Screen Recording Audio: System Sound, Microphone and Mixing
Picture problems in a screen recording are visible immediately; audio problems surface after the take. System audio and the microphone are two independent sources blended live into one track — understanding that blend answers most recording sound questions.
How two sources become one track
Browser recording draws audio from two places: system sound (game effects, meeting voices, playback) and the microphone (your narration). The tool mixes both live through an audio context into a single track — what lands in the file is a real-time sum of the two.
With one source selected, the signal goes direct with no mixing, which is the cleanest audio. With both enabled, the balance is locked at record time: the tool cannot rebalance afterward. If the mic input runs hot, your voice sits on top of the game audio forever — that is physics, not a rendering bug.
Three usual suspects for silence
A silent recording is checked by hit rate: first, the system audio never played — the source window was playing but the OS output device pointed elsewhere (Bluetooth headphones disconnecting is the classic). Second, the browser tab is muted — browsers have per-tab mute switches, and a muted tab records a silent stream. Third, permission denied — a rejected microphone permission fails silently as an empty track.
The cheap insurance: record a 10-second sample before the real take and listen for both sources and their balance. Ten seconds saves re-recording an entire meeting or live stream that cannot be repeated.
Echo and doubled voices
The same sentence appearing twice with a slight offset is the speaker-plus-mic echo: system audio plays through the speakers and is re-captured by the microphone milliseconds later. When narrating with system sound on, wear headphones — sealing the system audio in your ears keeps the mic from recording a second copy.
When you only need system sound (game captures without commentary, meeting archives without room noise), turn the microphone off and record the single system track — cleanest level and quality. For finished files that need audio surgery, extract the track first to confirm it is complete before editing or compressing.