The test I trust most for podcast audio has nothing to do with meters, it is what a listener does with their thumb, because if someone in a car has to turn your episode up when the guest speaks and back down when the music comes in, right, the mix has failed, and if they have to turn your whole show up compared with the episode they heard before yours, the master has failed, and they will rarely tell you either of those things, they just quietly stop listening.
So when people search for podcast mixing and mastering, what they are usually after is that thumb test, a show that sounds even, clear and the same loudness as everything else in the feed. Those are two different jobs that people often blur into one, and I want to separate them properly, explain what a service should hand back at each step, and be honest about when a tool does the job well enough and when you need people.
What is the difference between mixing and mastering a podcast?
Mixing is balancing everything inside the episode, the voices against each other, the noise underneath them and the music beds around them, while mastering is setting the finished file's overall loudness and peaks so it plays correctly on every platform.
The simplest way I explain it is that mixing is about the relationship between sounds and mastering is about the relationship between your episode and the outside world. A good mix means the quiet guest and the loud host sit at a similar level, the air conditioner hum is gone, and the intro music ducks under the first words instead of fighting them. A good master means the whole file lands at a sensible loudness, never clips, and is encoded in a format the platform expects.
The order matters, right, because you cannot master your way out of a bad mix. If one voice is far quieter than the other, raising the overall loudness just makes the loud voice louder too, so the imbalance survives, and that is why I am a little suspicious of any service that only talks about the final loudness number.
What should a proper podcast mix include?
A proper mix should include level balancing between every voice, noise reduction that does not make speech sound underwater, control of harsh sibilance and plosives, and music beds that sit under the speech rather than on top of it.
Let me go through how we actually think about it. That is one, voice balance, which is not just turning the quiet person up, because people get louder and softer within a sentence, so the levelling has to be gentle and continuous or the guest sounds squashed. Secondly, noise, where the catch here is that aggressive noise reduction removes the hum but also removes the breath and texture of the voice, so the goal is to remove only what a listener would notice. And the third thing is music, where intros, transitions and sponsor beds have to drop under speech fast enough that the first word is clear, and come back up slowly enough that it does not feel like a door slamming.
There are smaller jobs too, like crosstalk, where two people talk at once and the editor has to decide whose line the listener needs to hear, and room sound, where a remote guest recorded in a kitchen needs to sit next to a host recorded in a treated room without the jump being obvious. I wrote about the recording side of that in getting a remote guest to sound like they are in the room, and the cleanup service itself is described on podcast audio cleanup.
Levelling and noise cleanup come with every episode we edit, but the part I would really point at is that a person on our team listens to the whole thing first, so the tools do the cleanup while a human decides what counts as noise and what is part of the conversation, because a laugh, a sigh or a pen tapping during a thinking pause can be the most human moment in the episode, right, and a filter will happily delete it.
What loudness should a podcast be mastered to?
For spoken word, the most useful published reference is the one from Apple Podcasts, an integrated loudness of roughly -16 LKFS, give or take one decibel, with true peaks staying at or under -1 dBFS.
You will see LKFS and LUFS used for the same measurement, and in practice they mean the same thing, an integrated loudness measured across the whole programme rather than a single peak. Apple's guidance for podcast audio is the page I would keep open, and they also recommend AAC for RSS feeds, which is worth knowing because the format and bitrate you export at affect how the episode sounds after it leaves your hands.
The true peak number matters, right, because of encoding. When a file is converted to a compressed format, peaks that looked safe in the edit can overshoot and distort slightly, so leaving a little headroom below zero protects the episode, especially on the loudest laughs and the moments where someone leans into the microphone. And consistency matters as much as the target, right, because a show that lands at a different loudness each week fails the thumb test even if every individual episode sounds fine.
If your listener has to touch the volume once during an episode, the mix let them down, and if they touch it between episodes, the master did.
How do you mix a video podcast differently from an audio-only show?
The mix itself follows the same rules, but a video podcast also has to stay in sync with the picture across every camera angle, and its loudness has to work for a viewer watching on a phone speaker as well as a listener in earbuds.
On a video show the audio is usually recorded on a separate recorder or a remote platform and lined up with the cameras, and sync is one of the jobs we use tools for. When it drifts, which happens with long recordings, it has to be fixed before any mixing makes sense, and what to do when audio drifts out of sync explains the fix. On our side the YouTube export, the Spotify export and the audio master for your podcast host all come out of one mix, so the version people hear in the feed and the version they watch never drift apart in how they sound.
For audio-only shows the mix carries the entire experience, right, there is no face to watch while someone pauses, so pacing and clarity matter even more. We edit audio-only shows too, priced on the call, and audio only podcast editing goes through what a proper audio edit includes.
In what order should mixing and mastering happen?
Edit first, then clean and balance the mix, then add music and effects, and only then master the finished file for loudness, peaks and format, because each step depends on the one before it being finished.
The reason the order matters so much is that every edit changes the mix. If you master an episode and then cut four minutes out of the middle, the loudness measured across the programme can shift, and the music bed you ducked now ducks under nothing. That sounds obvious, but I see it constantly in shows where the edit and the audio are done by different people at different times, you see the problem, right, the last person to touch the file did not know what the first person decided.
Can an automatic tool mix and master a podcast for me?
An automatic tool can level and master a reasonably well recorded episode very well, and for a single microphone show with clean audio and a tight budget it may be all you need, but it cannot decide what is noise and what is conversation.
I want to be fair here, right, because automated loudness and levelling tools are genuinely good at what they do, and if you record solo on a decent microphone in a quiet room, a tool is often the better choice for you on cost alone. Where they struggle is judgment, the guest whose chair creaks every time they make a point, the moment both people laugh and one of them is three times louder, the music cue that should dip for a quieter line. I compared the two properly in Auphonic vs human audio mixing, and the short version is that the tool handles the numbers and a person handles the decisions.
At the other end, if you make a narrative or drama podcast with layered sound design, scored music and dozens of effects per scene, a dedicated audio post engineer is the better hire, because that is a different discipline from conversational editing. Most interview and business shows sit in the middle, and that middle is where we work.
What does podcast mixing and mastering cost with you?
At Pixel Samy Podcast the mix and the master are not a separate line item, they are part of every episode, so a single camera show of up to an hour at a fixed 199 dollars already includes the levelling, the noise cleanup and the master file for your host.
The other lengths, the multi-camera plan and the pack and monthly discounts are all on pricing, so I will not repeat the grid here. What matters for audio is the review loop, right, your first cut arrives with the mix already in it, within 48 hours of the upload, and the single round of notes you send afterwards is exactly where you tell us a voice still feels low or a music bed sits too loud, with the corrected version back a business day later. The session stays yours as well, so if you ever move to another editor the mix moves with you instead of being rebuilt from scratch.
If you are not sure whether your problem is a mix problem or a recording problem, the worst podcast audio I have fixed is a good honest look at the limits, and the podcast microphone guide covers what to change before the next recording. To be very honest, the cheapest audio fix is almost always at the microphone, and the edit can only do so much with what it is given.
How do I check the audio I get back?
Listen to the first five minutes on earbuds, then on a phone speaker and in a car if you can, and check the file's integrated loudness and true peak against Apple's published reference before it goes into your feed.
If you want me to listen to your current episodes and tell you which side the problem sits on, a 15 minute call is enough for that and for a quote, or you can write to samy@pixelsamy.com. But at the end of the day the meters only confirm what your ears already told you, so do the thumb test yourself, really really honestly, write down every moment you wanted to reach for the volume, and send us that list as your notes, because an episode nobody has to touch from start to finish is the entire point of mixing and mastering.
So yeah. That's my way of saying it.