Blog / Long-form and editing
How to edit a podcast video for YouTube
There is a mistake I see often in first podcast edits. It is cut like a video. The camera changes are careful, the titles are tidy, and the sound is an afterthought.
On YouTube that is backwards. YouTube's own page on creating a podcast in YouTube Studio says podcast content may be eligible for YouTube Music, where people can listen in the background and download it. So the same episode is being watched by someone who found it in search and listened to by someone walking the dog with their phone in their pocket. The listener never sees your camera changes. They hear every bad level.
This is the order I edit a podcast for YouTube in, and why each step comes where it does. It assumes two or more people, a camera on each (or one wide and one tight) and a microphone per person. If you recorded on one camera, most of it still applies; you just have fewer angles to cut between.
Step one: sync everything before you cut anything
A podcast shoot gives you a pile of files: two or three camera angles and a separate audio track per microphone, often from a recorder rather than the cameras. Before any creative decision, every one of those files needs to sit on the timeline lined up to the frame.
Most editing software can sync by matching the waveforms of the camera's scratch audio to the microphone tracks. It works well when every camera recorded sound. If one did not, a clap on camera at the top of the recording gives you a sharp spike to line up by hand. Check the sync at the start and again near the end: on long recordings, devices can drift apart, and a mouth that is a few frames off by minute forty is distracting in a way viewers feel before they can name it.
Then lay it out so you can see it: one video track per camera, one audio track per voice, labelled with names. Ten minutes spent here saves an hour of hunting later.
Step two: sound comes before picture
I fix the audio before I make a single cut, because every later decision depends on hearing the conversation properly.
The order on each voice:
- Pick the right microphone per person. Each person's own microphone is their voice; the other microphones pick them up too, quieter and roomier. Mute the spill so each voice comes from one place.
- Clean what needs cleaning. Hum, air conditioning, a chair squeak. Light noise reduction only; pushed hard, a voice starts to sound watery.
- Level the voices against each other. The quiet guest and the loud host should sit at the same perceived volume, so the listener is not reaching for the volume button every time the speaker changes.
- Control the peaks. Laughter and interruptions jump out. A compressor and a limiter keep them in line without flattening the voice.
When you upload, YouTube's recommended upload settings list AAC-LC or Opus audio, stereo, at a 48kHz sample rate. Work at 48kHz from the start and you never have to convert.
If the recording itself has problems, a room that echoes or a microphone too far away, there is a limit to what the edit can rescue. I wrote about that, and about getting it right on the day, in how to improve audio on phone video; the same rules hold in a studio.
Step three: cut the conversation, then cut the cameras
This is the step that separates an edit that drags from one that holds. Two different jobs get mixed up, and they should be done one after the other.
The conversation edit comes first. Listen through and cut for the listener: the false start, the tangent that goes nowhere, the "sorry, can you say that again", the two minutes spent finding a website on a laptop. I leave in most pauses, laughter and the odd "um", because they are what make it sound like two people talking rather than a script being read. What I take out is anything that would make a listener drift. On most conversations I edit, that is a few minutes per hour, sometimes much more; it depends entirely on the recording.
The camera edit comes second. With the conversation locked, you choose which angle the viewer sees, on top of a soundtrack that no longer changes. My rules of thumb:
- Cut to whoever is speaking when they start a new thought, not on every sentence.
- Cut to the listener for a reaction worth seeing: a laugh, a raised eyebrow, a pause before they answer.
- Use the wide shot to reset after a long stretch on one face, and to show both people when they are talking over each other.
- Do not cut for the sake of cutting. A long, strong answer can stay on one angle for a minute.
Doing it in this order means you never re-cut cameras because you changed the conversation underneath them, which in my experience is where podcast edits lose the most time.
Step four: rebuild the first minute
Podcasts recorded in a room tend to start slowly: a welcome, a bit of housekeeping, a "so, tell us about yourself". In my experience, the first minute is where a viewer from search decides whether to stay.
What I do on most episodes:
- Open on a moment, not a welcome. Pull a strong ten to thirty seconds from later in the episode, a clear opinion, a surprising number, a good story's turn, and put it first. Then the show's open, then the conversation.
- Get to the subject fast. If the guest's background takes four minutes, it can usually take one.
- Cut references to things the listener cannot see. "As you can see on this chart" means nothing to someone listening in their pocket. Either describe it in words in the edit, or put it on screen and keep the line short.
This is also where a podcast stops being only a podcast. The strong moment you chose for the open is very often your best clip for Shorts and reels. I cover finding and reframing those in how to convert horizontal video to vertical.
Step five: titles, chapters and the thumbnail
Keep on-screen graphics simple: a name and role the first time each person speaks, a few words when a topic changes, the occasional caption for a key line. A podcast is not a reel; nobody wants text bouncing on every sentence for an hour.
Chapters do more work than people expect on a long episode. YouTube's Video Chapters page sets the rules for adding your own in the description:
- the first timestamp must start at 00:00
- at least three timestamps, in ascending order
- each chapter at least 10 seconds long
YouTube can also add automatic chapters, and the page notes that your own timestamps override them. Write chapter titles as what the viewer gets ("Why the first hire was a mistake"), not as a table of contents ("Part 3").
The thumbnail is its own design job. For long-form, faces and a few words usually carry it; my approach is in YouTube thumbnail design tips.
Step six: export, upload and set it up as a podcast
Export to what YouTube asks for. Its upload settings say content should be uploaded at the same frame rate it was recorded, list MP4 with H.264 video, and give the audio spec above. If one camera recorded at a different frame rate from the others, decide on one timeline frame rate before you cut, not at export.
Then, in YouTube Studio, put the episode in a podcast. YouTube's podcast page says a podcast on YouTube is a playlist and each episode is a video in it. A few rules from that page catch people out:
- The podcast should only contain full-length episodes, in the order they should be watched. Clips and Shorts belong elsewhere.
- Give it a descriptive title. YouTube asks you to avoid generic names like "Full Episodes" or "Podcast", and says a title that is too vague gets replaced by your channel name in YouTube Music.
- The podcast takes its own square thumbnail, separate from each episode's thumbnail.
- MP3s cannot be turned into podcasts on YouTube. Every episode is a video.
Where I fit: long-form is the part of a month I can do the most with: I edit the full episode, then cut the strongest moments into clips for the rest of the month. If you record a show and want it edited, Book a call.
A short checklist for every episode
Before you hit publish, run through it once more:
- Is every camera in sync at the start and the end?
- Do all voices sit at the same volume, with no peaks that make you flinch?
- Would the episode make sense with your eyes closed?
- Does the first minute open on something worth staying for?
- Do the chapters start at 00:00, with at least three, each at least 10 seconds?
- Is it exported at the recorded frame rate, with 48kHz stereo audio, and added to the podcast playlist?
If you are handing this to someone else rather than doing it yourself, how to hire a video editor lists what an editor needs from you, and the video I make for clients is where my own work sits.