How to record a podcast that clips well
Most unusable clips were lost at the recording, not in the edit. The framing, tracks and guest setups that make vertical clips easy, from the editor's side.
Short answer
A podcast clips well when each speaker has their own camera angle framed with room to crop to vertical, their own audio track, and a sharp, well lit face that sits mid frame rather than at the edge of a wide shot. Remote guests need the same care on their end, which means a real camera, a plain background and a wired connection. Fix these at the recording and almost every moment becomes clippable; skip them and no amount of editing saves the clip.
Most clips that come back cramped, soft or badly cropped were lost at the recording, not in the edit. An editor can cut, reframe and caption almost anything, but they cannot create pixels that were never captured, separate two voices mixed into one file, or move a face recorded at the edge of a wide shot into the middle of a vertical frame.
The fix is a setup built with the crop in mind: one camera angle per speaker with room above the head and beside the shoulders, one audio track per speaker, a sharp face lit from the front, and remote guests held to the same standard on their end. Get those four right and nearly every strong moment in the episode is clippable. Miss them and the editor is choosing between the least bad options.
Where clips are lost: the recording faults an editor cannot fix
The same faults come up on show after show, and none can be repaired afterwards.
One wide shot of the whole table. A 16:9 frame with two people at either end has nothing usable in the middle. Cropped to 9:16 it gives you one person or half of each, so the choice is a static two shot full of table or a punch in so tight it goes soft.
Soft focus or low resolution. A 9:16 crop from a 16:9 master keeps roughly a third of the width. From a 1080p source that is a few hundred pixels being stretched to fill a phone, so anything slightly soft turns to blur.
One mixed audio file. Every fix applies to both speakers at once. A phone buzzing on the guest's desk stays in, because the filter that removes it also strips the host's voice, and the host's laugh over the guest's best line cannot be lifted out.
A remote guest on a laptop webcam. Low resolution, a lens pointed up the nose, a bright window behind. The host's half looks like a studio, the guest's half looks like a video call, and the clip cannot hide the join.
One angle per speaker, framed for the vertical crop
Give every speaker their own camera. Cutback's video podcast blueprint calls two matching single shots the most editable setup for a two person show and says a wide shot on its own limits clipping; the wide is an establishing shot, not clip material.
Frame each single for the crop that is coming, not for the 16:9 master:
- Camera at eye level, lens around the speaker's nose height
- Head in the middle third of the frame horizontally, not off to one side
- Mid chest to just above the head, the solo framing Cutback also recommends
- Clear space above the head and beside both shoulders, because the vertical crop keeps the height and throws away the width
Check the frame against a 9:16 guide before recording; if the guide forces a choice between the forehead and a shoulder, move the camera.
Record at the highest resolution you can store. A 4K master gives a 1080p vertical crop without upscaling, which is why 4K matters even when the full episode goes out at 1080p. How those singles get cut together afterwards is covered in turning horizontal footage into vertical.
Separate audio tracks and what they make possible in the edit
Record each speaker to their own audio track, and if the show is remote, have each participant recorded locally on their own machine. Riverside's guide to multitrack recording covers the setup; the short version is that each participant is captured on a separate track rather than combined into one.
For clips that matters in four ways:
- Noise removal per person. Traffic through the host's window is filtered on the host track only, and the guest's voice never goes through the filter.
- Crosstalk can be cut. A host's "yeah, yeah" over the guest's key line is removed from the host track and the line plays clean.
- Levels match. A quiet guest and a loud host are balanced separately, so the clip does not jump in volume on the cut.
- The transcript follows the speaker. With one voice per track, a host's interjection does not land inside the guest's sentence in the caption file, so the captions on a two speaker clip read as two clean lines rather than one tangled one.
Use a proper microphone per speaker, close to the mouth, kept in the same place in frame. Keep the raw tracks; a single mixed export is a fallback, never the deliverable.
Remote guests: the checklist to send before the call
Send this the day before and ask the guest to run through it fifteen minutes ahead of the call.
- A real camera or a phone on a stand, not the laptop webcam
- If it has to be the laptop, raise it on books so the lens is at eye level
- Face a window or a lamp; never sit with the window behind you
- Plain wall or a tidy shelf behind you, at least a metre back from it
- Wired headphones or earbuds, so the call audio does not bleed into your mic
- An external mic if you have one, close to the mouth
- Wired internet if possible; if not, sit near the router and take other devices off the network
- Other apps closed, especially anything syncing files
- Record on a platform that captures a local file on your machine, and keep the tab open until the upload finishes
Ask for a ten second test clip the day before, while there is still time to fix the setup.
Eyelines, lighting and backgrounds that survive a crop
Vertical is tall and narrow. It magnifies everything about the face and hides almost everything about the room.
Eyelines. A guest looking at a screen well below the camera looks at the floor in a vertical clip, so put the window showing the other person as close to the lens as possible. In a studio, Cutback's guide has speakers looking just off camera towards each other, which reads as conversation.
Lighting. Cutback recommends a lighting triangle: a key light at 30 to 45 degrees from the face, a soft fill on the other side and a subtle backlight, with the note that flat, even lighting is easier to work with in post. Hard side light leaves half the face dark; overhead room light alone puts shadows under the eyes.
Backgrounds. In a vertical crop the background is a narrow strip either side of the head, so it needs to be quiet. A plain wall or a shelf, slightly out of focus, is enough. Avoid a bright window, a busy poster, or a strong straight line that will cut through the head once the crop centres the speaker.
Recording habits that create clippable moments
Good footage still needs something worth clipping in it. What to agree before the episode is covered in the podcast to shorts workflow; these are the habits during it.
Let answers finish. The strongest line in an answer often arrives thirty seconds in. If the host jumps in with the next question, the clip has no ending.
Say the name, not the pronoun. A guest who says "it" and "they" for a full minute leaves a clip nobody can follow cold. When an answer gets going, have the host restate the product, the person or the figure, so the clip can open at that line without the setup.
Leave a beat after a strong line. A second of silence gives the editor a clean out point.
What to export and keep after the recording
Hand over source material, not a polished master: each camera angle and each speaker's audio as its own full resolution file, with the remote guests' local recordings alongside. The complete handover, from brand assets to the brief, is in what to send a clipper before the first episode, so this section only covers what belongs to the recording day.
Write down anything unusual while it is fresh: a camera that reset mid episode, a guest who shifted out of frame after the break, a name that was mispronounced and corrected, a section the guest asked to keep off social. Give the editor the whole episode, not a highlight reel trimmed in advance, which strips the context needed to find the moments that were not obvious on the day. Keep everything until the clips are live and approved; the line that needs recutting is always the one whose source was deleted.
VALORAE Arc works from the other end of this. We take the full episode, watch it, and return edited, captioned vertical clips distributed to TikTok, Instagram Reels, YouTube Shorts, X and LinkedIn. Recordings that arrive as matched singles with separate tracks and sharp faces are the ones where every strong moment is on the table, and a setup check before the next recording is a normal thing to bring to the starter call at https://calendly.com/valoraearc/valorae-arc-starter-call.
Frequently asked questions
Can you clip a podcast recorded with one wide camera?
Yes, with limits. The editor punches in on each speaker digitally, which only holds up if the source is 4K and the faces are sharp. From a 1080p wide shot the crop goes soft and the options shrink to a static two shot with a lot of table in it.
Do we need separate audio tracks if everyone is in the same room?
Yes. Each mic still favours its own speaker, so noise, crosstalk and levels can be handled per person. One mixed file removes that option for the whole episode, and it is the mixed file that makes a host's laugh over the guest's best line impossible to remove.
What resolution should we record at for clips?
4K if you can store it. A 9:16 crop from a 16:9 master keeps roughly a third of the width, so a 4K source gives a 1080p vertical clip without upscaling. A 1080p source works when the framing is already tight on the face.
Should a remote guest record on a phone instead of a laptop?
Usually the phone gives a sharper image than a laptop webcam, provided it is on a stand at eye level. The catch is the workflow: the guest still needs a local recording that reaches you, so either run the call on the phone through a platform that records locally, or use the phone as a webcam for the laptop.
Want your podcast turned into clips that hold?
We cut, caption and distribute short-form for podcasts and founders. Bring one episode and we will show you what comes out of it.
Book a call