Podcast and Spoken-Word Audio

How to match loudness between your podcast host and guest in REAPER

By StudioRescue Editorial

Published August 29, 2026

A host and guest sitting at different volumes is one of the most common complaints in two-person podcasts. The listener ends up riding their own volume knob: turning up for the quiet guest, then getting blasted by the louder host. It makes an otherwise good conversation tiring to listen to, and it is usually very fixable in editing.

The key idea is that loudness is not the same as peak level. Turning one track up until its peaks look like the other track’s peaks does not make the two voices sound equally loud. You need to match perceived loudness over time, which is what the LUFS scale measures. REAPER can measure and match it with stock tools. Here is the whole process.

Why the two voices do not match

Before you fix it, it helps to know what caused the gap, because some causes are fixable in the mix and some are not.

• Different mics and preamp gain. The host and guest almost never use identical gear or the same input gain, so their raw recorded level differs.
• Different mic distance and speaking level. A guest who leans back or talks softly lands quieter; a host who works the mic closely lands louder.
• Remote guests. A guest recorded over a call or on their own device often arrives at a completely different level from the locally recorded host.

Loudness matching in the edit handles the average level difference cleanly. What it cannot do is rescue a voice that is distorted, buried under bleed from a shared mic, or swinging wildly because the person kept moving. Those are separate problems, and we will be honest about them at the end.

Step 1: Get each speaker on their own track

You can only match two voices independently if they live on separate tracks. If your host and guest were each recorded to their own track, you are ready. If they are both on one track, or share one mic, skip to the limits section below, because there is only so much you can do.

For future episodes, recording each person to a separate track is the single biggest thing that makes leveling easy. If you are not doing that yet, see how to record two people in REAPER. Keep each voice track mono. A mono voice on a stereo track can read differently on a loudness meter and throw off your match.

Step 2: Measure each voice

REAPER ships with a stock loudness meter so you do not need to install anything. Add it to the first speaker’s track, play a representative stretch of the episode, and read the integrated value.

Step 3: Match them to a common level

Now bring both voices to the same integrated loudness. A good working target for spoken word is around -18 LUFS, which leaves headroom to set the final episode level later. There are two stock ways to do it.

Option A: Track gain (simplest)

Using your two readings, raise or lower each track’s volume so both meters land on the same number. If the guest reads -22 and you want -18, raise the guest track by about 4 dB. If the host reads -16, lower it by about 2 dB. Re-play and re-check the meter, because your ears and the reading should agree. This method is transparent and keeps each voice’s natural dynamics.

Option B: Normalize the items to LUFS

REAPER can also set loudness directly on the audio items. Select all of one speaker’s items, right-click, and choose Item processing, then Normalize items. In the dialog, set Normalize to to LUFS-I and enter your target, for example -18 dB. Choose the option to normalize the items together (common gain), not each item separately, so the clips keep their natural relationship instead of every clip jumping to the same level. Then repeat for the other speaker with the same target. Both voices now sit at the same integrated loudness.

Step 4: Hold the match with gentle compression

Matching integrated loudness fixes the average, but one person may still swing from loud to soft within a sentence, so the balance drifts moment to moment. A light ReaComp on each voice tightens that range so your match holds.

Start gentle: a ratio around 2:1 to 3:1, a threshold set so you only catch a few decibels on the louder peaks, and a medium attack and release. You are smoothing the swings, not squashing the voice. After compressing, glance at the loudness meter again, because heavy compression can nudge the integrated value and you may need to trim the track gain by a decibel to line them back up. If you are new to these controls, see vocal compression settings in REAPER.

Step 5: Set the finished episode loudness

With the two voices sitting together, set the level of the whole show. Put the loudness meter on the master bus and aim the full mix at the common podcast target of about -16 LUFS integrated, with true peaks kept under -1 dBTP. Apple Podcasts recommends -16 LUFS with a tolerance of roughly one decibel either way. A purely mono spoken-word show can sit a few decibels lower, nearer -19 LUFS, and still match the loudness of stereo shows on playback. For the final render step and the reasoning behind that target, see podcast loudness and the -16 LUFS target in REAPER.

What loudness matching will not fix

Leveling is powerful, but it only moves volume around. It does not:

• Separate two voices on one mic or one track. If both people bleed into a single recording, you cannot raise one without raising the other. That is a recording problem, not an editing one.
• Repair a distorted or low-quality guest. A crunchy phone-quality guest will just be a louder crunchy guest. That needs cleanup, covered in how to fix bad remote guest audio in REAPER.
• Fully tame a voice that never sat still. If someone drifted toward and away from the mic constantly, compression helps but will not perfectly even it. Volume automation, or a re-record, may be needed.
• Make two mics sound the same. Matching loudness is not matching tone. Different mics still have different character, and background noise rises with the level, so clean each voice first if needed. See how to reduce background noise in a podcast.