What Settings Are You Using For Narration / VO Channel Strip


I’m constantly tweaking and changing my channel strip settings on my VO tracks and can’t seem to really dial in settings I’m happy with over a longer period of time. I know every voice is different, and especially in music production there’s lots of subtle tweaks and of course EQ to really massage the voice. I’m not talking about that.

I would like to get a sense of what you’re doing for a narration / commentary (e.g. for documentary) / VO for movie track.

For example, I did a learning video recently where I had a commentary (i.e. "now move your mouse here and click the OK button to save the changes you’ve made to your document). I would like the voice to sound steady. If I speak softer or louder, there should not be too much dynamics so people need to set their computer volume level up or down all the time. Or when I record a audiobook character. Sometimes they talk softly, sometimes they scream. I want to smooth out that huge change in dynamics.

1. How do you do your gain staging?
What is the average loudness you set your clips to to enter the first plugin in your chain?

2. What plugins / settings do you use?
I’m first using an EQ to cut off unnecessary sounds in order not to trip the compressor. LowCut essentially.
Noise Gate to get rid of unwanted sounds.

Then EQ to form the sound of the voice a bit and make it shiny.

Then De-Esser.

Then a Compressor with pretty fast settings to compress transients and when I make dynamic changes quickly (e.g. suddenly a char in a book starts to scream). I play with ratios between 2 and 6.

Then a limiter to cut anything that still comes through.

And although I like the channel strip of Nuendo for this sort of bread and butter functions my favorite go-to plugin is iZotope’s Alloy 2.

3. My problems with this setup:
Sometimes in documentaries people breathe, and those are often below the threshold of the noise gate. So when people breathe, suddenly the noise gate turns off and the breath gets a very ugly volume bump. But when I set the threshold down, I get a lot more noise through when they move.

When I record narration for a documentary or learning video, I don’t seem to get a nice punchy, transparent sound. Somehow I think the compression has a negative effect. Maybe my compressor settings are too aggressive. Or maybe I’m just not happy with the way my mic sounds and I should get a different one. I’m currently recording commentary with a ElectroVoice RE20.

What do you use?
So that’s why I’d like to hear what you do. If you have a way that worked well for you. Especially the input signal loudness to your plugins and then Noise Gate and Compressor settings would be of interest. Or do you use multiple compressors? One for quick and one for slow changes? I’d really appreciate your experience to get some new ideas what to try.

Audiobooks are a great example for this sort of thing.

Most Go-To Chain, DX Group:

Low Cut on the Nuendo Channel Settings; general EQ

UAD API Channel Strip… mostly just the preamp & EQ (final touches). No compression.

Nuendo’s general de-esser.

Nuendo’s compressor (around 2:1, fast release)

Nuendo’s expander (light!)

UAD Precision Limiter

But the most important thing, always, is to deal with breaths, noise reduction and level in editing first. No way around it if you want to avoid artifacts and obvious processing. Liberal use of room tone.

Sometimes I will throw in a real-time instance of RX noise reduction if there are semi-consistent low-freq ambient problems, such as heaters and air conditioners.

There’s my very general setup.

*** After seeing Fredo’s comment on de-essing being the result of bad recording-- welcome to my world. I typically get tracks from home studio sources that arrive in a VAST variety of conditions; no control over it before the fact. Hence the de-esser and expander! Job one is making it all appear as transparent as possible. An often soul-sucking endeavor in a 10+ hour project…but we do what we can!


For voice my chain usually resembles something like this. The settings differ from project to project of course but the tools I use don’t.

On the channel itself I use a De-esser, Compressor and EQ. At this stage, the de-esser is only used if needed. The Compressor isn’t doing more then 2-3 db of gain reduction but is does add a little volume bump. The EQ might be adding a little low end or cutting annoying mid-range frequencies or both. Depends on what is needed. The channel also has the Nuendo hi-pass engaged and set at around 50-80hz. Gotta be careful with the Nuendo hi-pass as it is extremely steep. Very easy to cut all the “oommph” out of something with those filters.

This channel is then routed to a group.

On the Group I have another compressor, usually the UAD Neve 33609 with a low ratio. (2:1 or 3:1)
And I try to keep it from doing more then 4 db of gain reduction. I also boost the signal by 2-4 db on the output of the 33609. This is followed by another EQ, usually the UAD Neve 1081. I usually end up adding a little air up around 16khz. Maybe a dip at 700hz-3k and a little boost around 160hz-200hz. Then the UAD Precision Enhancer Khz. This plugin is amazing for voice recordings of all kinds. It has a preset called “Voice” that I start with. It brings a nice presence to the voice without ever getting harsh. Less EQ is required. I would call it an exciter but every exciter I have used always made things too harsh and grainy. There is none of that with the Precision Enhancer. But you can definitely over do things if not careful. Then another De-esser and finally a brickwall limiter. The D-esser is usually doing about 5-7db of gain reduction at the most. The limiter is set for a boost of anywhere between 2 and 4.5 db. However much you boost the limiter input the gain reduction should be slightly less. So if I boost the limiter input by 4db, the amount of gain reduction should never be more then 4db. Helps keep things clean.

So multiple stages of transparent compression, a little shaping with EQ/enhancer/exciter and some limiting.

If the track has excessive background noise and isn’t too long I will edit the noise out to avoid messing with gates. When I do use the gates on voice I usually set up a sidechain.

Try this: Duplicate the track to be gated. Remove all plugins on the duplicated track. Set up a side chain send from the duplicated track to the original’s gate key input. Now use the track offset time in the inspector and set the duplicated track to play a little early. Maybe 20ms-40ms earlier then the original track. This can help with the gate triggering and remove false triggering or triggering too late and cutting off the beginning of a word.

For me the process is generally;

  1. Dialog editing during which I a) take out breaths (I don’t like breaths in Narration), and b) actually set general levels using clip gain. I have my meters set so that when I hit a certain color range it sounds loud enough with my Control Room set to “reference” levels, and that in turn puts the Narration in a narrow range for further processing. I also do RX restoration if needed, which typically will include de-clicking and spectral repair for any obvious issues, and sometimes DeNoise for issues with the tone. I often EQ at this stage. Typically static, but occasionally I’ll have to automate if there are different sources for Narration (i.e. different day/mic/room but with same talent).

  2. If time permits I now do one pass of riding the level of the Narration.

  3. Turn on my instantiated but previously bypassed compressor that sits on the audio channel post-fader. It’s set to catch annoying sudden peaks.

  4. Move on to the group on which I have a De-Esser (Nuendo) which I use if necessary, plus another compressor that acts more slowly and makes sure that the phrases are more even. So the release in particular will be a bit dependent on the type of read. If it’s a faster read then I’ll use a faster setting for example. I also makeup gain here, and it’s usually very very close to what’s needed if my editing was on-point. Last in this chain is a limiter which just catches peaks again and makes sure it’s easier to hit the broadcast specs for peaks. I may also use an EQ on this group if need be. I generally have it bypassed and only use it once I’m in the mix stage if I feel that I need to do some minor general changes to the tone of the VO. It could be that it sounded mostly fine to this point, but that it now feels a bit dull or vice versa. So that’s when I do a bit of broad adjustment.

I’ve experimented with other plugins, and so far I’ve really liked the UAD Fairchild on Narration. It’s very smooth and sounds great. I use that as a more ‘round’ and ‘general’ compressor rather than a ‘harder’ one. I’ve also used their Pultec a few times, which is great for that last stage to give it broad warmth or air or whatever. Good sounding stuff. I’m actually using less of UAD plugins however, simply because I’ve got a laptop and have occasionally used it to do tweaks while in another studio. Since UAD doesn’t “travel” with it (it’s PCIe) I’ve sort of gravitated towards using only included plugins, which I find excellent to a large degree (well done Steinberg). But I do feel that the UAD stuff adds something extra that is sometimes warranted.

So… that’s my chain, generally speaking…

Agreed… which is why, for my travel rig, I’ve got one of those UAD PC express slot cards, which handles enough to take care of basic mixing, and was cheap on eBay… but keeps me in the laptop dark ages for now…


I was actually contemplating getting an external unit, satellite or whatever they’re called, but UA’s attitude towards Mac OS vs Windows has completely turned me off to the company. It makes zero sense for me to invest in something that works either only on one platform and not the other, or is limited on one. I just don’t comprehend how they couldn’t work it out better. Other companies have managed quite well.

Not to divert the thread too obnoxiously-- but how is UAD limited on Windows? I thought they had that all ironed out… or so they said. One of my studio partners uses it on his PC, and doesn’t SEEM to have any trouble, but I have a feeling he doesn’t tax it much. BTW, I’ve got a Satellite, too… just got tired of lugging it around.


Thanks everybody for their responses!

It’s a bit easier for VO coming from a booth or treated room. For a documentary, in an interview situation, when people are sitting in their own office or room, it’s always a bit harder. I also start with basic editing, cutting out annoying noises or breaths and adjust the clip gain so I generally end up somewhere near -23dBFS on average with the voice.

For VO coming from a booth or treated room, I apply EQ to make the voice shiny. If it’s a room or office or anything, I learned a trick from another guy. He does some surgical EQ. Meaning he has a narrow Q and turns up a band of the EQ to maybe +10 or +15dB. Then sweeps areas in the 100-400 Hz. Usually there are room resonances at these frequencies that make any recording sound boomy. So I add 1 or sometimes 2 very narrow bands and sweep the room for resonances, and then reduce them by 5-8dB. This makes an incredible difference on clarity of the voice. Then I add an EQ to make the voice more shiny and smooth.

I found that a gate and compressor work very badly on people who are introverts and speak very quietly. There’s not a lot of dynamic range to begin with, not a lot to compress as it’s all quiet and static in loudness, and the gate tends to swallow faint breaths as they’re right in the middle of the gate threshold. Maybe it’s best not to gate and compress those too much, as there’s nothing to do.

When I come across any really problematic zones I also turn to RX to clean things up. My process is quite similar to yours MatthiasNYC, so I can’t be that far off :wink:

Then De-esser and then compressor, as I wrote above. So far it’s “in general” not that different from your setups. I just think I might have too aggressive settings. You do ratios of 2:1 or 3:1 and I use aggressive 6:1. Maybe that’s too much and I fear peaking transients too much, I should let the limiter handle those short transients and keep more dynamic range to make the sound more transparent and natural. And then I’ll see if a second slow compressor is necessary, also at low ratios, to smooth out longer dynamic changes.

Also, I googled a bit and it seems there’s a less complicated solution to what you wrote Rotund. What you are describing with copying a track and side-chaining. Many DAWs, including Logic and Reaper offer a “Pre-Listen” parameter that does exactly that. They look into the future some milliseconds and can then react quicker to what’s coming. I haven’t found anything of that sort in Nuendo, though. Why is there no pre-listen? Also, to avoid the gate closing too early when a word fades out too gently is the “Hold” parameter. It keeps the gate open longer for that duration. Great to avoid cutting off words. But the built-in noise gate of Nuendo doesn’t have this, only the Studio Gate.

OK, so I’ll experiment a little bit further with those info. Thanks again, great read!

Well, using a limiter instead of a compressor is really more for convenience rather than the two tools being radically different, a limiter is really just a compressor with a fast attack/release and a hard ceiling. But anyway, I’ve worked on either premixed projects or in templates done by other engineers years ago on Pro Tools systems before they had clip gain. Those engineers had very aggressive compression on the audio channels which means they were pre-fader. The result was super-compressed dialog tracks (and VO) before they went to the bus. Clearly it was a way to save time and not have to do much work riding levels, and it gave it all a certain sound, but I hated that sound.

So I’m just mentioning it because apparently to some ears using a more aggressive ratio and even a relatively low threshold so you’re compressing a fair amount of the signal seems to sound fine.

  1. How do you do your gain staging?
    What is the average loudness you set your clips to to enter the first plugin in your chain?

-9 to -12dB

  1. What plugins / settings do you use?

Low cut @ 80hz – 6dB Slope
Slight High lift and other dips/boosts where needed. but hardly.

Waves RVox
I top off what is needed to “round/fatten up” the voice a bit and make it understandable on lower volume. But it all depends on the voices. Some of them remain untouched.

No de-esser (means it is badly recorded)
OK, the off-one out who has an sss-problem, I edit them out manually. Select the “sssss” and lower it x-dB in volume.
No gate (the most awefull plugin on a voice), no noise reduction, and most certainly no limiting.

Breaths. If they are too loud, means you compress too much. If it’s really needed, then edit them out by hand.
Here again, with a talent that suffers from asthma, you will need to edit the crap out of his/her performance.

I’m currently recording commentary with a ElectroVoice RE20.

OMG, no.
I know this is a popular microphone in the US. But it sounds dull, lifeless, boring and any other insult I can think of. Use a Sennheiser 416, Neumann TLM170, U87 KMR81, KM185 or the likes. With a good preamp like the Pendulum, Amek Channel In a Box.
And you will hardly need EQ or compression.


I never thought of using my senheiser 416 for a voice over… I’ll give it a try.

He…depends a lot on the room your talent is , and positionong. It is a great VO mic as well.

Well, it seems those were $500 not well spent… I kind of like it though, not really for my voice I have to say but I have speakers where I think it sounds punchy. Anyway, everything comes with a learning. And since I needed a lot of gain for that thing, I saw there are 2 products that you can use with not so good preamps. The CloudLifter and if you prefer a European company, the Triton FetHead. They’re based in the Netherlands and have a great, transparent gain booster that does about +20dB. I like it a lot. (See good in bad…)

Anyway, do you also think the SM7B is all the attributes above? Because this was what many said was the alternative as speaker mic.

Well I’m sure a mic that’s almost 7 times the price sounds good :wink: I’ll see if I can borrow a MKH416 as I’m most interested in that one at the moment. Thanks for making me feel miserable :wink: (joke)

I just asked a VO friend of mine from the US what he was using. He uses exactly that and says he likes it the most for his voice in his vocal booth.

Follow up question. Isn’t that on the “hot” side? I heard that plugins that model analog gear also model the distortion of those hardware equivalents (UAD makes big deal out of this). So the analog gear has its sweet spot, which is if my memory serves correctly, somewhere at about -20dB? Now it sounds you don’t use any of those anyway, but just out of curiosity, wouldn’t that be the case? So you’d “overload” the plugins a bit if the signal would come it at this level, correct?


OK, it’s a good mic. But it has pretty much the same characteristiscs as the RE20, though better.
You got the advice from people who are in broadcast.
These mics are mainly used because of their great isolation. You can place them in a noisy room and they won’t pick up much of the surroundings. Or interfere with other open mics. That is why they are used so much in radiostations.

You might have seen footage of Michael Mc Donald or James Ingram using a SM7, but if you look closer, you will see that they are used when they sing with a group of others, or in the middle of a life band. Again reducing bleed.
There are a few specific singers who use this mic, but these are the exceptions, like Phil Collins using a Beyer88, or Mic Jagger an SM57.

The matter of fact is that none of these mic are suited for studio work. And YES, if you want to do record studio work, you WILL need some kind of good sounding recording booth. And YES, you will need some good equipment. But there is a lot out there. Try a Rode. I personally dislike them at the highest level, because they have some kind of signature sound in the high frequencies which sounds like a layer of rubbish. But you will get much better results even with one of these. Look for An AKG C460; great mic too. Or even a 451.

. Thanks for making me feel miserable > :wink: > (joke)

I know … I am not trying to make you feel bad. But really, once you will have heard the difference …

Follow up question. Isn’t that on the “hot” side? I heard that plugins that model analog gear also model the distortion of those hardware equivalents

That is correct for the UAD hardware, because they are working on a Fixed Point chip. In a floating point engine (like Nuendo) you have unlimited headroom within the application. In a Fixed Point -integer- environment, you hit the 0dB peak wall, anything over will result in plain distortion.

So the analog gear has its sweet spot, which is if my memory serves correctly, somewhere at about -20dB?

That depends on the quality of the hardware. It’s all about headroom. I have a rack of Neve 1272’s here somewhere, well, you can drive speakers with these. So the better your hardware, the more you can push it. Voice recording doesn’t have many peaks, so minus 9 or 12 is very safe with good DA’s and preamps. The lesser quality your hardware is, the more headroom you haev to reserve yourself to be on the safe side.

Remember, you can not undo a bad recording. You also can’t make a recording “better”. Like you can not remove a drop of black paint from a bucket of white paint. So your recording chain needs to be as good as possibly can. That will save you hours and hours of work.


Agree with Fredo on the mic choices but I’d suggest that the RE20 is not so bad… I think it’s highly dependent on the voice you are recording and how. That mic works better up really close and for high energy sources like rap, for example. But yes, a U87, KM84 or Sennheiser MKH461 or MKH40 and other high end mics are likely to be better for straight narration. Also read good things about the Audio Technica 4047/SV but never tried one. IMHO it’s worth having a range of mics for voiceover anyway because different mics suit different voices.

For a VO (which I do not specialise in to be honest) I do like MG M930, or on some I really like a standard sennheiser mkh50 or a shoeps blue mic. With a great talent and room a good sounding shotgun can be the perfect weapon, but the talent really needs to know how to work a mic with them IMHO. A nice LDC is a easier mic to work with for many.

Great info guys!!
Mic choice. I settled with an AT4050 and Great River preamp. Its a workhorse combo since years, never let me down. We are doing tons of VOs per day, so I become lazy to swap the mics/pres. :slight_smile:
Plugin chain is quiet simple. EQ, then RVOX doing max -6dB. If I need the voice to sound crisp for radio I’d use rComp with Pensados Rap preset. Instant crunch. For more softer effect I’d go for TubeTech clone from Softube. I like to edit out “sssss” in Izotope RX. On the bus sometimes a dynamic EQ to deal with offending freq and another limiter with max 3dB reduction. And another one on the MasterBus. Multiple compressors doing a little work.

There are so many combinations with material, voices, mics, etc. One secret weapon I’ll use on audiobooks (because Audible has strict RMS standards) is to use upwards compression. This solves a lot of track and program-length levels, as well as bringing up soft parts of a performance. You can also use it for voice-overs in loud commercials where the tail-ends of words fall down into the music. Of course there are many other plug-ins and settings you can use in combination. But when you just can’t get it to ride up enough, try upwards compression on an effect channel and send your VO track to it.

“Neil K” : where the tail-ends of words fall down into the music.

That’s a big one and a very good point :wink: It is mostly noticeable on semi-pro or even amateur VO talent. I normally try to coach them and using clip gain on those parts before any processing.

Just saying…

I also ride gain on the VO during the mix. The really great talent lean into the mic and do it for you.