Podcast editing for musicians who actually play on the show

The difference is that half your episode is a performance, and performances cannot be edited like conversation. In a normal interview show I can cut a rambling answer in the middle and nobody notices. Do that inside a chorus and the whole thing collapses.

So the workflow inverts. Before anyone trims a single filler word, my editor scrubs the episode and marks every musical moment: the full songs, the half-verse someone plays to demonstrate a point, the noodling under an answer, the count-in before a take. Those regions get locked. Everything else is fair game for tightening.

A human watches the entire episode, start to finish. Nothing here is auto-cut by silence detection, and on a music show that matters more than anywhere else. Silence detection reads a held note, a rest between phrases, or a guitarist finding a chord as dead air and rips it out. I have seen tools shorten a 4 minute acoustic take to 2 minutes 40 and call it a win.

The second difference is audio handling. Talk and music want opposite treatment. Speech wants tight compression and a high pass filter. Music wants headroom and the low end left alone. When both live in one timeline, the settings that make your voice sound present make your guitar sound thin. The rest of the workflow is the same shape as any other show, and I walk through it in detail on podcast post production.

Nothing, in the best case. A good live take gets left alone and framed properly.

Here is the order my editors work in. First pass, watch the full episode and log timecodes for every performance and every real conversational beat. Second pass, cut the talk: false starts, the three minute tangent about parking, the part where someone's phone rang. Third pass, camera cutting across the performances. Fourth pass, audio balance, then titles and clips.

On the camera cutting: a song is where multi camera earns its money. During conversation, cutting every 8 to 12 seconds keeps it alive. During a performance you want the cuts to follow the music, which usually means holding a wide for the intro, moving to hands for the instrumental figure, and sitting on the face through the vocal. If you shoot two to four angles I cut them to the song structure rather than on a timer. That is the multi-cam editing workflow, and music shows are the clearest case for it.

When a take genuinely has a wreck in it, a dropped string or a blown lyric, I will ask before I touch it. Sometimes the fix is a cut to a wide on the flub and back. Sometimes you played it again and take two is better. Sometimes the flub is the best moment in the episode and it stays in. That is a call for you, not for me, so it goes in the notes I send with the first cut.

Because every automatic tool is optimised for speech, and a musician's episode is not speech. That is the whole story.

Auto silence removal cuts anything under a threshold for longer than a set duration. A rest is silence. A fade at the end of a song is silence sloping downward. A drummer counting in with sticks reads as noise, not speech, so some tools cut the count and leave the song starting cold.

Filler word removal is worse. It scans the transcript and deletes tokens. Sung lyrics get transcribed badly, so it finds "yeah" and "oh" and "hey" inside your chorus and removes them. I have watched a tool take four "oh"s out of a hook.

Auto clip generators pick highlights by transcript density, so they reliably skip the song and clip the part where you explained your pedalboard. Your best 30 seconds of content on a music show is almost always musical, and the transcript there is nearly empty.

I am not saying the tools are bad. They are genuinely fast at talk-heavy shows and I use editing software the same way anyone does, including Resolve, which I break down step by step in how to edit a podcast in DaVinci Resolve. The point is that automation cannot tell the difference between a pause and a rest, and on your show that difference is the whole product.

The fix is treating them as two separate jobs on the same timeline instead of one job with one setting.

Speech tracks get a high pass around 80Hz, compression, and levelling so every host sits at a consistent loudness. Music tracks get none of that. No high pass, because the bass guitar and the kick live under 80Hz. Light compression at most, because your take is already mixed the way you wanted it. If you hand me a board feed or a separate music stem, I use it and I do not touch it beyond level matching.

Loudness is where music shows trip up. Podcast platforms target around -16 LUFS for stereo. If I normalise the whole episode to that, the songs get squashed against the talk and lose their dynamics. What I do instead is set the speech to sit at target and let the music ride a couple of dB under peak on the loudest sections, so the transition into a song feels like a lift rather than a jump scare in your listener's headphones.

Room noise matters more on your show too. Instrument mics pick up more of the room than a broadcast mic on a boom does. I clean the noise floor in the talk sections and leave the performance floor alone, because noise reduction on a cymbal decay sounds like a bad phone call. Where a take has a real problem, I flag it and give you options instead of quietly processing it into mush.

Both, and the ratio that works is roughly two musical clips to one talking clip. Musical clips travel further, but talking clips are what make new people care who you are.

A 30 to 45 second slice of a live performance is the single best vertical asset a music podcast produces. It needs almost nothing done to it: reframe to 9:16, keep the hands or the face in frame, caption the lyric line, done. Those clips get watched to the end because there is a payoff built in, which is the note the phrase resolves on. Pick the hook, not the intro.

Talking clips need harder work. The good ones are opinions, not explanations. "I stopped using click tracks and here is what happened" beats "here is how our writing process works" every time. I pull those from the conversation passes and treat them the way any show does, which I cover in TikTok podcast clips.

One warning that is specific to you. If you play someone else's song on the episode, that clip is a copyright claim waiting to happen on YouTube and Instagram, and the platforms are getting faster at catching live covers. Originals are safe. Covers are a gamble and I will flag them in the notes rather than let you find out from a takedown.

Clips are $25 each as an add-on, so a typical music episode running six clips adds $150.

Starter is $199 per episode: single camera, 1080p, one revision round. Studio is $299 per episode: two to four cameras, 4K, unlimited revisions. Both are per episode with a 60 minute base. Under 30 minutes costs less, up to 90 minutes costs more.

For a music show I would push you to Studio, and not to upsell you. The unlimited revisions matter because performance edits are subjective in a way interview edits are not. You will want a cut moved by half a bar. That should not cost you extra, and on Studio it does not. The extra cameras matter for the same reason.

Discounts stack against consistency. A 4-episode pack saves 15%. Monthly saves 25%, which puts a weekly Studio show at roughly $224 an episode. Add-ons are vertical clips at $25 each, show notes and chapters at $35, and rush 24-hour delivery at $75 when you need an episode out for a release date.

Turnaround: first cut back in 48 hours, revisions the next working day. You own everything. Footage, project files and exports belong to you, so if you ever want to take the show in-house you take the whole timeline with you.

One more thing musicians ask about constantly, which is thumbnails. I do not do them. My sister studio ClickTheory does, at 8 for $250, and for a show where the hook is a performance the thumbnail is doing a lot of the work on YouTube.

Record a separate audio feed for the music. This is the single biggest thing. If your only music audio is the same mic that is capturing conversation, I am working with a compromise from the start. A board feed, an interface recording, or even one dedicated mic on the amp changes the ceiling of what the episode can sound like.

Clap or count in before each take, on camera, with everything rolling. Sync on a music episode is harder than on a talk episode because the tell for drift is different: a voice that is 2 frames out looks slightly odd, but a guitar strum that is 2 frames out is immediately wrong. A clap gives me a hard reference.

Keep one camera locked wide and never touch it. During a song that wide is my safety net for every cut I need to make. If someone is operating handheld and you only have two angles, and the operator drifts during the solo, I have nowhere to go.

Do not stop recording between songs. The talk between takes is often the most human content in the episode and it is where the good short clips hide. Storage is cheaper than a moment you did not capture.

Finally, send me the setlist and the timecodes you remember. Even rough ones. It saves my editor a pass and it means the intros and branding get built around what the episode actually is rather than a guess.