РусскийRUEnglishEN

Can't hear the other person or the microphone

In brief

Whisperer listens to a call from two sources: your voice — through the microphone, the other person's voice — through system audio. In the transcript they're marked as [Me] and [Other]. On macOS, system audio comes through the "Screen Recording" permission; on Windows, system audio is captured from the default output device without any permission needed — see Windows permissions. If one of the roles disappears or the transcription goes "silent", the problem is almost always in permissions, the choice of source/device, or session settings.

This article is a step-by-step diagnostic checklist: from the most common case (no other person = no "Screen Recording") to noise suppression and the transcription language.

When to use this

  • The transcript shows only [Me] — you can't hear the other person.
  • The transcript shows only [Other] — your voice isn't being recorded.
  • The transcription is empty or "patchy", even though audio is flowing on the call.
  • Text is recognized in the wrong language (garbled words).

Step-by-step (diagnostic checklist)

  1. No other person? On macOS this is reason #1 — check "Screen Recording": System Settings → Privacy & Security → Screen Recording → the toggle next to Whisperer is on, then restart the app (without screen recording, system audio isn't available). On Windows no permission is needed for this — make sure the default output device is the one the call actually plays through: system audio capture only picks up the default device. For details, see Windows permissions.
  2. No voice from you? Check the "Microphone". System Settings → Privacy & Security → Microphone → the toggle next to Whisperer is on.
  3. Look at the waveform indicator. The overlay's CommandBar has a volume/waveform indicator. As you speak, it should react to your voice; when the other person speaks, there should be movement too. No reaction on one side means the corresponding source isn't flowing (see steps 1–2).
  4. Make sure the session is running and not paused. In the CommandBar, the play/pause button should be in recording mode. While paused, audio isn't captured.
  5. Check the input device. If you have several microphones (built-in, headset, webcam), make sure the working microphone is selected in the system and on the call. A heavily noisy or muted microphone gives an empty [Me] track.
  6. Check noise suppression. The overlay settings have noise suppression. If speech is quiet and gets "eaten", try easing it off/turning it off; if there's a lot of background noise, do the opposite and turn it on.
  7. Check the transcription language. The language is set per session (default ru). If the call is in another language but is recognized as Russian, the words will be garbled. Set the correct transcription language (Whisper is multilingual) and start the session again.
  8. Restart the session/app. If something "hung up" after changing permissions or the device, end the session, restart Whisperer, and start over.

Screenshots

📸 [Screenshot: a transcript with only [Me] and no [Other] — the typical sign that "Screen Recording" is missing]

📸 [Screenshot: the waveform indicator in the CommandBar during active speech]

📸 [Screenshot: the overlay settings — noise suppression and transcription language]

Common mistakes

  • Can't hear the other person (macOS) → "Screen Recording" not granted. The microphone gives only your voice; the other person's voice is system audio, available only through screen recording.
  • Can't hear the other person (Windows) → wrong default output device. System audio capture only takes audio from the default output device. Set the device the call plays through as the default (see Windows permissions).
  • Permission granted, but no audio → the app wasn't restarted. macOS applies "Screen Recording" only after a restart.
  • Text is "gibberish" → wrong language. A mismatch between the session language and the call's actual language breaks recognition. The language is set per session.
  • Empty [Me] → wrong/muted microphone. Check the selected input device and that the microphone isn't muted on the call.
  • Waiting to upload a recording. Whisperer transcribes only in real time; you can't upload a finished audio file — the audio must flow during the session.

Best practices

  • Before an important call, do a 30-second trial session and make sure both roles — [Me] and [Other] — appear in the transcript.
  • Remember the mnemonic: [Me] = Microphone, [Other] = Screen Recording. This points straight at which permission to fix.
  • Set the transcription language to match the meeting's language ahead of time.
  • Use a stable microphone (a headset) and don't change the input device in the middle of a session.

Related articles