Improving bad dialogue recordings for podcasts

[Tappistry]

Hi guys! I did a thread search and went back a few pages and found a few tips, but I figured it was still relevant to make my own thread for this question.

I've recently made friends with a podcast producer. A lot of his interviews done remotely over the phone, so he sends people an external iPhone mic for better quality audio capture. Unfortunately, some of the interviewees don't use the mic, and there's no way to know that until after he gets the audio.

I offered to help him clean up some of his bad audio recordings that he's scrapped because I'm up for challenge. He's particular about audio quality, and he's ditched plenty of interview content because people didn't use the provided microphone or because of a bad recording environment. He also knows I'm an amateur with mixing/mastering, and is happy to give me some practice. So, here I am.

He sent me his first recording recently. It sounds like it was recorded with the iPhone voice memo recorder on the stock microphone. The voice (female) sounds nasal, honky, has no depth, etc. It almost sounds like it's out of phase, if that's possible with something recorded on speakerphone. And of course, all the plosives, the pops and breath noises ... yeah, it's bad.

For the sake of content propriety, I won't post any before/after clips here. But I'll share what I did to this track to clean it up:

1. Acon Digital Restoration tools to remove plosives, crackles, and the distracting noise junk.
2. ReaGate for the producer's part of the interview bleeding into her phone, de-ess, and Nectar Breath Control.
3. Impulse response of a good mic.
4. Desk saturation via Klanghelm SDRR, with some modest bass cut and treble boost dialed on the plugin.

I got a pretty good result, and he was impressed with the improvements enough that he said he'd give me more tracks to work on. But, after we discussed it further, my initial corrections aren't quite there yet.

What are some eq points I should focus on boosting or cutting to improve vocal clarity and depth from a bad microphone? What other effects are worth trying? Is this type of rescue work a lost cause? I realize that there's no magic wand for this, and I imagine that there's only so much improvement to be made to this kind of source material to make it good for production. But any tips or guides for this particular application that are worth reading would be welcome. Thanks!

[BertKoor]

What are some eq points I should focus on boosting or cutting to improve vocal clarity and depth from a bad microphone?

There are no fixed EQ freqs that work every time.
What does work, is the following approach to track the critical bands:

Use a parametric EQ, set it to boost. Slowly sweep its frequency and find the spot where it sounds most awful. That's the point where you now should cut! Repeat this with more bands when needed.

So the rule of thumb (on which there are plenty exceptions) is to cut instead of boost.

[Tappistry]

What are some eq points I should focus on boosting or cutting to improve vocal clarity and depth from a bad microphone?

There are no fixed EQ freqs that work every time.
What does work, is the following approach to track the critical bands:

Use a parametric EQ, set it to boost. Slowly sweep its frequency and find the spot where it sounds most awful. That's the point where you now should cut! Repeat this with more bands when needed.

So the rule of thumb (on which there are plenty exceptions) is to cut instead of boost.

Thanks for that. I’ll sweep around with ReaEQ and try to find the terrible spots.

[Tappistry]

Turns out a fair amount of overall midrange cut really did the trick.

[funky lime]

If you're trying to make it a regular thing, cleaning up audio, you might be interested in Izotope RX. Its spectral denoise module is pretty great, combined with the fact that you can "paint" with the various effects directly onto the spectrogram view of the audio.

It has gotten me out of a bind when I had to do something like remove a noisy child from the background of an interview, for instance. It's also great for constant noise (like ground hum, A/C unit, etc), mouth noises like plosives and sibilants, and has a whole lot of other neat functions as well.

You could also transcribe the dialogue and record voice-over yourself, if you really want to mess with your buddy.

[Tappistry]

If you're trying to make it a regular thing, cleaning up audio, you might be interested in Izotope RX. Its spectral denoise module is pretty great, combined with the fact that you can "paint" with the various effects directly onto the spectrogram view of the audio.

It has gotten me out of a bind when I had to do something like remove a noisy child from the background of an interview, for instance. It's also great for constant noise (like ground hum, A/C unit, etc), mouth noises like plosives and sibilants, and has a whole lot of other neat functions as well.

You could also transcribe the dialogue and record voice-over yourself, if you really want to mess with your buddy.

I actually sold my MPS bundle so I could pick up Acon Digital Acoustica Premium. I think it's pretty solid competition to RX6 when you consider the value of the full Restoration Suite that's included, as well as the other mixing and mastering plugins, and spectral editing. It's sort of like a mini DAW.

I can do spectral editing inside of Reaper as well.