audio-normalizer（音频标准化器）

Audio Normalizer

Normalize audio volume levels using peak or RMS normalization to ensure consistent loudness across files.

Purpose

Volume normalization for:

• Podcast episode consistency
• Music playlist leveling
• Speech recording standardization
• Broadcast loudness compliance

Features

• Peak Normalization: Normalize to maximum peak level (dBFS)
• RMS Normalization: Normalize to average loudness level
• Loudness Matching: Match LUFS target for broadcast compliance
• Batch Processing: Normalize multiple files to same level
• Format Preservation: Maintain original audio format
• Headroom Control: Prevent clipping with configurable headroom

Quick Start

from audio_normalizer import AudioNormalizer

# Peak normalization to -1 dBFS
normalizer = AudioNormalizer()
normalizer.load('input.mp3')
normalizer.normalize_peak(target_dbfs=-1.0)
normalizer.save('normalized.mp3')

# RMS normalization for consistent average loudness
normalizer.normalize_rms(target_dbfs=-20.0)
normalizer.save('normalized_rms.mp3')

# Batch normalize all files to same level
normalizer.batch_normalize(
    input_files=['audio1.mp3', 'audio2.mp3'],
    output_dir='normalized/',
    method='rms',
    target_dbfs=-20.0
)

CLI Usage

# Peak normalization
python audio_normalizer.py input.mp3 --output normalized.mp3 --method peak --target -1.0

# RMS normalization
python audio_normalizer.py input.mp3 --output normalized.mp3 --method rms --target -20.0

# Batch normalize directory
python audio_normalizer.py *.mp3 --output-dir normalized/ --method rms --target -20.0

# Show current levels without normalizing
python audio_normalizer.py input.mp3 --analyze-only

API Reference

AudioNormalizer

class AudioNormalizer:
    def load(self, filepath: str) -> 'AudioNormalizer'
    def normalize_peak(self, target_dbfs: float = -1.0, headroom: float = 0.1) -> 'AudioNormalizer'
    def normalize_rms(self, target_dbfs: float = -20.0) -> 'AudioNormalizer'
    def analyze_levels(self) -> Dict[str, float]
    def save(self, output: str, format: str = None, bitrate: str = '192k') -> str
    def batch_normalize(self, input_files: List[str], output_dir: str,
                       method: str = 'rms', target_dbfs: float = -20.0) -> List[str]

Normalization Methods

Peak Normalization

• Scales audio so highest peak reaches target level
• Preserves dynamic range
• Good for preventing clipping
• Target: typically -1.0 to -3.0 dBFS

RMS Normalization

• Scales audio so average level reaches target
• Better for perceived loudness matching
• Good for podcasts and speech
• Target: typically -20.0 to -23.0 dBFS

LUFS Matching

• Integrated Loudness Units relative to Full Scale
• Broadcast standard (EBU R128, ITU BS.1770)
• Target: -23 LUFS (broadcast), -16 LUFS (streaming)

Best Practices

For Podcasts:

normalizer.normalize_rms(target_dbfs=-19.0)  # Speech clarity

For Music:

normalizer.normalize_peak(target_dbfs=-1.0)  # Preserve dynamics

For Broadcast:

normalizer.normalize_rms(target_dbfs=-23.0)  # EBU R128 compliance

Use Cases

• Podcast Production: Consistent volume across episodes
• Music Playlists: Even loudness for continuous playback
• Audiobooks: Standardized narration levels
• Conference Recordings: Normalize different speakers
• Video Production: Match audio levels before mixing

Limitations

• Does not apply dynamic compression (use separate compressor)
• Does not remove DC offset (pre-processing recommended)
• Peak normalization won't match perceived loudness
• Doesn't fix clipped audio (distortion is permanent)