Podcast editing work for beginners is real, paid, and reachable inside a month, but it isn’t the same job as knowing your way around Audacity. Entry-level editors in India charge roughly 500 to 1,500 rupees a finished episode, and US beginners charge 25 to 50 dollars per finished hour of audio, on Fueler’s India-versus-US breakdown of 24 October 2025. The gap between an editor who gets rehired and one who doesn’t is rarely talent.
This article sets out the five-pass workflow, three real podcast editing problems worked end to end, and how the work is priced and won.
What clients are buying is a file that hits the platform loudness spec, arrives on the day it was promised, and needs no second pass from the host. None of that requires a studio, a degree, or a mentor.
It requires one editor you know cold, one loudness tool, and the discipline to measure what you send instead of trusting your ears. But there’s a fourth requirement, and it’s the one beginners skip: enough diagnostic vocabulary to name a problem before you fix it. Most people starting out can hear that something is wrong without being able to say what, which turns every job into guesswork and every quote into a nervous underbid. The three worked cases below exist to close that gap.
Set up your podcast editing workflow
Your podcast editing workflow is five passes run in a fixed order, on one editor plus one loudness tool, and the order matters more than the software does.
Pass one is intake: import the files, confirm the sample rate on every track, align them if there was more than one mic, and label every speaker before you touch a waveform. Pass two is cleanup, meaning hum, clicks, room noise and the worst of the breaths. Pass three is the content edit, which is the pass clients are genuinely paying for.
Passes four and five are sound design and the master, in that order, and neither one starts before pass three is signed off. Here’s why that sequence isn’t negotiable.
Drop a music bed under a section before the content edit is locked, and every cut you make afterwards shifts the music against the speech, which means rebuilding the bed from scratch. The same logic binds at the other end, because removing four minutes of quiet false starts changes the integrated loudness of the whole file.
Pass three is also where beginners are least consistent, so work to a written rule rather than instinct. Four lines are enough, and they belong in your project notes where you can re-read them at the halfway mark of every episode.
Cut false starts and repeated sentences, and cut filler only where it stacks (three “you knows” in a row, not one). Cut any tangent that doesn’t return to the question within 30 seconds. Never cut a pause the listener needs to absorb a number. Keep laughter, and keep the host’s mistakes when the guest reacts to them.
With that written down, minute 40 gets treated the same way as minute four. And predictability across a season is what quietly converts a one-off job into a retainer.
So which software should you actually learn? Audacity is free and open source and will carry you through your first paid episodes without a rupee spent, which makes it an honest starting point rather than the compromise it usually gets presented as.
Learn one editor properly before you go looking at a second.
REAPER gives you 60 days of full functionality free, then costs 60 dollars for the discounted licence, which covers personal use or commercial use where yearly gross revenue stays under 20,000 dollars. The full commercial licence is 225 dollars. For a beginner that means a fully licensed professional editor for the price of a month’s groceries.
Descript takes a different approach with text-based editing, where you delete words from the transcript and the audio follows, which most beginners find faster than learning to read a waveform.
The free plan gives you 1 hour of media a month, Hobbyist is 24 dollars a month (16 if billed yearly) for 10 hours, and Creator is 35 dollars a month (24 billed yearly) for 30 hours. Pick it if you think in sentences rather than in shapes.
Two free tools carry most of the cleanup and mastering load, and between them they cover what a beginner can’t yet do by ear.
Adobe Podcast’s Enhance Speech free tier accepts files up to 30 minutes and 500 MB, one at a time, with a one-hour daily ceiling and no strength control. That’s enough to rescue a badly recorded guest track, though not enough to run a whole season through.
Auphonic is free for 2 hours of processed audio a month and applies loudness normalisation for you, so your very first delivery can hit spec before you understand the maths behind it. Two hours covers roughly two episodes, which is enough to learn the target by ear against the meter. All these prices and limits are current at the time of writing in August 2026, and every one of them is the vendor’s to change without notice.
And normalisation is where amateur files get caught, because it’s the one part of this job that is measured rather than judged.
Apple Podcasts asks creators to precondition audio so overall loudness sits around -16 dB LKFS with a plus-or-minus 1 dB tolerance, and so the true-peak value doesn’t exceed -1 dB FS. Both are calculated per the ITU-R BS.1770-5 recommendation.
That preconditioning has to happen before encoding, because compression algorithms don’t change loudness and will clip a signal that arrives too hot. Export stereo MP3 at 128 to 256 kbps, mono at 96 to 128, both at 44.1 to 48 kHz. Spotify normalises playback to -14 dB LUFS under the same standard.
The practical reality is that this one number, -16 LKFS, is the fastest credibility signal a beginner has, and almost nobody applying for entry-level work mentions it.
Put your measured integrated loudness and true peak into the delivery message of every episode you send, even when nobody asked. It tells a client you work to a standard rather than to taste.
If you write for shows as well as cutting them, our guide to writing podcast scripts covers the other half of that workflow.
[INFOGRAPHIC-01]
Loudness and true peak are calculated per the ITU-R BS.1770-5 recommendation, and the preconditioning has to happen before encoding. Compression algorithms do not change loudness, and they will clip a signal that arrives too hot.
Three podcast editing problems and their fixes
Three problems account for most of what actually lands in a beginner’s inbox, and each one looks unfixable until you can name the cause. What follows is each of them worked end to end: the symptoms as the client describes them, the diagnosis, the fix in order, and what you write back afterwards.
Read them as templates rather than as anecdotes.
The value sits in the sequence, not in the specific numbers, and the sequence is what you’ll reuse on files that look nothing like these.
Case one: the untreated room and the level gap. The brief arrives as “the audio sounds cheap and my guest is too quiet”, which is two separate faults the client has bundled into a single complaint. Open the file and you’ll usually find a low continuous buzz under everything, a guest track sitting well below the host, and speech that sounds like it was recorded in a stairwell.
Most beginners reach for a noise gate at this point, which chops the tails off words and makes the result sound worse than the original. But the buzz is the first thing to diagnose, and you diagnose it by frequency rather than by ear.
Electromagnetic interference picked up on an unbalanced cable sits at the local mains frequency, which is 50 Hz in India and 60 Hz in the United States.
So an Indian editor working a US client’s file who reflexively notches 50 Hz will find nothing there and conclude the hum is unfixable. It isn’t. It’s simply sitting ten hertz higher than the one place they looked.
Acoustic hum radiating from a transformer or an appliance behaves differently again, because the fundamental is twice the line frequency: 100 Hz on Indian power, 120 Hz on American, with harmonics stacked above it. Work out which of the two you have before you touch a single control, because the fixes land in different places.
The good news is that a notch at 50 or 60 Hz does almost no collateral damage, because it sits below every adult speaking voice.
Measured speaking fundamentals put adult men at a mean of 116 Hz with a range of 93 to 135 Hz, and adult women at a mean of 205 Hz across a range of 162 to 238 Hz, on the Holmberg, Hillman and Perkell figures.
A narrow notch at the line frequency plus its first two harmonics, followed by a high-pass filter rolling off below 80 Hz, removes both the buzz and the desk rumble. No voice in the room gets thinned in the process, which is exactly why this is the first move rather than the last resort.
Fix the level gap next, and fix it before you go anywhere near the reverb.
Auphonic’s adaptive levelling or Descript’s Studio Sound will pull the guest up to match the host across the whole episode rather than at one fixed gain, which matters because the gap widens and narrows as a guest leans in and out of the mic.
Only once the levels sit together should you run the room sound through Enhance Speech, since de-reverb algorithms behave differently on a quiet track than on a levelled one. Finish with a de-esser if the enhancement has pushed sibilance forward. The hiss on S and F sounds lives roughly between 5 and 10 kHz, and enhancement tools routinely exaggerate it.
Then tell the client what you did, in their language rather than in yours.
Two separate problems in that file. The buzz was 60 Hz mains interference on the guest’s cable, notched out at 60, 120 and 180 Hz with a high-pass below 80. Your guest was running about 9 dB under you, so I levelled you both across the whole episode instead of boosting one gain and hoping. Delivered at -16.1 LKFS, true peak -1.2 dB FS.
Case two: the double-ender that drifts out of sync. A double-ender is two people recording locally at their own ends of a call, which gives far better quality than recording the call itself. It also produces the single most common panic in a beginner’s first month.
The symptom is that the two tracks line up perfectly at the start and are visibly, audibly apart by the end. The client’s message usually says the recording is ruined.
It almost never is, and the fix takes about fifteen minutes once you’ve done it twice.
But which of the two possible faults are you actually looking at? They look similar on a waveform and have nothing to do with each other, so separating them is the whole job.
A genuine sample-rate mismatch, where a file recorded at 44.1 kHz is being read as 48 kHz, isn’t drift at all. It plays about 8.8 percent fast, which makes an hour-long recording finish roughly five minutes early and lifts the speaker’s pitch by around a semitone and a half. That fault is obvious inside ten seconds, and it’s fixed by re-interpreting the file at its true rate rather than by stretching anything.
True drift is the subtler fault, and it happens when both devices record at the same nominal rate while running on independent clocks that differ by a hair.
The tracks separate slowly, which is why minute one sounds fine and minute fifty doesn’t.
Double Ender, a free macOS utility, describes the mechanism plainly and corrects it by patching or removing areas of silence to offset the accumulated difference. Its more recent versions scan the entire file, so sync errors of thirty minutes or more still get caught.
On Windows or Linux there’s no equivalent one-click tool, so you do it by hand.
Align both tracks on the countdown or clap at the top, then re-align at three or four natural pauses through the episode, slipping each segment into place rather than time-stretching the whole track. And charge for that work as a named line item rather than absorbing it quietly.
A client who never sees the cost assumes the problem didn’t exist, and sends you the same broken setup again next week.
The better approach, in our view, is to fix it once, bill it once, and then tell them how to prevent it. Matching sample rates in both recording apps before they hit record removes the mismatch case entirely, and one loud clap at the top gives you a hard sync point for the drift case.
Case three: one session, two deliverables. More clients now record video alongside audio and expect both back, and this is the job that separates an editor who gets the next season from one who gets a single episode.
The demand behind it is measured rather than assumed.
Edison Research’s Infinite Dial 2026 found that 57 percent of Americans aged 12 and over have both listened to and watched a podcast, and that 58 percent (about 167 million people) consumed one in the last month. A beginner who can only deliver audio is quoting on a shrinking half of the available work.
The trap here is cutting the video first, and almost every beginner falls into it once.
Edit the audio to lock exactly as in passes one to three above, then conform the video to that audio edit, because audio is the thing listeners abandon over and video is the thing they forgive. Cut picture first and every subsequent audio decision either breaks a visual cut or forces you to accept a worse edit to protect one.
You’ll feel that constraint tighten with every minute of the episode, and by the halfway mark you’re editing to defend earlier mistakes.
But none of that matters if the export is wrong, so deliver the video to the platform’s published numbers rather than to whatever your editor chose by default. Spotify’s video specs recommend MP4, 1080p or higher, 16:9, H.264 or H.265, and 25 Mbit per second constant bitrate for a 1080p source. Those five values alone rule out most default presets.
The rest of that spec matters just as much: a frame rate of 24, 25, 30, 50 or 60 fps, roughly one keyframe per second, an audio track as AAC-LC at 128 kbps or higher, and a file under 10 GB and under four hours.
Hitting those numbers takes a deliberate export and a check, which is precisely why video gets quoted separately.
Audio master to Apple’s -16 LKFS spec, plus a 16:9 1080p video cut conformed to the locked audio edit and exported to Spotify’s published video spec (H.264, 25 Mbit CBR, AAC-LC audio at 192 kbps). Audio and video delivered together, 72-hour turnaround from file receipt.
Two deliverables, one line, and a number the client can approve without booking a call.
That last part matters more than the price itself, because every quote that needs a meeting to explain it loses to a quote that doesn’t.
Find and price podcast editing work
The three cases you just read are also your portfolio, which is the part beginners get backwards by waiting for clients before building samples. A podcast editing portfolio is three before-and-after clips, each under 90 seconds, each with one short paragraph naming the fault and the fix.
Those three faults are a deliberately good spread: a cleanup problem, a sync problem, and a multi-format delivery.
A host screening editors is listening for a fix they can hear inside a minute, and a full 40-minute sample proves endurance rather than ears.
Getting raw audio to practise on is easier than it looks. Record a 20-minute conversation with a friend on two separate devices and deliberately keep the mistakes, which hands you a real double-ender and a real level gap out of a single session.
Or pull public-domain and Creative Commons interview audio and treat it as a client file.
Better still, find a local show that publishes unedited and offer to cut one episode free on the condition that you can use it as a sample, because the show is real and the host becomes a nameable reference. The sample note does more work than the audio in almost every case, because it demonstrates diagnosis rather than button-pushing.
Write it the way you’d write a delivery message to a paying client.
Raw file had 50 Hz mains interference on an unbalanced cable and a guest running about 9 dB under the host. Notched 50, 100 and 150 Hz, high-passed below 80, then levelled both speakers across the full episode rather than at a fixed gain. Cut 4 minutes 20 seconds of false starts and one tangent. Delivered at -16.2 LKFS, true peak -1.1 dB FS, 26 hours from file receipt.
Notice what that note doesn’t contain: no adjectives, no claim that you’re passionate about audio, no software brag.
Name the files so a client never has to ask which is which, using something like showname-ep14-BEFORE-60s.mp3 and showname-ep14-AFTER-60s.mp3 every single time. Host them where the client can press play in the browser rather than behind a link that forces a download, since every extra step between a prospect and your work is a step some prospects won’t take.
Build all three while you’re still employed if that’s your situation, because starting a freelance practice without quitting your job is the lower-risk route. And the sequencing advice in how to become an SEO freelancer in India applies almost unchanged here.
At this stage you are not short of ability. You’re short of evidence.
Now, why will anyone hire an editor at all when the tools keep getting easier? Because hosts are drowning in the hours, not in the software.
Alitu’s Independent Podcaster Report 2025, a survey of 558 independent creators, found that 52 percent do everything themselves with no help whatsoever. The same survey found DIY podcasters spending 4 to 8 hours on an average episode, and 45 percent of audio-only creators spending six hours or more against 36 percent of video podcasters.
Those are your clients: people giving up a working day per episode to the part of the job they least enjoy.
The work itself comes from three channels, and beginners usually try only the most crowded one.
Marketplaces like Upwork and Fiverr are worth using for your first two or three paid jobs, because they solve the trust and payment problem for a stranger who has never heard of you. But they’re a poor place to stay, since you’re bidding against a global floor that has nothing to do with your quality. Treat them as a credential factory: three completed jobs with good feedback, then start building elsewhere.
The second channel is the industry jobs board.
Podnews, run by Podnews LLC since 2017, calls itself the podcast industry’s biggest jobs board, lists roles free, promotes them to 33,230 subscribers, and absorbed the old podjobs.net. Its listings skew towards producer and production-coordinator roles rather than pure editing, which is exactly where an editor’s career tends to move next anyway.
But read it weekly even when you’re not applying, because the job descriptions tell you what shows currently pay for.
The third channel is direct outreach and it pays best, so how do you choose which shows to write to? Find the ones with a broken publishing rhythm: 40 episodes, then a six-week gap, then two in one week.
That gap is almost always the host running out of editing time rather than losing interest. Then write to the gap rather than to yourself.
I listened to episodes 38 and 41 of your show. In 41 your guest’s track sits about 8 dB below yours from the 12-minute mark, so anyone on earbuds is riding the volume knob. I’ve cut a 60-second fix and put it here. If it’s useful, I edit at 6,000 rupees an episode with a 48-hour turnaround.
No portfolio link dump and no paragraph about your love of audio. One observation the host already half-suspects, one fix they can hear, one number. That structure works because it demonstrates the service instead of describing it.
On the number itself, price per finished episode rather than per hour once you know your own speed.
Hourly pricing punishes you for getting faster, and hosts budget per episode anyway, so the two of you end up negotiating against a unit neither side wants. Fueler’s October 2025 breakdown puts Indian entry-level editors at 500 to 1,500 rupees a finished episode, mid-level editors with two to five years at 1,500 to 4,500, and experienced editors at 4,500 to 10,000 or more, with hourly work between 1,000 and 3,500 rupees.
The US picture from the same source runs from 25 to 50 dollars per finished hour of audio for beginners, to 50 to 100 for mid-level, to 100 to 200 or more at the top. Per-episode rates there start around 50 dollars for simple edits and reach 500 dollars and up for complex ones.
Read those as two different markets rather than one ladder, because an Indian editor billing a US client is pricing into the second.
ZipRecruiter’s US listing data averaged 31.60 dollars an hour for freelance podcast editors in July 2026, with most listings between 21.39 and 39.66 dollars. Our recommendation for a first international quote is to take the bottom of the US beginner band, quote per episode, and hold it.
Editing, mixing and mastering to Apple’s -16 LKFS spec, one round of revisions, 48-hour turnaround from file receipt: 60 dollars an episode for shows up to 60 minutes raw.
Two clauses in that quote save more money than any rate increase, and both are about definitions rather than numbers.
Define a revision round as one consolidated list of timestamps rather than an open conversation, because “one round” without that definition reliably becomes four. And state that the turnaround clock starts when all files land rather than when the client says the episode is ready. A missing guest track or a forgotten intro read is the most common reason a first delivery slips, and without that clause the delay becomes your fault.
If you’re pitching US clients, the proposal structure US buyers expect is worth reading before your first quote, and reaching international startups directly covers the outreach side of the same problem.
Frequently asked questions
Do you need a degree or certification to get podcast editing work?
No. There’s no licensing body for podcast editing and no credential that hosts check, so what they evaluate is a sample and a delivered file. A sound-engineering diploma helps you learn faster, but three before-and-after clips and a file that measures to spec beat any certificate.
How long does it take to edit one podcast episode as a beginner?
Expect four to six hours for your first few 45-minute episodes, dropping to about 90 minutes once the five passes become automatic. Alitu’s 2025 survey found DIY podcasters averaging 4 to 8 hours an episode. Quote 48 to 72 hours, not 24, until your speed is predictable.
What is the difference between podcast editing and podcast production?
Editing is post-production only: cleanup, cuts, mix, master, export. Production covers the episode before it exists, including guest research, booking, recording supervision, show notes and distribution. Production pays more, but start as an editor: it’s the job clients hand a stranger.
Will AI tools replace podcast editing work?
AI already handles the mechanical passes well, and filler-word removal, transcript-based cutting and speech enhancement are close to solved. What isn’t solved is the judgement in pass three: which tangent earns its 90 seconds, which pause to keep. Learn the tools, charge for the judgement.
This article is for informational and educational purposes only and does not constitute professional, financial, legal, or career advice. Tool prices and platform specifications were verified in August 2026 and change without notice. Confirm current rates and requirements before acting on them.
Allow notifications
