What parameters should I apply to each effect on my voice-over to achieve a lively and natural sound?

[ruidito]

Greetings, everyone.
Some time ago, I reached out to these forums seeking assistance to kickstart a voice-over recording project. The excellent recommendations I received led me to purchase the hardware I currently own and to install the software I am currently using:

Hardware

• Shure SM7B
• Cloudlifter CL-2
• Shure by Gator Stand
• Solid State Logic SSL 2 Plus MKII
• 2 Mogami Gold Studio 06 XLR-to-XLR 4-conductor cables
• Sennheiser HD 600 (already owned)
• Audio-Technica ATH-SR50 (already owned)
• MacBook Pro M3 Max (already owned)
• Windows 11 PC (already owned)

Software

• Final Cut Pro (already owned)
• Audacity (on Mac and Windows 11)
• Adobe Audition (cracked version 25.6.0.97, Windows 11)

Now then—after asking which hardware to buy—I asked you all which effects I should apply to the recorded voice-over, and in what order. You recommended the following:

Noise Reduction (Cleanup)
High-Pass Filter (Clean up low frequencies)
Compressor (Add body and consistency)
Equalization (Add brightness/clarity)
Normalize/Limit (Final volume)

That covers the advice I received from you all. Now, to proceed further—since I know absolutely nothing about this subject—I turned to Claude. I needed to know what specific settings to apply in each case, and it generated a guide for me that I followed to the letter. However, while the audio sounded beautiful, it also felt somewhat flat—a bit congested, lacking "air," and devoid of sparkle or expressive nuances. It felt as though one of the effects had stripped some of the life out of the recorded narration.
This is where I am once again asking for your assistance. I would like to know what specific settings I should apply for each effect. I understand that it may be difficult to help me without actually hearing the audio files. If you need them, I have them—both the raw recording and the resulting file after applying the guide. I sincerely hope you can guide me, as AI—at this point—is not only not free, but obviously not human; and I want that human warmth to edit my human voice.

Thank you very much.

 

[dasdoing]

Try to AB every step and adjust. developing "that ear" is a long and continuous way, though.
The HP might be too high, if you have a deep voice > AB. in theory, it should be redundant anyway if the noise-reduction did a good job.
In the noise reduction step, listen for artifacts. How do they sound? People generally describe them as "metallic-sounding". btw, I use this free plugin for that: https://github.com/werman/noise-suppression-for-voice But I use it for live purposes (meetings) and never tried it in post-production.
The EQing is hard to learn. play with the sliders and listen to what every one of them does. Once you hear what it does, fit it in into the context. I think it might be better for a beginner to just play around with high and low shelves.
The compression is another beast that is hard to hear at the beginning. You have to AB a lot to actually hear what it does. Keep an eye on how much it actually reduces. it shouldn't be too much.
The limiter stage shouldn't really alter the sound. Dial it in so that you just can't hear it. That's the goal. You only need it to be able to make the signal louder.

all in all: I was the one suggesting AI tools in the other topic.....because IT IS hard to learn. At the beginning, you will do more harm than improvement.

 

[olds1959special]

Noise Reduction (Cleanup)

Only needed if you hear things like hiss, hum, pops, or excessive reverb.

High-Pass Filter (Clean up low frequencies)

You can use a 80Hz HPF (24dB/octave)

Compressor (Add body and consistency)

Boost the input gain a few dB, maybe +3dB to +7dB. You may be able to leave the other settings at default.

Equalization (Add brightness/clarity)

Only needed if you have resonances in your space you need to remove, (or nulls you need to boost,) but cutting works better than boosting.

You can do a sweep with boosted narrow EQ to hear where there are big resonances in the room and then cut those.

Normalize/Limit (Final volume)

Set your limiter slightly lower than 0dB (-1dB to -3dB) to prevent peaking.

 

[ruidito]

Try to AB every step and adjust. developing "that ear" is a long and continuous way, though.
The HP might be too high, if you have a deep voice > AB. in theory, it should be redundant anyway if the noise-reduction did a good job.
In the noise reduction step, listen for artifacts. How do they sound? People generally describe them as "metallic-sounding". btw, I use this free plugin for that: https://github.com/xiph/rnnoise. But I use it for live purposes (meetings) and never tried it in post-production.
The EQing is hard to learn. play with the sliders and listen to what every one of them does. Once you hear what it does, fit it in into the context. I think it might be better for a beginner to just play around with high and low shelves.
The compression is another beast that is hard to hear at the beginning. You have to AB a lot to actually hear what it does. Keep an eye on how much it actually reduces. it shouldn't be too much.
The limiter stage shouldn't really alter the sound. Dial it in so that you just can't hear it. That's the goal. You only need it to be able to make the signal louder.

all in all: I was the one suggesting AI tools in the other topic.....because IT IS hard to learn. At the beginning, you will do more harm than improvement.

Thank you so much for breaking down each effect; this helps me familiarize myself a little with their objectives.

 

[ruidito]

Noise Reduction (Cleanup)

Only needed if you hear things like hiss, hum, pops, or excessive reverb.

High-Pass Filter (Clean up low frequencies)

You can use a 80Hz HPF (24dB/octave)

Compressor (Add body and consistency)

Boost the input gain a few dB, maybe +3dB to +7dB. You may be able to leave the other settings at default.

Equalization (Add brightness/clarity)

Only needed if you have resonances in your space you need to remove, (or nulls you need to boost,) but cutting works better than boosting.

You can do a sweep with boosted EQ to hear where there are big resonances in the room and then cut those.

Normalize/Limit (Final volume)

Set your limiter slightly lower than 0dB (-1dB to -3dB) to prevent peaking.

Thanks.

 

[ChrisG]

How close are you to your mic?
Do you have any other mics available?

The SM7b is known for a few things, but HF clarity is not one of them.

If you're pretty close to your mic, then you're probably getting some LF buildup due to proximity effect.
A steep 80Hz HPF will take out the plosives etc, but doesn't do much to reduce the rising low-mid/bass from being close to the mic.
Try: 200Hz 12dB/octave, or 500Hz 6dB/octave.

EQ: You can apply considerably more EQ if you wanted. Try +6dB high shelf at 5kHz. Feel free to send me a file to work on.

 

[olds1959special]

How close are you to your mic?
Do you have any other mics available?

The SM7b is known for a few things, but HF clarity is not one of them.

If you're pretty close to your mic, then you're probably getting some LF buildup due to proximity effect.
A steep 80Hz HPF will take out the plosives etc, but doesn't do much to reduce the rising low-mid/bass from being close to the mic.
Try: 200Hz 12dB/octave, or 500Hz 6dB/octave.

EQ: You can apply considerably more EQ if you wanted. Try +6dB high shelf at 5kHz. Feel free to send me a file to work on.

You can remove the built-in pop filter from the SM7b to get a little more clarity (just use a separate pop filter). 1-2 feet away from the mic is about right I think.

 

[ruidito]

You can remove the built-in pop filter from the SM7b to get a little more clarity (just use a separate pop filter). 1-2 feet away from the mic is about right I think.

Thanks.

 

[ruidito]

How close are you to your mic?
Do you have any other mics available?

The SM7b is known for a few things, but HF clarity is not one of them.

If you're pretty close to your mic, then you're probably getting some LF buildup due to proximity effect.
A steep 80Hz HPF will take out the plosives etc, but doesn't do much to reduce the rising low-mid/bass from being close to the mic.
Try: 200Hz 12dB/octave, or 500Hz 6dB/octave.

EQ: You can apply considerably more EQ if you wanted. Try +6dB high shelf at 5kHz. Feel free to send me a file to work on.

You're very kind, honestly. I'll send you the raw and final versions with the effects chain applied. They're small, lightweight files. What's your opinion?

AUDIOS

 

[ChrisG]

Okay, some notes:
- Your noise reduction has worked well.
- Both files are bass-heavy. I think you are quite close to the mic.
- If I apply strong settings with the compressor/limiter, I can hear some room reflections quite clearly. I would suggest trying to address those for really professional results.

The thing I wanted to play with the most was EQ. I tried playing around with some HF boost for extra clarity, but any HF boost I tried just made the sibilants sound weird. This is probably the SM7b's non-linear HF response making life more difficult.
The biggest improvement that I found was a 6dB/octave highpass at 400Hz. This is my preferred approach for voices when the mic is used relatively close-up: the 6dB/oct slope tends to match well to the proximity effect, leaving a flat overall response. 500Hz left you sounding very thin, and I can tell there is some LF in your voice naturally, so 400Hz was IMO a good balance.
I also found some benefits in the grahic EQ: +3dB boost at 2.5kHz and 3.15kHz added a little bit more "clarity" to the sound, without anything sounding harsh/weird.

Honestly, though, the 400Hz HPF was the biggest improvement by far.
Your other clips will sound very bass-heavy in comparison, so I'd recommend listening to the 400Hz HPF version a couple of times so you acclimatise to that sound, and then try listening to the other takes. I trust you'll see where I'm coming from.

The comp/lim you'd done was fine. No obvious issues there.

For me, the biggest issue was the overal tonal balance, and I hope my recommendations above will be useful for you.

Your vocal performance was IMO good - while I don't understand the language, I could tell that you were enunciating clearly and the cadance felt about right.

Edit to add: I monitored with my Shure SRH1540 headphones, which I use frequently for my live sound work.