets here that you can try.
Sometimes for things like car horns.
Remove Long Tones is a good one.
You can see that there's various presets to remove, sort of, you know, some of those sounds, some of the glottal sounds that you might experience when recording voice over.
Basically these are just here to get you started.
And it's a great idea always to play around with the parameters to really just experience what each one of these things does.
You'll also see that we have a checkbox here: Enhance for Speech.
Because this is dialogue, I will always keep this checked.
And it really tries to minimize any kind of artifacts to speech content when removing noises.
So I can simply now click Apply.
It applies the effect, wind it back, take a listen.
And the horn has gone.
Don't worry about this little dropout right there that just existed.
Because when we actually remove the background noise content that's going to go away.
We're not going to have to worry about that.
So again very quickly now we were able to identify where the horn was, make multiple selections using the Paintbrush and the Shift key, go into the Sound Remover, learn the sound model, go into the Sound Remover and simply take it out.
So that's using the Sound Remover tool to remove specific sounds from your audio.
Now in the next example I'm going to show you how to remove even more complex sounds when they're over top of a voice.
So not just before, but actually at the same time as the speaking voice.
So stay tuned for Part 3.
Get rid of unwanted sounds that are mixed into your main audio
A more difficult scenario is when you need to remove a noise that occurs at the same time in the clip as sound that you want to keep in your audio file.
In this video, you will learn how to use the Paintbrush selection tool to select the unwanted noise, create another noise profile, and again use the Sound Remover effect to remove a bell that rings at the same time that someone is talking. You will also learn how to remove ambient noise by using the marquee selection tool to select the background noise, capturing a noise print, and using the Noise Reduction effect to remove it.
Welcome to Part 3 of 'Removing noise with Adobe Auditon CC'.
Now, in the last episode we talked about using the Sound Remover tool on a sound that was not in conflict for the voice-over with another sound that you're trying to preserve.
In this case now we actually have the bell which sounds like a street car, which is directly over top of the voice itself.
And once again you can see that we're on our Spectral Frequency Display so that we can visually identify those frequencies and where they live.
And if you take a look at the marker section that I made for you here that says "Bell", I'm just going to click my time selection tool
Remove intermixed noise
just before it, and play back and take a listen.
Now, honestly, that's probably a really challenging thing to remove.
And in many cases, you might just have to deal with it.
Maybe you'd have the opportunity to re-record this.
But if not, we can certainly try the Sound Remover tool which can really do some amazing things.
So once again I'm going to choose my Paintbrush selection tool.
If I want to make adjustments to the size, I can use my bracket keys.
And what I'm going to do - again, I'm looking for those horizontal lines -, I'm going to start to draw some horizontal lines, like this, and I can take a listen.
And you can already start to hear the bell sound.
So, once again, I'll hold down my Shift key, so I can make multiple selections.
And you can see, I'm just selecting those horizontal lines that represent the harmonics of that bell.
I'm only going to choose the ones that I can see.
Remember that it probably has upper harmonics that are in much higher frequencies, but I can't see them.
And we don't really hear them here.
And if we just take a quick listen to what I've done so far.
That sounds pretty good.
So, now that we have this, once again I will right click, Learn Sound Model, deselect, T for Time selection tool.
Let's just select this area here.
Noise Reduction/Restoration, Sound Remover (process)...
And let's just take a quick listen.
Now, while it's not entirely gone, here's the before: And here's the after: Significantly attenuated.
And now again, once in the context of an actual documentary, maybe with some music or underscore underneath, you'll never know that it was ever there.
So I can simply click Apply.
It's gone.
Take a quick listen.
And just like that, we've taken it out.
Now, at this point of course we also want to tackle
Remove background noise
removing some of that background noise.
So a good way to do that is to find a section that has just noise, and take a profile of that, so I can choose something like this section here.
And I can now choose my Marquee selection tool.
And I'm simply going to select an area, like this.
Now again, I can extend this a little bit higher.
Now you don't want to select too far into the black, because if it is basically all black, that's telling you that there really isn't any noise there.
That's the brilliance of this effect, that you can choose the specific area where the noise lives, so you only remove areas where there are noise.
If you try and take noise out of something where it doesn't exist, you end up with artefacts.
So now that I've made my selection, I can right click and I can choose Capture Noise Print.
And this will tell you that it's being captured and loaded into the noise reduction effect.
So I'm going to click OK, deselect, because we want to apply this to everything, go to Noise Reduction / Restoration, Noise Reduction (process)...
Now there's lots of parameters in here.
I'm going to basically let you stick with the defaults, with a Spectral Decay Rate of 65%, your Precision Factor maybe 9 or 7.
Again, we're just going to leave these at default settings.
And the only one we're really going to worry about here is this Noise Reduction slider.
So let's turn it all the way down, - and start playing back.
And while I'm doing that, I'm going to increase - the amount of noise reduction.
So again, here is before: And here is the after.
I'll click it in now.
So, click Apply, take the noise out, play it back.
And just like that we're able to remove noise with the noise reduction effect.
Now I'm gonna take this a little bit further in the next episode and also talk about adaptive noise reduction techniques.
But as you can see, it's as simple as finding where noise lives, taking a profile, applying some default settings, and very quickly and easily attenuating or softening some of that noise to really bring out the content underneath.
Remove all annoying background noise in one step
Sometimes you will have video or audio files where the background noise is uneven or inconsistent. You could use the technique you learned in the previous video to remove each piece of inconsistent noise in several separate steps. However, Audition has a feature that will allow you to remove all the inconsistent noise in one step.
In this video, you will use the Adaptive Noise Reduction effect to analyze the audio in the clip and adapt to changes in the noise over time. This will allow you to remove all the noise in one step. You will also review the steps you learned in the previous video using the Noise Reduction effect.
Hello and welcome back.
This is Part 4 of 'Removing noise with Adobe Audition CC'.
Now in the previous episodes I showed you how to use the Sound Remover.
And I also gave you a brief look at using the actual noise reduction effect, where basically you visually identified noise, took a profile or captured a profile and then simply removed it all at one time.
Now that's fine if the background noise is consistent.
But in this particular file - we're looking at file 03 Interview_AdaptiveNR.wav -, what you can actually see now if you're starting to look at the Spectral Frequency Display a bit more, you'll start to see and start to understand a bit more of what's going on.
And basically we've got noise.
Well, it's throughout the file, but where there's no talking is here and here.
And if you just take a look at these sections you can see that sort of the reddish orange area extends to just under 2 kHz in this section.
But here it actually goes a bit higher, it's almost about 2.5, maybe even as high as, yeah, I guess around 2.5, 2.7 kHz.
And then again this sort of upper mid-range noise; it just visually it looks different.
And if we take a listen...
There seems to be just more sort of low-end, muddy sounding noise in this section.
And it's slightly brighter, just slightly different in this section.
So this is a perfect opportunity to use the noise reduction effect known as Adaptive.
Because this will essentially adapt to the changes in noise over time.
And this also means that you don't have to take a profile at all.
It does it automatically.
So I can go up to my Effects menu,
Adaptive Noise Reduction effect
Noise Reduction / Restoration, Adaptive Noise Reduction...
And once again let's just go ahead and go into our (Default) settings here.
And let's go ahead.
And start with it on and click Play.
Now you're going to hear it actually kick in, because it has to kind of figure it out.
It takes about a second or a second and a half for it to figure out what it wants to do.
And then you'll hear the noise kick in.
So take a listen.
So here it is with it off.
And then we'll turn it on.
Now I created one myself which I called InterviewNR.
Adjust reduction parameters
And I did that just by clicking on this Save settings as a preset.
And simply by adjusting the reduction amount, a little bit of my noise floor and then a bit more of the spectral decay.
The more Spectral Decay Rate that you apply, you're just going to have smoother fall-off, smoother reduction, a bit less artifacted, preserving a bit more of that background noise.
So take a quick listen to this.
Apply the effect
Go ahead and click Apply.
And again just like that it attenuates the noise.
Now you can see there's still some here, right?
But again it's not that bright orange color that it was a moment ago.
It's a softer reddish orange.
And you can already see visually that it just looks cleaner.
And it obviously sounds a lot cleaner.
Now again if you want to go back to that first file, number 1, that we worked on at the very beginning, again this one has lots of very consistent noise throughout.
And in this case I could actually take a profile.
Something like this: Ok.
And you can see that goes up to just around 7k.
I can right-click, choose Capture Noise Print.
And now I could go into something like...
Oh, let's deselect here as well.
Always important to deselect.
Go into Noise Reduction (process)...
And by the way, if you don't deselect, you'll see that we do have a Select Entire File option; basically the same thing.
And once again now we can play around with the settings here.
So I'll start with the Noise Reduction off and then kick it into gear.
Ok.
Click Apply.
And just like that.
Now this is taking out the lion's share of the background noise.
And remember we had a little sort of dropout happening here.
Well, now when you take a listen:...
And now it just sounds a lot better, a lot cleaner.
So again you can really use either one of these effects.
It really just depends on the situation.
Sometimes Adaptive is just a quick and easy way to really attenuate some noise very quickly.
And if you have the opportunity to take an actual profile and apply it, typically that works a little bit better, maybe a bit smoother, with a few less artifacts.
But you have a lot of options in Adobe Audition CC.
And that's really the key.
So this is was Part 4 of 'Leveraging noise reduction'; in this case a combination of Adaptive Noise Reduction and our standard noise reduction effect in Adobe Audition CC.
“Heal” your files automatically and export them
In this video, you will use both the Marquee selection tool and the Paintbrush selection tool to select a noise, and then use the Auto Heal feature to attenuate the noise. You will also learn the different settings to export your clip correctly in the proper format.
Now most of the noise should be removed from your audio.
Hello and welcome to Part 5 of 'Removing noise with Adobe Audition CC'.
In this final episode in the series I'm going to show you how to now use the Photoshop-style Paintbrush tool to heal your audio, basically identifying where a sound is, and healing it out or attenuating or deleting it out.
And then I'll show you how to actually export or save out these files.
So now that you've removed all of your noise in your files, how do we get them back out and how do we share them with people?
So the first thing is this file here: SqueakyDoor_HealPaintbrush.wav.
And if you take a listen as I mentioned in the very first episode, this is a big metallic door that is being opened.
And you'll hear a squeaky sound.
Take a listen.
Now since you've been looking at the Spectral Frequency Display for the last four episodes, you probably have a good idea of where it is.
Lasso selection tool
Now if I go ahead and grab my Lasso selection tool, if you were guessing that it was somewhere right around here, you were absolutely correct.
So I'm going to make a selection around this.
And you can see it actually looks like a little dachshund puppy right there.
If I go ahead and play this back, you can hear exactly the frequencies of just that squeak.
Now if we wanted to go very aggressively to take this out, I could simply delete that.
And when I do that you'll see that it actually leaves a hole.
But because it's confined to a very small frequency range, when I play it back... you don't really hear anything missing, except that the squeak is gone.
And that's awesome.
And sometimes that's just the way to do it.
But if you want to be a little more careful and do it a little more Photoshop-style with a bit more finesse and control... first thing I'm going to do is I'm going to zoom in vertically which I'm just using a gesture on my Magic Mouse here.
I'm going to take my Paintbrush selection tool.
And once again I can adjust my brush size by using the bracket keys.
And I'm simply going to paint like this on that squeak.
But instead of deleting,
Auto Heal effect
I'm simply going to go up to my Favorites menu and choose Auto Heal.
Now you might have to do this a couple of times, because this is not going to leave a hole like in the other example.
This is going to follow a much more Photoshop-style technique of removal, where it's going to simply analyze all the pixels or the audio before and after and basically smooth it out.
So take a listen now.
So there's just the faintest hint of it still in there.
And again we could make another selection on this if we wanted.
And at this point maybe we could do a slight attenuation, maybe something like 9 dB.
Just a little bit to just again bring that down a bit more, wind it back.
And now it's effectively gone.
There's no hole; it's great.
And we've now saved the day.
So again just another amazing noise reduction technique in Adobe Audition CC.
So now that we've made all these changes,
Export the files
how do we get these files out if we want to share them back to our editor or perhaps put them into our favorite audio program, whether it's iTunes or Windows Media Player or what have you.
So if I go up to the File menu, you can either go to Save As... - which is what I will do 99% of the time - or you also have the option to go to File..., Export File, which just uses the word "file" too many times for me, so I don't typically go here.
But effectively they're doing the same thing.
So if we choose that, you give it a name, you give it a location and then you choose the Format.
Now if you want to keep this, what is known as "uncompressed", i.e. the highest quality possible without loss, you're going to save back to the default, which is Wave PCM.
This is what everybody uses as a standard.
And again if you want to keep it uncompressed, choose Wave PCM.
We also have the very popular AIFF.
Again you see this a lot with other digital audio workstation applications, you can save out to AIFF.
If you're going to something like iTunes or you're wanting to listen to this on a device, you don't want to leave these uncompressed.
These files are going to be a little bit too big.
So in that case we would choose something like MP3 Audio.
And when you do that, then of course you have all the options to modify your settings.
Now because this came from video, these are 48K, 48,000 Hertz.
That's not the iTunes CD standard.
The standard is 44100 Hz.
So here's where again you could click on Sample Type, Change...
And this is where you can modify things like Sample Rate.
You can change the channel configuration.
And then you can convert down from 32 to 16 bits. 16-bit is what you want for iTunes, for YouTube, for Vimeo, all of your audio coming from video or just audio alone, whether it's going to SoundCloud should be 44100, 16-bit stereo.
Now in terms of MP3, here's where you can change the bitrate.
And again: the higher the bitrate, the better the quality. 320 kbps is what I will save out all of my MP3s that I bring into iTunes or share with anyone.
So this is what you'll want to do too.
But again: you can place or choose any particular bitrate.
You'll see that you also have options here for Constant CBR or Variable VBR.
So again 320 kbps is going to give you the best quality MP3.
And you'll be able to play these literally on any device, on any machine, on any system.
So again when you're exporting, if you want to keep things uncompressed in high quality: Wave PCM.
If you've been requested to use AIFF files, you can choose that here.
You'll also see that we have other popular sharing formats, like Monkey's Audio, OGG and FLAC.
These are also part of Audition CC.
And then lots of various other sort of more esoteric sound file types.
But your standard ones are Wave PCM, MP3 and AIFF.
Again File, Export File, click OK.
And it will save that out.
If you've placed any markers or meta data inside the file that you want to preserve you want to keep that box checked.
That's also essential for MP3s, because that will contain any artist, album title, year information that you placed in the meta data as well.
So that's essentially how you use the Paintbrush.
And then again in a combination of leveraging all the various noise reduction tools that I showed you whether the standard noise reduction effect, the Adaptive Noise Reduction or the Sound Remover, you can very quickly remove or attenuate your noise and save your files back out to an uncompressed or to a lossy file format to share with others very quickly with Audition CC.
Now don't forget there's lots of other opportunities to learn more about Audition CC.
You can go to the Creative Cloud and find more training on Adobe Audition CC by yours truly and some of my colleagues.
So we invite you to check that out.
And we hope to see you again next time.
Contributor
