How to Record and Edit a Podcast: The Complete Guide

Listeners forgive a lot. They will forgive a rambling introduction, an awkward interview question, and a joke that does not land. What they will not forgive is audio that is unpleasant to listen to. Harsh, echoing, hissy, or wildly uneven sound makes people leave within the first minute, and no amount of good content rescues it, because the listener is usually wearing headphones on a train and simply cannot make out what you said.

The good news is that most of the quality gap between an amateur recording and a professional one comes from a handful of decisions that cost nothing: where you sit, how far you are from the microphone, how loud you record, and how carefully you cut. This guide covers the entire chain, from choosing a room to exporting the final file, with specific numbers rather than vague advice. It applies equally to a solo show, a two person conversation, a remote interview, and an internal company podcast nobody outside the building will ever hear.

Plan the Episode Before You Press Record

Editing time is decided before recording starts. A structured conversation takes twenty minutes to edit. An unstructured one takes three hours, because you are looking for the episode inside the recording rather than trimming an episode that already exists.

A workable plan is short. Decide the one idea a listener should leave with. Write three to five beats that carry them there. Write the first sentence out in full, because openings are where people freeze, and write the closing line too, because endings are where people ramble. Everything in between can be bullet points.

For interviews, send the guest the themes rather than the exact questions. Themes produce thoughtful, natural answers. Exact questions produce rehearsed ones, and rehearsed answers are noticeably flatter. Agree the running time in advance so nobody is checking a clock mid sentence, and tell the guest explicitly that mistakes are fine because you will edit them out. That one sentence relaxes people more than anything else you can say.

Finally, decide the format before recording, not after. A tight solo episode, a long interview, and a panel discussion require different pacing, and trying to convert one into another in the edit rarely works.

Recording: Room, Microphone, and Levels

This section contains the largest quality wins available, and none of them require spending money.

The room beats the microphone

When you speak, sound travels to the microphone directly and also bounces off every hard surface and arrives slightly later. Those reflections are what make a recording sound distant, boxy, or like it was made in a bathroom. Reflections cannot be removed afterwards in any convincing way, because the echo is baked into the same waveform as your voice.

Soft, small, and cluttered is what you want. A bedroom with a bed, curtains, and a rug is excellent. A carpeted room with a full bookshelf is excellent. A kitchen, a bathroom, an empty office, or a room with a glass table and bare walls is terrible. If your options are limited, record while sitting in a wardrobe surrounded by hanging clothes, which sounds ridiculous and produces genuinely professional results. Throwing a duvet over a clothes airer behind you achieves most of the same effect.

Microphone choice, briefly

Dynamic microphones pick up less of the room and are more forgiving of imperfect spaces, which is why broadcast studios use them. Condenser microphones are more detailed and more sensitive, which means they capture your voice beautifully and also the neighbour's lawnmower. For untreated home rooms, dynamic is the safer choice.

Beyond that, spend as little as you need. The difference between a modest microphone and an expensive one is real but small compared with the difference between a good room and a bad one, or between correct and incorrect distance. Even a decent pair of wired earphones with an inline microphone, kept close to the mouth, will beat a laptop built in microphone by a wide margin, because the laptop microphone is far away and pointed at the room.

Distance and angle

Sit roughly one hand span from the microphone, about 15 to 20 cm. Closer produces a fuller, more intimate sound and less room, but risks plosives and heavy proximity effect. Further away lets the room back in, which is exactly what you are trying to avoid.

Speak slightly across the microphone rather than straight into it, so the blast of air from p and b sounds passes by rather than striking the capsule. A pop filter helps, and so does a sock over the microphone in an emergency, but angle alone solves most of it.

Setting levels

Digital audio has a hard ceiling at 0 dBFS. Anything louder is clipped, which is permanent distortion that cannot be repaired. So you record with headroom: speak at your normal presentation volume, including the loudest moment you expect, and set the input gain so peaks land between minus 12 and minus 6 dBFS. The signal will look quieter than you expect, and that is correct, because loudness is set later in a controlled way.

Always wear headphones while recording. Without them you cannot hear a cable buzz, a fan hum, a phone vibrating on the desk, or a guest whose microphone has quietly switched to the wrong device. Fixing those during the recording takes seconds. Discovering them afterwards means the recording is compromised.

Recording Remote Guests Properly

Remote interviews fail in predictable ways, and all of them are avoidable with two habits.

The first is to record each side locally. Call audio is compressed heavily for real time transmission, and it drops out when a connection wobbles. A local recording on each participant's own device captures full quality regardless of the call, and you combine the tracks afterwards. Ask the guest to record a voice memo on their phone held near their mouth if nothing else is available, which is a surprisingly good backup.

The second is to keep the call recording anyway as a safety net. It will sound worse, but it is a complete recording of the conversation, and it will save an episode when a guest forgets to press record or their file is corrupted.

To synchronise the tracks later, have everyone clap once at the start. The spike in the waveform is unmistakable and lines the tracks up in seconds. Ask the guest to wear headphones, without exception, because otherwise your voice plays out of their speakers and back into their microphone, producing an echo of you inside their track that cannot be removed.

Before the real conversation begins, record thirty seconds of small talk and listen back. That test catches wrong input devices, weak levels, background noise, and echo, all of which are trivial to fix in the first minute and impossible to fix in the last.

Editing: Trimming, Joining, and Pacing

Editing is mostly subtraction. The best episodes are not the ones with the cleverest processing, they are the ones where nothing is wasted.

Cut the top and tail first

Every raw recording starts with setup chatter and ends with a slow wind down. Removing both takes seconds and improves the episode more than any effect. Podcast openings in particular should reach the actual subject fast, because the first thirty seconds decide whether the listener stays.

Trimming to the exact section you want is the single most common audio task there is, and it does not require a full editing suite. A browser based tool to trim an audio file shows the waveform, lets you set the start and end points precisely, and exports just that section as MP3 or WAV, with the file never leaving your computer.

Cut tangents without mercy

Conversations wander. A three minute detour that felt engaging in the room is usually the part where listeners stop. Ask of every section whether it advances the point of the episode. If it does not, delete it, even if it was interesting, and especially if you personally enjoyed it.

Edit pauses carefully

Removing every gap makes speech sound unnatural and airless, like a machine reading a script. Removing long ones tightens the pacing considerably. A practical rule is to shorten pauses longer than about a second, leave the shorter ones alone, and always keep the breath before an important point, because that breath is what makes it land.

Filler words are the same story. Cutting every "um" produces something that feels sterile and slightly unsettling. Cutting the clusters, where three appear in a row while somebody thinks, does the work without stripping the personality out of the voice.

Joining segments

Most episodes are assembled from parts: an intro recorded separately, the main conversation, an advertisement or announcement, and an outro. Joining them is straightforward, but two details matter. Add a short fade of a few milliseconds at each join so the transition does not click, and make sure the segments are at similar loudness before joining, or the listener will be reaching for the volume control halfway through.

When you have separate files to assemble, a tool that lets you merge audio files into one track in whatever order you choose handles the assembly without a project file or a timeline, which is often all a straightforward episode needs.

Editing a two person conversation

When you have separate tracks per person, you can edit them independently, which is a large advantage. Silence the guest track while the host talks, and vice versa, and the background noise from each room disappears during the other person's speech. Leave overlaps intact where people talk over each other naturally, because removing those makes the conversation sound artificial.

Cleaning Up the Sound

Processing should be light. Heavy handed correction is more noticeable than the problem it was correcting.

Noise reduction handles steady sounds well: fan hum, air conditioning, computer noise, mains hum. It handles irregular sounds badly. Push it too far and voices develop a watery, underwater quality that is far more distracting than the original hiss. Apply the smallest amount that makes the noise unobtrusive, then stop. It is not a substitute for turning the fan off before recording.

A high pass filter is the most useful and least risky processing available. Rolling off everything below roughly 80 Hz removes rumble, desk bumps, traffic, and plosive energy without touching the voice, since almost nothing in speech lives down there.

Gentle compression reduces the distance between the loudest and quietest moments, which is what makes professional voices sound consistent and easy to hear in a car. Modest settings do the job. Extreme compression pulls up every breath, every mouth click, and every trace of room noise between words.

Equalisation should be corrective rather than decorative. A small cut in the low midrange can clear boxiness, and a gentle lift in the upper midrange can add clarity, but large boosts usually mean the underlying problem is placement or the room, which no equaliser can undo.

Loudness, Levels, and a Clean Mix

Loudness is where amateur episodes give themselves away. Two guests at noticeably different volumes, or an episode far quieter than everything else in the listener's feed, both read instantly as unfinished.

Peak level and perceived loudness are different things

Peak level is the highest instantaneous value in the file. Perceived loudness is how loud it actually feels over time, and it is measured in LUFS. Two files can peak at exactly the same value while one sounds dramatically louder, which is why matching peaks does not make episodes consistent.

The widely used targets for spoken word are approximately minus 16 LUFS integrated for stereo files and minus 19 LUFS for mono, with true peaks kept below minus 1 dBTP so that lossy encoding does not push anything into distortion. Hitting that range means your show sits comfortably alongside professionally produced ones rather than forcing listeners to adjust the volume.

Balance the voices first, then the episode

Work in the right order. Match each speaker to a similar level, then match any music or clips to the speech, then normalize the finished episode to the target. Normalizing first and balancing afterwards undoes the work you just did.

For the practical version of this, a tool that can boost, reduce, or normalize an audio track to a standard loudness in one step handles the common cases: a guest recorded too quietly, an intro that is jarringly loud, or a finished episode that needs to sit at a consistent level before publishing.

Do not chase maximum loudness

Pushing an episode as loud as it will go squashes the dynamics until everything sits at the same intensity, which is exhausting over an hour. Speech needs light and shade. Emphasis only works if there is something quieter to contrast against it.

Music, Intros, and Transitions

Music sets expectations in seconds, which makes it powerful and easy to overuse. A few practical guidelines keep it working for the show rather than against it.

Keep the theme short. Ten to fifteen seconds is plenty, and regular listeners will skip anything longer. Use the same theme every episode, because familiarity is the point. Mix it clearly below the voice when the two overlap, and duck it further under speech rather than fighting for the same space.

Licensing is not optional. Using commercial music without a licence risks the episode being pulled, monetisation removed, or a legal complaint arriving. Royalty free libraries and Creative Commons sources both work, but read the terms, because many require attribution or exclude commercial use.

For transitions between segments, restraint wins. A brief musical sting or a short pause is enough to signal a change of subject. Elaborate sound design tends to date quickly and pulls attention away from the conversation.

Exporting and Compressing the Episode

The export settings determine how quickly the episode downloads, how much bandwidth it consumes, and whether every app can play it.

Choosing a bitrate

For mono speech, 96 to 128 kbps is the practical range and sounds fine. For stereo with music, 128 to 192 kbps is sensible. Going higher inflates the file with no audible benefit for spoken word, and every extra megabyte is bandwidth your host charges for and download time your listener waits through on a phone connection.

The arithmetic is easy: a 45 minute mono episode at 128 kbps is about 43 MB, and the same episode at 320 kbps is about 108 MB. The second file is two and a half times the size and sounds the same to a listener on earphones.

Converting between formats

You will need conversion regularly: a guest sends an M4A voice memo, a music bed arrives as FLAC, a host requires MP3. Converting between MP3, WAV, OGG, M4A, and FLAC with an audio converter that runs in your browser keeps the files on your machine and avoids uploading unreleased episodes to a service you have not vetted.

One rule protects you from slow damage: always keep an uncompressed master. Editing an MP3 and exporting it as MP3 again re-compresses already compressed audio, and repeating that over a season leaves a noticeably degraded archive.

Publishing, Metadata, and Show Notes

The episode file is only part of what gets published, and the text around it does most of the work of being found.

Episode titles should describe the content, not amuse the host. A listener scanning a feed decides in about two seconds, and an inside joke gives them nothing to decide with. Put the specific subject early in the title.

Show notesdeserve real effort. They are the only text a search engine can read, since audio is opaque. A useful set of notes includes a short summary, the main points with rough timestamps, links to anything mentioned, and the guest's details. Chapters, where your host supports them, let listeners skip to what they want and measurably improve completion rates.

File metadatashould be filled in before upload: title, artist or show name, album for the series, track number, year, and cover art. Some players display these instead of the feed information, and an episode showing as "track 3" in somebody's library looks unfinished.

Transcripts are the most underrated asset in podcasting. They make the show accessible to deaf and hard of hearing listeners, they give search engines a full page of relevant text to index, and they let you pull quotes for social posts without listening back. Automatic transcription has become accurate enough that a quick manual pass is usually all that is needed.

Choosing a Setup on Three Budgets

Equipment questions dominate podcasting forums and matter far less than the room and the placement, but the question is reasonable, so here is an honest answer at three levels.

Nothing spent

Wired earphones with an inline microphone, kept close to the mouth, recording into a phone voice memo app in a bedroom. This sounds better than most people expect and better than a laptop built in microphone in a good room, because proximity beats hardware. Avoid wireless earbuds for recording: when they are used as a microphone, the audio drops to a low quality call mode that sounds thin and compressed.

Modest budget

A USB dynamic microphone on a small desk stand or boom arm, plus closed back headphones. This is the level where the equipment stops being the limiting factor. A boom arm matters more than it sounds like it should, because it lets you position the microphone correctly and keeps desk vibration out of the recording. Add a pop filter if plosives persist after adjusting the angle.

Serious setup

An audio interface or a dedicated recorder with XLR microphones, one per person, recording separate tracks. The advantage is not really sound quality, it is control: separate tracks mean you can fix one speaker without touching the other, silence a guest's background noise while the host talks, and rescue an episode where one person was recorded badly. Add basic acoustic panels behind and beside the recording position before adding anything else.

At every level, spend on the room before spending on the next microphone. A cheap microphone in a treated space beats an expensive one in a bare one, and this remains true no matter how much money is involved.

Interviewing: Getting Better Material to Edit

The edit can only work with what was recorded. Better interviewing produces less editing, and the techniques are simple enough to learn in one session.

Ask one question at a time.Stacked questions produce answers to whichever part the guest remembers, usually the last one, and the interesting first part is lost. Short questions also edit cleanly, whereas a rambling question cannot be cut without cutting the answer's context.

Leave silence after the answer. Most people stop talking before they have finished thinking. A two second pause is uncomfortable in the room and regularly produces the best line in the episode, because the guest fills it with something they had not planned to say.

Ask for the specific instead of the general. "What is your advice for new founders" produces platitudes. "What did you do in the week you nearly ran out of money" produces a story. Stories are what listeners remember and what makes clips worth sharing.

Do not react vocally.Saying "mmm" and "right" over a guest's answer is a natural conversational habit and a nightmare in the edit, especially on a single shared microphone, because your reactions are permanently mixed into their sentences. Nod instead. Guests adapt to this within a minute.

Let people restart. When a guest stumbles, tell them to take the sentence again from the beginning rather than repairing it mid flow. A clean second attempt cuts invisibly. A repaired sentence usually does not.

Record a wrap up question. Ending every interview with the same closing question gives you a consistent structural ending, which makes the outro easy to assemble and gives the show a recognisable shape.

Hosting, RSS, and How Distribution Actually Works

Podcasting is unusual among modern media because it still runs on an open standard. Understanding the plumbing removes most of the confusion about where episodes go.

You upload the audio file to a podcast host. The host stores the file and generates an RSS feed, which is a text document listing your show details and every episode with a link to its audio file. You submit that feed address once to each directory, such as the major podcast apps. Those directories do not store your audio, they read your feed and point listeners at the file on your host. That is why a new episode appears everywhere shortly after you publish it, and why moving hosts requires a redirect rather than resubmitting the show.

Several consequences follow from that design. Your feed is the show, so protect the account that controls it. Changing the file after publishing is possible but messy, since apps may have cached the old version. Bandwidth is billed by downloads multiplied by file size, which is the practical reason to export at a sensible bitrate rather than the highest one available. And download statistics are inherently approximate, because a download is a request for a file, not proof that a human listened to it.

Two settings on the host repay attention. Fill in the show level metadata carefully, since the title, description, category, and artwork are what people see when browsing a directory. And set the episode numbering and season fields correctly, because apps use them to decide play order, and a show that plays in the wrong order confuses new listeners immediately.

Repurposing an Episode Into More Content

A recorded conversation is raw material for far more than one file. The marginal cost of the extra assets is small because the hard part, getting people to say interesting things, is already done.

• Audiograms. A sixty second clip of the best moment, with captions, works well on social platforms where audio is often muted.
• Quote graphics. Pull three strong lines from the transcript and set them as images.
• A written article. The transcript, restructured and edited, becomes a post that ranks for the topic and links back to the episode.
• A newsletter section. A short summary with a link converts subscribers who never open podcast apps.
• A clips feed. Individual answers published as short standalone episodes serve people who will not commit to an hour.

The workflow is consistent: trim the segment, level it to match, export at the right format for the destination, then write the text around it.

Solo Episodes: Recording Yourself Without a Guest

Solo recording is harder than interviewing, which surprises people who assume the opposite. There is no second person to react to, no natural rhythm to follow, and nothing to stop you drifting into a monotone. A few adjustments fix most of it.

Do not read a full script unless you can read well. Written sentences are longer and more formal than spoken ones, and most people read them flatly. Bullet points with a fully written opening and closing give you structure without the flatness. If you do script, read it aloud twice before recording and rewrite anything you stumble on, because a sentence that is awkward to say is awkward to hear.

Talk to one specific person. Picture an actual listener, or record as though explaining it to a friend who asked. Addressing a crowd produces a presenting voice that sounds distant. Addressing one person produces the conversational tone that works in headphones.

Stand up, or at least sit forward. Posture changes breath support audibly, and slumping in a chair produces a quieter, duller sound. Keeping water nearby matters too, since a dry mouth creates clicks that are tedious to edit out.

Use the punch in technique. When you make a mistake, pause for two seconds, then take the sentence again from the start. The silence gives you an obvious visual marker in the waveform, and the retake cuts cleanly against it. This is faster than trying to fix a stumble in the moment and produces a far better episode.

Record the introduction last. Once the episode exists you know exactly what it contains, which makes the opening specific rather than vague. Vague openings are the most common reason listeners leave a solo show in the first minute.

Common Problems and How to Fix Them

The recording sounds echoey or distant

The room, the distance, or both. Move closer to the microphone and record somewhere softer. This cannot be repaired convincingly afterwards, which is why it is worth solving before you record anything you intend to publish.

There is a constant hiss or hum

Hiss usually means the gain was too low and the recording had to be raised afterwards, amplifying the noise floor along with the voice. Hum usually means an electrical issue, a cable running alongside a power lead, or a laptop charger. Try recording on battery power, move cables apart, and set a healthier input level next time.

Plosives thump on p and b sounds

Air hitting the capsule directly. Speak slightly across the microphone, add a pop filter, and increase the distance a little. A high pass filter reduces what remains but cannot fix a heavy thump.

One speaker is much quieter than the other

Balance the tracks separately before combining them, then normalize the finished episode. Where a track was recorded too quietly, raising it will also raise its background noise, so a small amount of noise reduction afterwards may be needed.

The audio is distorted and crackly

The recording clipped, meaning the signal exceeded the digital ceiling and the peaks were cut off. This is permanent. Reduce input gain and re-record if the material allows it. Prevention is the only real fix, which is what headroom is for.

The file is enormous

You are probably publishing a WAV or an unnecessarily high bitrate MP3, or exporting speech in stereo. Mono at 96 to 128 kbps is the standard answer and cuts the size dramatically without a listener noticing.

An Episode Checklist

1. Plan the single idea and the three to five beats that support it.
2. Choose a soft, small room and turn off fans, air conditioning, and notifications.
3. Set the microphone one hand span away, slightly off axis, with headphones on.
4. Record a thirty second test and actually listen back to it.
5. Set peaks between minus 12 and minus 6 dBFS and record to WAV.
6. Record locally on each side of a remote call and clap once for sync.
7. Trim the top and tail, then cut tangents and long pauses.
8. Apply a high pass filter, light noise reduction, and gentle compression.
9. Balance speakers, then normalize the episode toward minus 16 LUFS stereo.
10. Export MP3 at 96 to 128 kbps mono, keeping the WAV master.
11. Fill in metadata and cover art before upload.
12. Write show notes with timestamps, links, and a transcript.

Frequently Asked Questions

What microphone do I need to start a podcast?

Any dynamic microphone with a USB connection or an audio interface is enough to begin. Placement and room treatment affect the result far more than price does, and a modest microphone in a soft room beats an expensive one in an echoing kitchen every time.

What loudness should a podcast be?

Approximately minus 16 LUFS integrated for stereo and minus 19 LUFS for mono, with true peaks below minus 1 dBTP. That range keeps your episode consistent with other shows so listeners are not adjusting the volume between them.

Should I record in mono or stereo?

Mono for speech. A single voice carries no stereo information, so mono halves the file size with no audible loss. Reserve stereo for shows where music or spatial sound design genuinely matters.

What bitrate should I export at?

96 to 128 kbps for mono speech and 128 to 192 kbps for stereo with music. Higher settings make downloads slower and hosting more expensive without any improvement a listener can hear on spoken word.

How do I remove background noise?

Fix the source first: turn off fans, close windows, and move closer to the microphone. Then apply the lightest noise reduction that makes the remaining hum unobtrusive. Heavy processing produces a watery voice that is worse than the noise it removed.

How long should an episode be?

Long enough to cover the subject and no longer. Interview shows commonly run 30 to 60 minutes and solo shows 15 to 30, but a tight 22 minute episode outperforms a padded 50 minute one, because completion rate matters more than duration.

Can I edit a podcast without installing software?

For the common tasks, yes. Trimming, joining, adjusting volume, and converting formats all run in a browser, which has the added benefit of keeping unreleased audio on your own device rather than uploading it to a server.

What format should I publish in?

MP3, because every podcast app and player supports it without exception. Keep a WAV master for archiving and future re-editing, and publish the MP3 version.

Do I need a transcript?

You do not need one to publish, but it is the highest value optional asset available. It makes the show accessible, gives search engines text to index, and provides ready made material for articles, quotes, and social posts.

Bringing It Together

Good podcast audio is not the product of expensive equipment. It is the product of a soft room, a microphone close to the mouth, a level with headroom, an edit that removes everything unnecessary, and an export at a sensible loudness and bitrate. Every one of those is free, and together they account for most of the distance between a recording that sounds amateur and one that sounds professional.

Start with the parts that cannot be fixed later. Get the room and the placement right, record with headroom, and keep the uncompressed master. Everything after that is adjustable, repeatable, and quick, which means you can spend your effort on the only thing listeners actually came for, which is the conversation itself.