How to Edit a Podcast in Premiere Pro (My Actual Workflow)

Seven steps, in this order: build the project, sync the cameras and the separate audio, do one full listening pass and mark it, cut for content, cut for camera, treat the audio, then export. Doing them out of order is the single biggest time sink I see.

Here is why the order matters. If you start cutting before you have listened all the way through, you will make decisions about minute 8 without knowing that the same story lands better at minute 41. If you polish audio before the picture cut is locked, you redo the polish. If you colour before the cut is locked, same thing.

The listening pass is the step everyone skips, and it is the step that makes the episode good. I watch the whole thing at 1.0 speed with markers on. Green marker for a cut point, red for a hard kill, blue for a moment worth pulling as a clip later. That is one hour of my life for a 60 minute episode and it saves two hours downstream.

My team works the same way on every job. Human editors watch the whole episode, and nothing gets auto cut by silence detection. If you are still planning your show, the podcast launch checklist covers what to get right before any of this matters.

Match your sequence to your camera, not to your delivery. Shoot 4K 25p, edit in a 4K 25p sequence, then let the export handle the rest. Making a 1080p sequence out of 4K footage throws away your reframing headroom on day one.

My folder structure inside the Premiere project, every time: 01 Footage, 02 Audio, 03 Sequences, 04 Graphics, 05 Music and SFX, 06 Exports. Numbers so they sort. Bins nested no deeper than two levels or nobody can find anything six months later.

Sequence naming matters more than it sounds. I use EP012_MAIN, EP012_CLIP_01, EP012_CLIP_02. When a host emails asking for a fix in the third clip, I know where to go without opening five timelines.

Set your scratch disks to the same drive as the media. Turn on auto save at five minutes with twenty versions kept. And do a proxy pass if you are on a laptop: 1/4 resolution ProRes Proxy or H.264 proxies. A 90 minute multi-cam episode at full 4K will make most machines stutter, and stuttering playback makes you cut badly because you stop trusting what you hear.

For anything past two cameras, multi-cam editing has its own set of rules.

Select every camera clip plus the audio recorder files, right click, Create Multi-Camera Source Sequence, and set Synchronise Point to Audio with the track channel set to Mix Down. That works for maybe 90% of episodes if every camera recorded scratch audio.

The other 10% is where people lose an afternoon. Sync fails when a camera stopped and restarted mid recording, when scratch audio is too quiet or too clipped for the algorithm to match, or when one camera drifted because it was recording 29.97 while the others ran 30.

Three fixes, in the order I try them. First, split the broken camera at the restart point and sync the halves separately. Second, sync by timecode if the cameras were jam synced. Third, do it by hand: find the clap or the loudest laugh, line the waveform peaks up visually, and nudge frame by frame until the flam disappears.

On drift, check the whole episode, not just the start. A camera that is in sync at minute 2 and eight frames off at minute 58 needs a speed change of about 99.9%, not a re-sync.

Once synced, I mute all camera scratch audio and keep only the recorder tracks. Scratch audio left live under a good mic is the most common reason an episode sounds hollow.

Cut for content first, camera second. They are two different jobs and mixing them doubles the time.

Content pass: I work off my markers on a single wide angle. Remove the false starts, the tangent that went nowhere, the two minutes of setup chat, the repeated question. I do not remove every um. Stripping all filler makes people sound like androids, and hosts notice. I pull the ones that stack up: three ums in one sentence, the long stall before an answer, the throat clear.

Camera pass: now I lay in angles. Rules I follow, and yes they are rules and not vibes. Hold each angle at least four seconds. Cut on the incoming speaker, not two seconds after they started. Do not cut back to the same angle you just left. Use the wide for laughs and crosstalk, because two singles cutting fast during overlapping speech reads as chaos.

Use Premiere's multi-cam monitor and record the angle switches live at 1.0 speed, then clean up. It is far faster than dragging cuts, and it makes the rhythm feel like a conversation rather than a spreadsheet.

Tag every voice track as Dialogue in Essential Sound, then build the chain in this order: noise reduction, EQ, compression, de-esser, limiter. Order matters because each stage feeds the next.

My defaults, which you should adjust by ear rather than copy blindly. Noise reduction around 20 to 30%, higher than that and voices start sounding like they are underwater. A high pass filter at 80Hz to kill desk rumble. A gentle cut around 250 to 400Hz if the room boxes up the low mids. A small lift at 3 to 5kHz for clarity. Compression at roughly 3:1 with about 4 to 6dB of gain reduction on the loudest sections. De-esser only if you actually hear sibilance.

Target loudness for a video podcast is around minus 14 LUFS integrated with true peaks under minus 1dBTP. Use Essential Sound's Loudness match to get in the neighbourhood, then check with the loudness radar rather than trusting your headphone volume.

One more thing. If a guest joined remotely on a bad connection, no plugin chain saves it. Fix that at the recording stage. Music beds sit at roughly minus 24 to minus 20 LUFS under speech, and duck them properly instead of leaving them flat.

For the main episode on YouTube: H.264, Match Source, VBR 2 pass, target bitrate 35Mbps and maximum 45Mbps for 4K, or 16 and 20Mbps for 1080p. Audio at AAC 320kbps, 48kHz. Tick Use Maximum Render Quality only if you are scaling, because otherwise it just costs you time.

For vertical clips, do not just crop the 16:9 export. Go back into the timeline, make a 1080x1920 sequence, and reframe each shot on purpose so the speaker sits in the upper third with room for captions. Auto Reframe is a decent starting point on single speaker sections and genuinely bad during crosstalk, so check every clip it touches.

Burn captions into vertical clips. Most people watch them muted. Premiere's Transcribe then Create Captions is accurate enough that you are proofreading rather than typing, though it will still mangle names and product terms every single time.

We cut vertical clips at $25 each as an add on, and the picking matters more than the cutting. Choosing which moments actually travel is a judgement call, and what works on TikTok is not what works on LinkedIn. Thumbnails are their own craft, which is why our sister studio ClickTheory does 8 for $250.

Premiere Pro is genuinely the right tool for video podcasts. Multi-cam is mature, Essential Sound gets you 80% of the way on dialogue without opening Audition, and the transcript based editing has got good. I am not going to pretend otherwise to sell you something.

The honest question is not whether you can learn it. It is what an hour of your time is worth. A clean 60 minute two camera episode is four to six hours once you know the software, and closer to twelve the first few times. Weekly, that is roughly twenty hours a month before you have made a single clip.

So run the numbers. We charge $199 an episode for Starter, single camera, 1080p, one revision round, and $299 for Studio, two to four cameras, 4K, unlimited revisions. Up to 60 minutes is the base, up to 30 minutes costs less, up to 90 minutes costs more. A 4 episode pack saves 15%, monthly saves 25%. A typical agency is $300 to $400 an episode and then about $400 a month again for clips.

You keep everything either way. Footage, project files and exports are yours. How it works walks through the handoff, and founders running a show alongside a company tend to feel this maths fastest.