Audio normalization balances human perceived loudness rather than simple electrical peak voltages:

• Why Peak Normalization Fails: Two songs with identical 0 dBFS peak spikes can sound drastically different in volume. A dynamic classical piece with one loud drum hit will have quiet speech throughout, while a heavily compressed pop song will sound four times louder.
• ITU-R BS.1770 / EBU R128 Loudness (LUFS / LKFS): Measures perceived loudness using K-weighting frequency filters (modeling human ear sensitivity to mid-range frequencies between 1 kHz and 5 kHz) with gating to exclude prolonged silent passages.
• Standard Streaming Platform Loudness Targets:
• Spotify & YouTube: $-14.0\text{ LUFS}$ ($\pm 1.0\text{ LUFS}$, Peak ceiling: $-1.0\text{ dBFS}$)
• Apple Podcasts & Amazon Music: $-16.0\text{ LUFS}$ (Industry podcast standard)
• EBU R128 (European Television & Radio): $-23.0\text{ LUFS}$
• ATSC A/85 (US Television Broadcast): $-24.0\text{ LUFS}$
• Automated Gain Calculation: Calculates required gain adjustment: $\Delta G = \text{Target LUFS} - \text{Integrated LUFS}$, scaling the buffer and applying true-peak limiting to prevent inter-sample clipping.