- Unsupervised Learning
- Posts
- Chasing the Perfect Podcast Microphone Sound
Chasing the Perfect Podcast Microphone Sound
I’ve written like 5 articles about getting the right sound for my podcast. I talk about the different mics, the different settings, recommendations from NPR professionals, etc.
You can see the other articles in the related posts section below.
I’ve once again had an epiphany on this topic, so I thought I’d share it here in case it’s useful to someone.
Here’s how my basic evolution of thinking has gone:
Whatever, sound doesn’t matter that much (2015)
I guess I’ll get a Yeti mic because they look awesome (2015-2016)
It has to be all about the mic, so let’s figure out what the pros are using
More bass is authoritative
The RE27 N/D is the radio workhorse, so that’s the answer
Oh but wait, NPR uses the U87ai, so that’s even better
More bass sounds cool in the first 30 seconds, but is crap to listen to over time
Ok, Sam Harris has my preferred sound and he supposedly uses an SM7B
What this has brought me to is where I should have started with, but lacked the fundamentals to arrive at. Namely, it’s all about the impact to the listener.
Well, duh. But what does that mean?
First, the heavy bass sound that so many men go for is just bad. Don’t do it. It produces a dull, throbbing sensation that assaults the ear over time. And even worse, it is hard to hear when you have low-pitch noise that competes with it.
A sound that’s in the mid to higher range cuts through background noise better, and is easier to listen to.
Second, forget how podcasters sound and ask yourself how audiobooks sound. They’re designed to be listened to for hours at a time. And since I listen to lots of audiobooks, I started paying attention to their sound.
What do they tend to have in common?
They aren’t bassy at all, but they’re also not crisp. They’re mid to mid-high end in tone.
Next, they don’t have any high-fidelity character in the voice. It’s a neutral voice sound, so that you aren’t distracted by the speaker and can focus 100% on the content itself.
And that’s what got me thinking about Sam Harris’s podcast and all the various mics and sounds that I’ve tried.
I own the Yeti, the RE27, and the U87. And I just bought the SM7B.
When I was using the Yeti I didn’t know anything about sound, so that one doesn’t count. With the RE27 I was just learning, but I liked the fact that it had a wide dynamic range. And with the U87 I was obsessed with the best possible range and the fact that NPR used it.
But here’s the thing—it requires a studio to get the most out of it because it’s so damn sensitive. And even if I had a studio, it turns out I don’t want to hear perfect detail in my voice!
In fact, I just want to hear certain parts of it, in a clear mid-to-high range. And I want that sound to be extremely flat. Not flat in tone, but flat in character. I want it super dry. No noise. No clicks or pops.
Unsupervised Learning — Security, Tech, and AI in 10 minutes…
Get a weekly breakdown of what's happening in security and tech—and why it matters.
No distractions.
This is precisely what I’ve loved so much about Sam Harris’ sound for so long.
He has a bassy voice, but the audio isn’t bassy. It’s flat and dry. And it doesn’t sound super accurate or super sensitive. It’s just clear.
I just set up my SM7B, which is the mic that Joe Rogan and he use, and I turned it down to NPR-like settings. Meaning, as little messing about as possible.
Cloudlifter
Bass rolloff on the mic
No presence enhancement on the mic
Standard pop filter (not the more bassy one)
Gain set to 40 on the RODECASTER Pro
Compressor, De-esser, Aural Exciter, Processing, High-Pass filter, and Noise Gate on RODECASTER Pro
No Bass Boost on RODECASTER Pro
No equalizer in post (Hindenburg)
Moderate compression in post (Hindenburg)
So I’m reducing bass on the SM7B itself, reducing more bass by using the High-pass filter on the RODECASTER Pro, and purposely not using the Big Bottom setting either.
And then I’m not using any equalizer in Hindenburg in post.
And I think these settings may have finally done it. The sound I just got is really close to what I’m looking for, with its unassuming forwardness and clarity.
Here it is for a listen.
I’ll continue to update this post as I make improvements. But hopefully this is fairly close to where I’ll be for a while. It’s time to land and focus completely on content rather than accouterments.
Anyway, hope this helps someone.
Summary
The bassy sound is not your friend
It’s possible to have too much clarity because it can distract the listener
My landing place is unassuming, dry, clarity, like what you hear above
Notes
June 29, 2020 — The other thing I realized after doing even more comparisons, and talking to some smart people on Reddit, is that there is such a thing as matching a tone to content or to one’s preference. I like the sound above because I think it matches my content and my “brand”, even if it’s not perfectly loud and clear and bassy. Someone on Reddit pointed me to some audio that sounded super loud and clear and crisp, and it was someone doing reviews of microphones. They were upbeat and friendly and told a few jokes, and the sound of the audio was great. It truly was. But I tried to imagine my audio sounding like that and realized it wouldn’t be a great fit. I like the subdued, quieter, less assuming, less forward, less clear, less loud sound. In short, I like the vibe of a late-night conversation with a friend, with some wine, about a deep topic—not the vibe of someone enthusiastically showing their followers the latest camera from their favorite manufacturer. In short, it’s not just about the audio; it’s about whether the audio matches your brand and content.
July 7, 2020 — I have since had yet another advance in my thinking on this. Or at least I hope it’s an advance. So I once again listened to the above comments on not having enough bass in the clip I previewed here—from someone on Reddit—and came away thinking I had taken away too much bass. But then I relistened to Joe Rogan’s show, and a ton of NPR and NYTimes podcasts. They’re all super-low on bass. Like massively. And another thing: they sound really good in a car. They’re very clear. So what I did was I took my perfect Sam Harris sound profile that I got to my own sound and I just removed a ton of the bass. Probably half or so. This dropped me right in the middle of NPR and Sam Harris. And I really like where that ended up. I am also using more plugins from Izotope which really cleaned up my noise gate flapping and I removed extra processing and compression. So the sound is actually more natural than the clip above. Here’s a show done using the new processing. Bottom line: I have a sound I like which is super clean and muffled and bassy, and I have a sound that I know the industry recommends, which is not what the Reddit guy said with lots of bass, and that actually has very little bass. And what I’ve elected to do is find a middle-ground that still has a branded sound I like (different than Sam’s), but that also follows the rules that NPR have discovered over decades of experience. For now it’s the hybrid that makes me happy!