Enhancing Your Podcast Audio with AI: The Pros, Cons, and Best Practices

[00:00:10] Hey Clipped listeners, do you ever feel like you're juggling your script, your show prep, your questions, and monitoring your recording on the fly? I get it. I know what it's like and that's why I'm here to tell you about Riverside's new teleprompter feature. It's built right into Riverside's platform so you can have your show prep on screen and scroll as you record all within the session window. No more tabs switching, app juggling. It's a pain doing that on the fly and now you're not going to have to because this teleprompter feature is game changing.

[00:00:39] You can pause it, slow it down, speed it up, even replay whatever you need to do to stay organized and calm during your session. I think it just makes podcasting a lot easier. And for me personally, it makes me less anxious when I'm in the middle of a recording. So if you're looking to smooth out your delivery and keep things organized while podcasting, you should check out Riverside's teleprompter feature. I think you're going to like it. And hey, you can use promo code CLIPED, that's C-L-I-P-P-E-D, to

[00:01:09] snag 20% off of any individual membership plan. That's CLIPED for 20% off any individual plan and you're going to start flowing like never before.

[00:01:31] What up potty people and welcome back to CLIPED. I'm Eric, your host, and this is a podcast where I'm dropping podcast production tips, resources, industry insights, podcasting knowledge for my podcasting pals. And it's all aimed to help make your podcasting life easier. I'm on a mission to help you start, grow, and monetize your show.

[00:01:50] And hey, today we're talking about something fun. And that's the role of AI in processing your podcast audio, talking about like speech enhancement, beefing things up, where AI can take you from sounding like you're in a phone booth to sounding like you're big and full on the microphone.

[00:02:09] And so this episode is all about my philosophy on speech enhancement via AI, how it can boost and improve your dialogue. I'm giving you the pros and cons and also some tips for those of you that are overdoing it with AI and some tips for those of you who have yet to try AI. So without further ado, let's give it a go.

[00:02:33] Okay, so the role of AI in audio production is fairly new. I want to say a year or two, maybe less than that. But really, what these AI tools do, and there's two that I'm going to get into in this episode, and that's Descript Studio Sound, which I have used many, many times. And then there's Adobe Speech Enhancement, which I haven't used, but that's also very popular.

[00:02:58] And the role of these is really to take, I don't want to say terrible audio, but take like mediocre audio and improve it and get it to sound like it was recorded on a better microphone or it was recorded like in an actual studio. It beefs it up, it improves like EQ problems, the frequencies of the voice, plays a huge role in noise reduction and like de-reverb.

[00:03:21] So that's what we're talking about when we're talking about like AI speech enhancement.

[00:03:26] This is a popular topic and it's being used by pros, by beginners, by intermediate. Seems like almost everyone is using some form of AI. And I don't always do that. I'm going to get into more of that as we go on here.

[00:03:41] So I mentioned Descript Studio Sound and Adobe Speech Enhancement, and I actually have never used the Adobe tool. I have used Descript Studio Sound.

[00:03:51] And a little bit about how that works is once you import your audio to Descript or Descript, however you like to say it, it's going to transcribe your audio.

[00:04:00] I've done episodes on that. But then there's a little button you can click and it's called Studio Sound and then a slider comes up like, and that slider is from zero to 100%.

[00:04:11] And so you can apply this AI studio sound from 1% to 100%. And the more you pump it up, the more your voice starts to sound like,

[00:04:22] Hey, everybody, welcome to the podcast haven. And it kind of boosts it. It reduces any inconsistencies with the leveling.

[00:04:32] Again, noise reduction, EQ tends to add a lot of low end in my experience. Descript does have a free plan. Don't know if you get access to that.

[00:04:41] I have a paid plan. I think it's $14.99 a month or $15.99 a month. I'll put a link to Descript. It's an affiliate link in the show notes.

[00:04:48] I'm going to put a sample here and you can listen to this that was recorded on my iPhone.

[00:04:54] And I'm going to run it through Descript Studio Sound so you can hear how this sounds.

[00:04:58] All right. So here's the raw recording. Check. Hey, hey, this is Eric with the podcast haven.

[00:05:05] And now here's the same at 30% with Descript Studio Sound. Check. Hey, hey, this is Eric with the podcast haven.

[00:05:22] 90% check. Hey, hey, this is Eric with the podcast haven. And 100% check. Hey, hey, this is Eric with the podcast haven.

[00:05:32] And so it's a huge time saver if you don't know how to do that manually. And even if you do know how to do a lot of that stuff, EQ, compression, noise reduction manually, it's going to expedite that.

[00:05:45] And it's like this helping hand that their algorithm, quote unquote, knows like how that voice should sound.

[00:05:52] And so it's going to apply that stuff. And you're going to be off to the races with a decent piece of audio pretty quickly.

[00:05:59] And that's great for just like consistency. If you can sort of match like someone that was recorded on a microphone sounding good.

[00:06:08] And then somebody, say it was on Zoom, just talking into the laptop mic, it's going to kind of match them up.

[00:06:14] It just does wonders. And I know I'm being repetitive, but it really is remarkable what these AI processes can do.

[00:06:22] And it's also just going to open doors for like beginners, which I am a fan of because it lowers the barrier to entry.

[00:06:29] Let's say like you're not an audio engineer, but you want to get into podcasting, but you don't know how to make the voice sound good and EQ it and level it and mix it.

[00:06:37] It's going to do that for you. And so on one hand, I think that's great because it opens the door for new podcasters and people that kind of want to learn about audio.

[00:06:49] But there are some cons to that and some opinions that I have, which I'm going to get into now.

[00:06:55] The first is the risk of over processing.

[00:06:58] I find that people that don't really know about audio, like that new person that's getting in and they're excited.

[00:07:05] Oh, I can make my audio sound great with this AI.

[00:07:07] It tends to over process. It tends to make your voice sound like you are a broadcaster in the 1970s.

[00:07:16] And yes, that's probably an improvement from how it was if you didn't know how to record and you got crappy sound on the way in.

[00:07:23] But that over processing also does another thing, which is it can actually start to cut off words.

[00:07:31] It does really weird things with breaths where we'll reduce that breath, but it'll kind of cut the tail end of the word going into that breath and the top end of the word coming out of that breath.

[00:07:43] And that's because people over process.

[00:07:45] Like I said, on Descript, there's that slider from zero to 100.

[00:07:49] If you crank that thing up to 100, you're going to get some weird little blips in the audio.

[00:07:54] The algorithm isn't like perfect. It's good.

[00:07:57] But you can tend to get some like skipping sounds and digital blips.

[00:08:01] It just sounds kind of fake and not only fake, like I mentioned with the boosted voice.

[00:08:07] But if you don't like listen back, there's going to be many little glitches in the audio that you can hear or at least I can hear.

[00:08:17] Maybe the average podcast consumer either can't hear them or don't care or both.

[00:08:22] But the risk of over processing, in my opinion, is something that you need to steer clear of.

[00:08:27] And then there's like the dependency dilemma.

[00:08:30] People, I think, are becoming too dependent on it.

[00:08:34] It's great on one hand, like I mentioned, because people are going to be able to get into podcasting and produce some decent sound without knowing anything about audio.

[00:08:43] But people are starting to come too dependent on it.

[00:08:46] And I've even seen people that like run podcast production agencies that don't necessarily have the skills or they haven't put the hours in to learn about like audio and to learn how to adjust a voice and EQ it and mix it and level it.

[00:09:05] And I don't know why, but it kind of bugs me because I think you should learn that.

[00:09:09] And I guess that goes against what I said of it opens that barrier to entry.

[00:09:14] But I think there just has to be a balance there.

[00:09:16] I don't think that it's okay to just 100% rely on AI.

[00:09:22] I think it's a tool.

[00:09:23] And if you use that tool correctly, and you're not too heavy handed in over processing, you can actually improve some of the audio that you get quicker and possibly better than you could of just using like your plugins and your tools that you typically edit and mix with.

[00:09:41] So is it a crutch?

[00:09:42] I think it can be.

[00:09:44] And I think that's why here I want to talk about not being too heavy handed with it.

[00:09:48] Striking a balance is what I like to do.

[00:09:50] There's actually one show, I'm not going to mention the name, that I've been working on for over a year.

[00:09:54] And I have just never been able to get the two girls' voices right.

[00:09:59] And it's not like their recordings are terrible.

[00:10:01] They're actually decent.

[00:10:03] I don't know if it's like the mic on that person's voice or the two girls' voices.

[00:10:08] So I honestly run them through Descript.

[00:10:10] But here's what I do.

[00:10:11] And I recommend everybody start here is I don't crank that AI up to more than like 30 or 40%.

[00:10:18] And I think that's a great starting point.

[00:10:20] And yes, you can apply a little bit more or even less.

[00:10:24] That puts you in a good place to get like a decent starting sound.

[00:10:28] Then you can go into your DAW and apply the EQ and fine tune and put a little compression and listen for the breaths and listen for things.

[00:10:38] So I think striking that balance is important.

[00:10:40] But the reason I'm pointing this out is just to show you that like I'm not anti by any means.

[00:10:44] Oh, and another thing is that certain voices respond better to AI.

[00:10:49] It seems like those two girls, it doesn't like make miracles.

[00:10:54] But then I see like these videos on YouTube from like Descript or Adobe or these other services.

[00:10:58] And they'll have like a person on an iPhone in traffic talking into it.

[00:11:02] And then they drop it in apparently to their AI thing.

[00:11:06] And it sounds like so much better.

[00:11:08] And I don't know if maybe I'm still new and I'm learning how to use these AI processors and like what works and what doesn't work and what type of environment it works better on than others and certain mics or male, female.

[00:11:22] I still have some learning to do and some experimenting.

[00:11:26] And maybe that's why I'm not recommending you or myself crank up these AI sliders to like 80, 90, 100%.

[00:11:36] Maybe as I experiment more, oh, I'll be like, yeah, maybe you can like crank it up to 60 if you do these certain things first.

[00:11:44] Or if it was recorded on this mic, it's going to sound better versus that mic, etc.

[00:11:51] And so I just want to let you know that I have some room to grow with that as well.

[00:11:56] That's it for me today, guys.

[00:11:57] Wanted to talk about my philosophy and those pros and cons on AI.

[00:12:03] So hopefully you have a good idea of where you stand.

[00:12:07] And I think if you haven't tried it out at all, I highly recommend you at least give it a go and start experimenting a little bit.

[00:12:14] And again, start low on the processing and crank it up and see how it sounds.

[00:12:18] The key with anything, guys, is to use your ears.

[00:12:21] And then if you've been using too much of it and you're not happy and you don't know why it doesn't sound the way you want,

[00:12:28] that's because you're using too much of it and it's damaging your audio more than it's improving it.

[00:12:33] So if you're overdoing it, I recommend dialing it back.

[00:12:37] If you haven't done it at all, crank it up a little bit and get into it.

[00:12:40] And the two tools I recommend are Descript Studio Sound, which I have personal experience, and Adobe Speech Enhancement.

[00:12:47] To be completely honest, I haven't used Adobe Speech Enhance.

[00:12:52] I mentioned that earlier.

[00:12:53] But the reason that I'm recommending it is because it's highly reputable.

[00:12:57] Adobe products are top of the line.

[00:13:00] And I've heard a lot of stuff that podcasters have produced with it and have heard samples.

[00:13:05] And so although I haven't used it, that's the one I'm going to recommend in addition to Descript.

[00:13:11] The free version of Adobe Speech Enhance, you don't get to adjust how much AI you put into it.

[00:13:20] So I don't know how that works.

[00:13:22] It just does its thing with the algorithm and spits it back out to you.

[00:13:25] It's my understanding that the paid plan, you can adjust the AI processing from a little bit to medium to a lot.

[00:13:32] And so that's my call to action, guys.

[00:13:34] Give it a go or dial it back.

[00:13:36] And I think that should give you some insight and help you on your way to crafting better edits, better mixes, better recordings.

[00:13:45] And just remember that AI is a tool.

[00:13:48] And if you use it correctly in the right circumstances, the tool will do the job that it's supposed to do.

[00:13:55] So thank you guys for listening.

[00:13:57] It's been a blast.

[00:13:59] Head on over to the website, thepodcasthaven.com.

[00:14:02] Tons of articles on the blog, over 100 articles.

[00:14:06] They're ranking stuff about education, how to podcast, what equipment to buy, how to grow your audience,

[00:14:13] how to become a better host, a better guest.

[00:14:17] Talking about everything in the podcast industry.

[00:14:19] That's thepodcasthaven.com.

[00:14:22] Check out that blog.

[00:14:23] Give me some feedback.

[00:14:24] And I look forward to connecting with you next week.

[00:14:27] Keep listening.

[00:14:28] Keep engaging.

[00:14:29] Keep creating.

[00:14:30] I'm out.