Sound effect delayed ~4-5s on first recording

**Describe the bug**
When triggering a recording via the Fn double-tap hotkey while the app is not in focus (another app is active), the sound effect plays ~4–5 seconds after the recording actually starts. On subsequent recordings (when the app was already active), the sound effect plays immediately as expected.

**To Reproduce**
1. Make sure Sound Effects are enabled in Settings.
2. Open a different app so the transcription app is not in focus (no prior recording in this session).
3. Double-tap the Fn key to trigger a recording.
4. Notice the recording starts, but the sound effect plays ~4–5 seconds later.
5. Stop the recording, then trigger a second recording — the sound effect now plays immediately.

**Expected behavior**
The sound effect should play immediately when the recording starts, regardless of whether the app was in focus or not.

**Desktop (please complete the following information):**
- Version 0.6.10

**Additional context**
The issue only occurs on the first recording after the app has been inactive / another app was in focus. Subsequent recordings behave correctly. This suggests the audio system or sound resource may need to be initialized or "warmed up" when the app regains focus.

Start by tracing the Fn double-tap recording path and sound-effects initialization in the Swift macOS app, focusing on the first recording after the app is inactive. Reproduce with Sound Effects enabled and verify that the sound plays immediately on the first recording as well as on subsequent recordings.

Written by the indexing model from the issue text.