• Welcome to ASR. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

What effects, and in what order, do I apply to my recorded voiceover?

ruidito

ruidito

Member
Joined
Jun 5, 2024
Messages
61
Likes
12
Location
Argentina
A few days ago, I asked for your help through a post, requesting simple tips on how to record my voice and then overlay it onto an existing audio track.
You recommended programs and microphones, gave me advice, taught me many things, and I'm very grateful.

So, from all the wise advice you gave me, I chose to begin my rookie journey using Audacity, since it's compatible with Windows and macOS.

I'd like to ask you a simple question: in what order and what effects should I apply to the recorded voice before doing the voiceover with the existing audio track?

I know I need to compress, remove background noise, etc. So, what effects/adjustments, and in what order, should I apply them to my recorded voice before the voiceover?

Please give me simple, to-the-point, and very basic answers; remember, I'm just starting out.

Thank you so much for this wonderful space.
 
Good question as there is a correct order to this.

0) normalize
1) noise removal
2) EQ, as needed
3) Compression (go easy on ratio, but also don't be scared to turn up the threshold so it stays active.)
4) Reverb (if appropriate)

I recommend finding a voice over clip that sounds really good to you, and comparing your clip to that step by step to guide your application of effects.

You might also consider using a DAW like Reaper, FL Studio, Logic, Ableton, etc for this process. They work in real time, and come with more numerous and more flexible effects, so tweaking and undoing the effects is much faster. If you plan on doing this kind of work a lot, it's a worthwhile step up.
 
Ideally, a good recording shouldn't need any "effects". ;)

In the real world, most pro recordings have some EQ, compression/limiting, plus reverb if it's music. Those are the "big-3" effects. Generally it's all done by-ear unless you have a loudness target and then you might need to measure LUFS while applying limiting and compression.
It's generally a good idea to normalize ("maximize"). That's just a linear volume adjustment so It doesn't change the sound and I don't consider it to be an "effect". If you are using other effects or processing, normalize last.* I'm talking about regular (peak) normalization. Loudness normalization is trickier and it can potentially push the peaks into Clipping (distortion).

You can EQ by-ear depending on how your voice and microphone sound. And compare to a known-good reference. Reference recordings help to "keep your ears calibrated".

If you need Noise Reduction, run it before making certain parts of the recording louder or quieter, changing the background noise. Noise reduction works best when you have a constant low-level background noise... When you don't really need it. If the noise is bad you can get artifacts/side effects and "the cure can be worse than the disease". It's something you have to experiment with.

Limiting is a "fast kind" of compression and with make-up gain (or normalizing after) it can make your voice sound stronger-louder. In most cases, it works better than regular compression and there are fewer settings to mess-with and mess-up. Audacity's limiter is very good. It uses look-ahead so it doesn't distort the waveform. Note that since compression and limiting are normally used with make-up gain, they tend to bring-up the background noise (the signal-to-noise ratio is reduced).

Assuming you have a good microphone and audio interface (or a good USB "podcast mic) acoustic noise is usually the biggest difference between a home recording and recording in a soundproof studio. A lot of people make audiobooks or other voice-overs in "home studios" but noise is usually a battle and you might have to turn-off your refrigerator, HVAC, and anything else that makes noise, and record at quiet times during the day (or night) etc. And some noise reduction is still often needed.

and then overlay it onto an existing audio track.
Click to expand...
Mixing is done by summation (analog mixers are built-around summing amplifiers) so you have to lower the track volumes and watch out for clipping. If you mix two (or more) 0dB normalized tracks you'll get distortion. Audacity doesn't make mixing "easy" because it doesn't have a master volume control. One solution is to export as floating-point WAVE (which has no upper or lower limits). Then open the stereo (or mono) floating-point mix and normalize it before exporting again to your desired format.



* Except when you are making an ACX audiobook there are specific requirements for peak and RMS levels, so limiting is usually the last step. Note that ACX will reject you for "overprocessing" if the recording is "too quiet". The maximum noise level is -60dBFS but they don't publish the "too quiet" spec.
 
Last edited:
What I would do:

I wouldn't bother with noise reduction, instead focusing on recording clean vocals in a quiet environment with good signal-to-noise ratio (you want your dB meter to be "in the middle" near -12dB at the loudest: not near the bottom at -100dB where the noise is, not near the top at 0dB where the distortion is).

Once the voiceover is recorded, I would then apply a high-pass filter that cuts anything from 80Hz and below.

Here there are two philosophies: compression into eq and eq into compression. I lean towards compression into eq because the compression is going to squash the eq curve you set up so it's more difficult to keep track of what's happening.

Compress just enough so that you don't feel the need to raise the volume when the voice is quiet and lower it when it's too loud.

For eq, if the voice is muddy or boomy you can cut the lows and boost the mids (2.5kHz is where clarity in vocals is perceived the most). If the voice is too shouty or bright, you can apply the opposite curve, boosting the lows and cutting the mids. There's also a sort of "radio host" eq that boosts lows and highs while scooping the mids if you're into that (I wouldn't bother with it).

Optionally a touch of reverb if you feel that the voice is too dry or unnatural.

I always have a hard time doing this with Audacity because it doesn't have real-time metering for its included effects, but it's possible. Maybe others have more experience with the software and can help you get where you need.

Watch out for clipping while you record or apply processing. You don't want to hit 0dB (which means the signal is too loud). To achieve that, lower the input (track) volume. You shouldn't touch the playback volume in the app.
 
I would recommend automatic algorithms.
First RX11 - repair assistant (accept noise removal; play with deess on/off because of the next step)
than smart-deess (incredibly good - probably just accept and go)
than smart-EQ (accept the suggestion, play with intensity)
than smart-compress (accept the suggestion, play with intensity)
optionally smart-limit at the end

it will sound better than 90% that is out there.
 
ruidito said:
I know I need to compress, remove background noise, etc. So, what effects/adjustments, and in what order, should I apply them to my recorded voice before the voiceover?
Click to expand...

EQ and compression, that’s all. Noise reduction shouldn't be needed in a decent recording chain, unless where talking about environmental noise (e.g. someone closinga door in the background). Forget about noise gates, that’s where a lot of rookies go wrong.

(Recording voice overs for trailers, commercials and documentaries was one of my main jobs for 10 years).
 
kemmler3D said:
Good question as there is a correct order to this.

0) normalize
1) noise removal
2) EQ, as needed
3) Compression (go easy on ratio, but also don't be scared to turn up the threshold so it stays active.)
4) Reverb (if appropriate)

I recommend finding a voice over clip that sounds really good to you, and comparing your clip to that step by step to guide your application of effects.

You might also consider using a DAW like Reaper, FL Studio, Logic, Ableton, etc for this process. They work in real time, and come with more numerous and more flexible effects, so tweaking and undoing the effects is much faster. If you plan on doing this kind of work a lot, it's a worthwhile step up.
Click to expand...
Thank you so much for taking the time to give me so many details.
 
DVDdoug said:
Mixing is done by summation (analog mixers are built-around summing amplifiers) so you have to lower the track volumes and watch out for clipping. If you mix two (or more) 0dB normalized tracks you'll get distortion. Audacity doesn't make mixing "easy" because it doesn't have a master volume control. One solution is to export as floating-point WAVE (which has no upper or lower limits). Then open the stereo (or mono) floating-point mix and normalize it before exporting again to your desired format.
Click to expand...
Thanks a lot!
I didn't understand this, and I realize it's important.
 
tccalvin said:
What I would do:

I wouldn't bother with noise reduction, instead focusing on recording clean vocals in a quiet environment with good signal-to-noise ratio (you want your dB meter to be "in the middle" near -12dB at the loudest: not near the bottom at -100dB where the noise is, not near the top at 0dB where the distortion is).

Once the voiceover is recorded, I would then apply a high-pass filter that cuts anything from 80Hz and below.

Here there are two philosophies: compression into eq and eq into compression. I lean towards compression into eq because the compression is going to squash the eq curve you set up so it's more difficult to keep track of what's happening.

Compress just enough so that you don't feel the need to raise the volume when the voice is quiet and lower it when it's too loud.

For eq, if the voice is muddy or boomy you can cut the lows and boost the mids (2.5kHz is where clarity in vocals is perceived the most). If the voice is too shouty or bright, you can apply the opposite curve, boosting the lows and cutting the mids. There's also a sort of "radio host" eq that boosts lows and highs while scooping the mids if you're into that (I wouldn't bother with it).

Optionally a touch of reverb if you feel that the voice is too dry or unnatural.

I always have a hard time doing this with Audacity because it doesn't have real-time metering for its included effects, but it's possible. Maybe others have more experience with the software and can help you get where you need.

Watch out for clipping while you record or apply processing. You don't want to hit 0dB (which means the signal is too loud). To achieve that, lower the input (track) volume. You shouldn't touch the playback volume in the app.
Click to expand...
Thank you so much. I find similarities between what you're telling me and research I've done previously.
 
dasdoing said:
I would recommend automatic algorithms.
First RX11 - repair assistant (accept noise removal; play with deess on/off because of the next step)
than smart-deess (incredibly good - probably just accept and go)
than smart-EQ (accept the suggestion, play with intensity)
than smart-compress (accept the suggestion, play with intensity)
optionally smart-limit at the end

it will sound better than 90% that is out there.
Click to expand...
Thank you very much. Where do I get these automated algorithms and how do I apply them?
 
Geert said:
EQ and compression, that’s all. Noise reduction shouldn't be needed in a decent recording chain, unless where talking about environmental noise (e.g. someone closinga door in the background). Forget about noise gates, that’s where a lot of rookies go wrong.

(Recording voice overs for trailers, commercials and documentaries was one of my main jobs for 10 years).
Click to expand...
Thanks a lot.
 
One other tip I've found useful when editing VO content in Audacity is to normalize the audio piece by piece, like individual phrases or sentences. This levels out the volume well without losing natural dynamics on a short time scale. It makes it a lot easier to get a good result when you add compression and you may not need to use as much.
 
kemmler3D said:
One other tip I've found useful when editing VO content in Audacity is to normalize the audio piece by piece, like individual phrases or sentences. This levels out the volume well without losing natural dynamics on a short time scale. It makes it a lot easier to get a good result when you add compression and you may not need to use as much.
Click to expand...
I understand; but if they are long texts, wouldn't it take a lot of time to do it that way?
 
ruidito said:
I understand; but if they are long texts, wouldn't it take a lot of time to do it that way?
Click to expand...
Yes. It's time consuming, maybe not practical if you have more than a few minutes of audio.

In that case you can rely on a compressor with slow release time and gentle ratio to even out dynamics between words.
 
kemmler3D said:
Yes. It's time consuming, maybe not practical if you have more than a few minutes of audio.

In that case you can rely on a compressor with slow release time and gentle ratio to even out dynamics between words.
Click to expand...
Thanks.
 
Some voice over software and other effects. 41 programs $1500 worth for $20 and helps charity. FYI Ends in six days.
www.humblebundle.com

Audio Alchemy: A Premium Collection of Sound FX, Music, and Game Engine Plugins from Ovani Sound Encore

Download a library of royalty and subscription-free sounds for your game, film, or creative project. Pay what you want and support charity!
www.humblebundle.com www.humblebundle.com

All Bundles

Our Bundles are limited-time collections of games, books, software, and more. Pay what you want and choose where your money goes, including to charity.
www.humblebundle.com www.humblebundle.com
 
You must log in or register to reply here.

Similar threads

ruidito
At what point should I apply the effects to the recorded voice?
Replies
0
Views
317
ruidito
ruidito
ruidito
I need simple tips for recording and editing my voice
2 3
Replies
53
Views
2K
ruidito
ruidito
C
Looking for a Good System-Wide EQ for macOS (Apple Silicon)
2
Replies
39
Views
2K
ChaoscripT
C
E
EV RE-20 Sounds Muffled/Muddy to me - Please Help
Replies
8
Views
1K
Mr. Swordfish
Mr. Swordfish
A
Separate DAC + Amp vs Wiim Amp Pro // macOS Volume Control Confusion
Replies
1
Views
544
v-kooistra
V
Back
Top Bottom