Podcast and Spoken-Word Audio

How to cut a long podcast interview down to length in REAPER

By StudioRescue Editorial

Published September 22, 2026

This is the least glamorous problem in podcasting and one of the most common. The conversation went well, it ran 75 minutes, and the slot is 40. Somewhere in there is a good episode, and the job is to find it without making the two voices drift out of alignment and without leaving the whole thing sounding chopped.

It is worth being clear about what this article is and is not. This is the structural edit: deciding what leaves the episode and removing it cleanly. The word-level cleanup, the ums and the breaths, is a different job with different tools, and it is covered in how to edit out filler words and breaths in REAPER. Doing them in the wrong order is one of the main reasons a long edit takes all day.

Get the ripple mode right before you cut anything

An interview is usually at least two tracks, one per voice, recorded from the start of the session. If you delete ten seconds from the host track only, everything after that point on the host track slides ten seconds earlier while the guest track stays where it was. The conversation now overlaps itself, and it gets worse with every cut.

REAPER solves this with ripple editing, which has three modes:

• Ripple editing off. Deleting leaves a gap. Moving an item moves only that item. This is the default and it is the wrong mode for cutting an episode down.
• Ripple editing per track. Material later on that track slides back to close the gap. Other tracks do not move. Useful for one-track work, still wrong for a multi-track interview.
• Ripple editing all tracks. Everything later in the whole project moves together. This is the mode for cutting an interview down to length.

You can reach the modes from the Options menu, which in current REAPER has a dedicated Ripple editing submenu, or from the ripple button in the main toolbar. Keyboard shortcuts for ripple editing have changed between REAPER versions and depend on your key map, so check the menu rather than trusting a shortcut you read somewhere.

Step 1: Fix the target number, then measure what you have

Decide the finished runtime before you start, including intro, outro and any music. A target that exists only as a vague sense that the episode is too long produces an edit that never ends.

Then find out how far over you are. Press End to jump to the end of the project and read the position. Now you have a real number: not “too long” but “we are 34 minutes over”. That number is what tells you whether this is a segment problem or a trimming problem, and the answer is almost always segment.

Step 2: Map the episode before you cut it

Listen through once, at speed, and mark the structure. You are not editing yet. You are building a table of contents so you can see what the episode is made of.

Step 3: Cut in passes, biggest first

The order matters more than the technique. Each pass should only start when the one above it has run out of material.

Pass 1: Whole segments

The tangent that went nowhere. The question that got the same answer as an earlier question. The five minutes of setup before the conversation actually started. These are the only cuts that move a runtime by double-digit minutes, and because you are removing a complete unit of conversation, the join is usually easy to make invisible.

Pass 2: Inside answers

Most spoken answers have a run-up, a point, and a restatement of the point. The run-up and the restatement are usually removable. Cut into the answer, not around it.

Pass 3: Dead air and long pauses

Thinking pauses, interruptions, the pause while someone finds a document. Shortening rather than removing these keeps the conversation sounding natural. A pause of a second or two often reads as thoughtful; six seconds reads as a technical fault.

Pass 4: Breaths and filler words, only if you are still over

This is the slowest possible way to recover time and the easiest way to make a conversation sound artificial, so it goes last. One published REAPER podcast workflow describes taking a slice out of each breath, in ripple editing all tracks mode, specifically as a last resort for hitting a hard runtime limit. If you get here, the method is in the filler words and breaths article.

How to make the actual cut

With ripple editing all tracks on, the fastest structural cut in REAPER is a time selection plus a ripple delete.

1. Make a time selection across the section you are removing, by clicking and dragging in an empty part of the track area, or with the time selection start and end keys. Esc clears a selection you are unhappy with.
2. Check both edges by playing across them before you commit. Zoom in. A cut that starts a fraction too early clips the last consonant of the previous word.
3. Remove and close the gap. Right-click in the arrange area or on the timeline ruler and choose Remove contents of time selection (moving later items). If you cannot find it in the menu, open the Actions list with ? and search for “remove contents”. This is not the same as a plain delete, which leaves a hole.
4. Listen across the join immediately, before the next cut. Fixing one bad seam now is trivial. Finding it after forty more cuts is not.

If you prefer working with items rather than time selections, the alternative is to make the time selection, press Shift+S to split every selected item at both edges of the selection, then select and delete the middle piece with ripple editing on. It is a few more keystrokes and it gives you a clearer visual of exactly what is leaving.

What this will not do

Structural editing only removes. It is worth being honest about the limits before you spend a day on it.

• It will not make a dull conversation interesting. A tighter version of an interview with no good answers in it is a shorter dull episode. Editing cannot add a point that was never made.
• It will not fix a missing answer. If the guest never said the thing the episode needed, no arrangement of the tape produces it. That is a follow-up question, not an edit.
• It does nothing for how the episode sounds. Levels, noise, host and guest loudness mismatch and final delivery loudness are all separate jobs that happen after the cut is locked. Start with measuring and hitting your loudness target.
• It will not catch a cut that orphans a reference. Removing a question but keeping a later “like I said earlier” leaves a hole in the logic that sounds fine technically and confuses the listener. Only a listen-through catches these.
• It will not protect music you placed early. Ripple editing all tracks moves the music bed along with everything else, which is rarely what you want. Lock the spoken edit first, then add intro, outro and beds, and only then set up ducking under the voice.

If something has gone wrong

1. The tracks have drifted out of sync. Undo back past the bad edit rather than trying to nudge tracks back into place, then check the ripple mode before redoing it. Nudging almost never lands back on sample-accurate alignment.
2. Your markers no longer line up with the audio. That is the regions and markers ripple setting described above. Set it, then rebuild the marker map from the current edit rather than trusting the old one.
3. Every join clicks. Auto-crossfades are off, or the crossfade is too short for the material. Turn the toolbar toggle on and re-make the worst joins.
4. The episode is on time but sounds hacked. You probably did Pass 4 work before Pass 1 work. Go back to the region list and look for a whole segment to lose instead, then restore some of the breathing you took out.
5. The edit is taking forever. One published REAPER technique is to raise the project play rate slightly while editing. The engineer who described it reports starting to miss editing artefacts above about 1.25x, and warns to reset it to 1.0 before final previews and rendering. Treat it as a personal preference to test, not a setting to leave on.