Back to blog

Tab audio vs system audio vs microphone

Three ways to get speech into a caption app, what each one can hear, and where each one fails.

Nov 4, 2026TanLens Editorial
Tab audio vs system audio vs microphone

Before you choose a captioning tool, decide where the sound comes from. The three capture routes behave differently, and most "it did not work" reports come from picking the wrong one.

Microphone

The microphone hears the room: your voice, the people around you, the speakers, and the fan in the corner. It needs no extension and no special permission beyond the browser's microphone prompt.

Use it for in-person meetings, classrooms, and any call where you can put the call on speaker. Headphones help: without them the microphone hears the remote participants twice, once directly and once through your speakers, and the recognizer produces doubled words.

Browser tab

Tab capture takes the audio of one browser tab and leaves everything else alone. It is the right choice for YouTube Live, Twitch, TikTok Live, a web meeting, or a course player.

It requires a browser extension because a normal web page cannot read another tab's audio. Chrome also requires a user gesture: you click Start, the browser asks to share audio, and only then does capture begin. If you reload the extension or close the tab, the session ends.

The important caveat: DRM-protected players can block audio capture or show a black frame. Netflix and similar services are not guaranteed. Test the specific title you care about before building a workflow around it.

System audio

System audio means everything the operating system is playing, including native apps. Browsers do not expose this directly. On macOS and Windows you can route system audio into a virtual input device and use microphone mode, or share the app's audio through the extension's picker where the OS allows it.

This route is more fragile because the audio path depends on the OS mixer, the virtual device, and the app. Expect to spend a few minutes setting it up once, and keep the route documented for the next person.

A quick decision table

| Source | Use microphone | Use tab capture | | -------------------------------- | ---------------------------- | ------------------------------ | | Person in the room | Yes | No | | YouTube Live on the same machine | No | Yes | | Zoom/Meet in a browser tab | No | Yes | | Zoom desktop app | Yes, via speakers/headphones | Only with system-audio routing | | DRM streaming service | Maybe, via speakers | Not reliable | | Uploaded recording | No | Use file transcription |

What to check when nothing appears

  1. Is the language pair right? Auto-detect is convenient but explicit is better for names.
  2. Is the tab actually playing audio? Muted tabs produce silence, not captions.
  3. Did the browser grant microphone or audio-capture permission?
  4. Is a second session already running on a free plan? Free allows one concurrent session.
  5. Are you out of minutes? The overlay stops and the balance shows zero.

Captions are only as good as the audio you feed them. Choosing the right route is half the job.