By the ReScreen team · · 5 min read
The picture is perfect and the sound is a mess. Your voice is a murmur, the app you are demonstrating chimes at full volume over the one sentence that mattered, and somewhere underneath there is a laptop fan. You search for how to fix bad audio in screen recordings and the advice is "buy a better microphone", which does nothing for the take you already have and little for the real cause.
People forgive soft video more readily than audio that is hard to listen to. For a course creator that is an abandoned lesson; for a sales rep, a demo nobody finished; for a support team, an answer that produced another ticket.
Why is the audio in screen recordings so often bad?
Because "bad audio" is three different problems, and many recorders weld them together.
Your voice is too quiet
A microphone half a metre from your mouth records at a modest level, and that level moves every time you lean back or turn to a second monitor.
The app is too loud
System audio arrives at full digital level. A notification, a video call joining sound or a product's own sound effects sit far above your narration. Turn the whole recording up to hear your voice and the chime becomes painful.
The room is in the recording
Fans, air conditioning, a mechanical keyboard, traffic. You stop noticing them. A viewer in headphones does not.
One flattened track makes all three permanent
If the recorder mixes microphone and system audio into a single track as it captures, every fix hits everything. Raising the volume raises the chime, and filtering the noise dulls the app audio. The only real repair is recording again. We go deeper on this in why separate audio tracks save recordings.
What actually fixes bad audio in a screen recording?
The fix is structural: capture each source on its own, and decide the mix afterwards. ReScreen, a native macOS recorder and editor, records four independent tracks: screen, camera, microphone and system audio. The audio controls live in the editor's Audio inspector, and the same tab in Settings holds your defaults.
System audio is captured through macOS itself, with no virtual audio driver or loopback device to install. The setup is in our guide on how to record system audio on a Mac.
Balance voice against the app
Each track has its own "Volume" from 0 to 200%, plus "Mute". System audio defaults to 60% so it sits under narration without any action from you. If your voice is still low, raise the microphone above 100%. If the app need only be faintly heard, drop system audio to 30 or 40%.
Remove the one sound that ruined the take
If a notification landed mid-sentence and the app audio is not needed, "Mute system audio" silences that whole track without touching your voice. If the app audio matters elsewhere in the video, keep it and cut the affected moment instead: set In and Out points around it and use "Remove Range", which closes the gap. The timeline editor handles this without an export and import cycle.
Clean up the voice
The microphone track has three cleanup controls, all off by default so nothing is processed unless you ask.
- "Reduce background noise" softens quiet background sound between phrases, which is where fan and room noise is most audible.
- "Level voice volume" evens out the difference between leaning in and leaning back.
- "Voice enhancement" at Light or Strong applies a high-pass filter, a soft noise gate, levelling and compression. Light suits a decent microphone; Strong is for a built-in microphone or a noisy room.
The original tracks stay separate and unchanged. Processing is applied in preview and export, so you can switch it off again.
Smooth the edges
"Fade in" and "Fade out" run from 0 to 5 seconds per track.
Which audio settings should you use?
| Problem | Setting | Suggested value |
|---|---|---|
| Voice quieter than the app | System audio "Volume" | 30 to 60% |
| Voice quiet overall | Microphone "Volume" | 120 to 160% |
| Fan or room hum between sentences | "Reduce background noise" | On |
| Level drifts as you move | "Level voice volume" | On |
| Thin or boomy built-in microphone | "Voice enhancement" | Light, then Strong if needed |
| App audio not needed at all | "Mute system audio" | On |
| Abrupt start or end | "Fade in" and "Fade out" | 0.3 to 1 s |
By scenario: for tutorials, keep system audio around 30 to 40% and switch on "Level voice volume". For product demos where the product makes sound, stay at 60% and add short fades. For courses, start from the built-in "Course Recording" preset, which records at 1080p with noise reduction on, and keep the same settings every lesson. Consistency across lessons matters as much as quality, a point we expand on in screen recording for online courses.
What to fix before you press record
- Watch the live input meter before you start. If it barely moves when you speak, move closer or raise the input level.
- Wear headphones when recording system audio. Otherwise your microphone hears the speakers, the app sound lands on both tracks, and ReScreen has no echo cancellation to separate them.
- Use the microphone you have. All audio is resampled to 48 kHz, so AirPods and high-sample-rate interfaces work without configuration.
- Choose Mono for a single voice. "Capture channels" also offers stereo when the device supports it.
- Use ⌥⌘M to mute the microphone during a cough or an interruption without stopping the take.
If you appear on camera, the same record-first, decide-later approach applies to the picture: the webcam overlay shapes guide covers shape, size and position.
What separate tracks cannot do
There is no AI voice isolation, no echo cancellation, no transcription or captions, and no music library. If the microphone clipped, or a loud voice in the room was picked up alongside yours, no slider will restore it. Captions are best added with your publishing platform's own tools.
One more detail: the ready-to-play MP4 written when you stop contains microphone and system audio mixed into one track. The separate tracks live in the project, so rebalance there and use Export for the corrected version.
Next step: make the mix a decision, not an accident
Open your most recent recording and listen with headphones. If you wince, ask whether you could repair each source alone. The camera and audio feature page shows what ReScreen captures and how the tracks are kept apart.
Every plan includes every audio control, with no watermark. Compare monthly, yearly and lifetime licences on the pricing page, or download ReScreen for macOS.
Filed under