Mic Too Quiet, System Audio Too Loud: Balancing the Two Tracks

Narration buried under app sound? Here's how to set mic and system audio levels before you record and rebalance them per segment in the editor.

You record a walkthrough with your voice over an app that plays notification sounds, a video, or a conference call. On playback, the app drowns you out in some places and disappears in others, and your own narration sits somewhere between a whisper and a shout depending on which sentence you’re on.

This isn’t a recording failure. Both sounds made it into the file. What’s missing is a balance between them — and because a screen recorder captures two completely unrelated sources, nothing sets that balance for you unless you do it.

Here’s how the two tracks behave, what to set before you hit record, and how to fix the mix afterward without re-recording.

Why the Two Sources Drift Apart

Your microphone and your computer’s audio output have nothing in common except that they end up in the same video.

System audio level is whatever your apps decided. It’s the product of the app’s own volume, your operating system’s output volume, and how loud the source material was mastered. A YouTube video, a Zoom call, and a UI click sound can be 20 dB apart from each other without you ever touching a slider.

Mic level is set by physics and gain. How close you sit, which direction you face, and the input gain on the device all change it. A laptop’s built-in mic two feet away produces a much weaker signal than a USB mic six inches from your mouth.

Nothing normalizes them. Recorded has no loudness normalization, compressor, limiter, or noise reduction. The levels in your export are the levels you recorded, scaled by whatever you set in the editor. That’s predictable, which is good — but it does mean the balance is your decision, not the app’s.

The practical consequence: a recording that sounded fine in the room can play back with narration pinned under app audio, especially for viewers on laptop speakers, which have little headroom to spare.

The Two Tracks Stay Separate

The important structural fact is that Recorded does not mix the sources at record time. When you enable both, the microphone is written to its own audio file and system audio to another one, alongside the video. They stay separate all the way into the editor, which is why you can change their relative levels afterward at all.

That also means two things that trip people up:

  • If you recorded with the Mic toggle off, there is no mic track to turn up later. The same goes for System Audio.
  • If one of the sources was captured badly — clipped into distortion, or so quiet it’s buried in hiss — no slider recovers it. Boosting a near-silent track raises its noise floor by exactly as much.

Both toggles live next to each other in the main window before you start, labeled Mic and System Audio, and both show a small live level meter while they’re enabled.

Set Levels Before You Record

Two minutes here saves an editing session later.

1. Watch the meters, not the room. Enable Mic and System Audio in the main window. The meters animate in real time with the actual signal. Speak at your normal presenting volume and watch the mic meter move; play a few seconds of whatever you’re going to demo and watch the system meter.

2. Aim for “clearly moving, never pinned.” You want consistent movement with headroom left over. A meter that barely twitches means you’ll be boosting noise later; one that sits at the top means the signal is clipping, and clipping is permanent.

3. Turn the source down, not the mic up. If app audio dominates, lower it at the source — the app’s own volume slider, then the system output volume. Recording system audio at a sane level and leaving your narration alone is far easier than fixing the opposite.

4. Fix mic distance before mic gain. Moving closer to the mic improves the ratio of voice to room noise. Raising gain amplifies both. Audio recording best practices goes deeper on placement and room treatment.

5. Wear headphones. If system audio comes out of your speakers while your mic is live, the mic re-records it. You get a doubled, slightly delayed version of every app sound, and no amount of editing separates it back out.

6. Record fifteen seconds and play it back. Talk over the loudest thing in your demo, stop, and listen. This single test catches almost every level problem while it still costs fifteen seconds to fix.

If system audio isn’t being captured at all, that’s a different problem from a balance problem — see recording internal audio on a Mac or screen recording on Windows 11 with audio.

Rebalancing in the Editor

Open the recording in the editor and look at the timeline. Each video segment draws the waveforms of the audio it contains, so you can see loud and quiet stretches before you hear them. The waveform reflects your current volume and mute settings, which makes it a usable visual check on the mix.

To change levels, select a segment on the timeline and open the Segment tab. If the recording has audio, you’ll find a section with separate controls for System Audio and Microphone: a volume slider for each, and a switch to mute each one. A control only appears if that track exists in the recording.

The sliders run from 0% to 400%. Anything above 100% is amplification, with the caveat above: you’re amplifying hiss, keyboard noise, and fan noise at the same rate as your voice. Treat 150–200% as a rescue for a too-quiet track, not a default.

Crucially, these settings are per segment, not global. That’s what makes a real mix possible.

Mixing in Sections Instead of Setting One Level

Most walkthroughs don’t need one fixed balance — they need different balances at different moments. The workflow is to split the timeline where the intent changes, then set levels on each piece.

Move the playhead to the boundary and press C to split at the playhead. Now you have two segments that carry independent audio settings. Select one, set its levels, select the next, set those. You can select several segments at once and move a slider to apply the same change to all of them.

A recipe that works for most product walkthroughs:

MomentSystem AudioMic
You’re narrating over the UI20–40%100%
The app sound is the point (alert, notification, video)100%mute or low
A video or call plays and you stay silent100%mute
Intro and outro to cameramute100%

Muting the mic over sections where you aren’t talking is underrated. It removes breathing, chair creaks, and typing from exactly the stretches where viewers are trying to hear something else.

If splitting for audio also leaves you with sections worth removing entirely, trimming and cutting screen recordings covers how to cut them cleanly.

Where Background Music Fits

If you add a track from the Music tab, it becomes a third element competing for the same space. Music has its own volume control, and it should sit well under everything else — far lower than feels right while you’re setting it, because you already know what the narration says and your viewers don’t. Using background music in screen recordings covers choosing and placing a track.

A rough order of priority: narration first, functional app sound second, music a distant third. Any mix where you have to concentrate to follow the words is wrong, no matter how good the track is.

Check the Mix Before You Export

Levels that sound balanced on studio headphones can fall apart on a phone speaker. Before exporting:

  • Play the whole thing through once. Not scrubbing — actually listen, start to finish. Audio problems hide in the sections you skip.
  • Listen once on laptop speakers. This is what most viewers use, and it’s the least forgiving device for quiet narration under loud app audio.
  • Check the transitions between segments. If you set very different levels on adjacent segments, the jump can be jarring. Soften it by moving the split to a natural pause.
  • Confirm audio is actually included. In the export panel, Include Audio lives under Advanced Settings. If it’s off, you get a silent MP4 regardless of everything above.
  • Remember GIF has no audio at all. Exporting to GIF discards the mix entirely. If sound matters, export MP4.

Quick Troubleshooting

SymptomLikely causeFix
Narration buried under app soundSystem audio recorded too hotLower System Audio on narrated segments; next time, turn the app down before recording
Voice sounds distorted or cracklyMic clipped during recordingRe-record with lower gain or more distance; clipping can’t be undone
Voice is clear but faint, with audible hiss when boostedMic too far away or gain too lowBoost to 150–200% as a stopgap; fix placement before the next recording
Every app sound is doubled with a slight echoSpeakers playing while the mic was liveWear headphones and re-record
Mic slider missing in the Segment tabThe recording has no mic trackThe Mic toggle was off at record time
Export has no soundInclude Audio turned off, or exported as GIFRe-export as MP4 with Include Audio on

The Short Version

Two sources, no automatic balance. Set them with the live meters before you record, keep headphones on so they don’t bleed into each other, then split the timeline where the intent changes and give each segment its own levels. The editor can rescale what you captured — it can’t invent headroom that wasn’t there.

If you haven’t set up the app yet, you can download Recorded for macOS or Windows and try a short walkthrough with both tracks enabled.