Module quality 

Audio quality assessment with SNR estimation and spectral analysis.

This module provides multi-dimensional quality metrics for audio preprocessing validation.

Metrics

• SNR (Signal-to-Noise Ratio): Measures signal power vs noise floor (dB)
• RMS Energy: Root-mean-square energy as baseline quality indicator
• Spectral Centroid: Weighted average frequency (brightness measure)
• Quality Score: Unified score in [0.0, 1.0] combining all metrics

Performance

• Target: <10ms per second of 16 kHz audio
• Memory: Minimal allocations (reuses frame buffers)

Example

use speech_prep::preprocessing::QualityAssessor;

let assessor = QualityAssessor::new(16000);
let audio_samples = vec![0.5f32; 16000]; // 1 second at 16 kHz

let metrics = assessor.assess(&audio_samples)?;
assert!(metrics.snr_db.is_finite());

Structs

QualityAssessor

Audio quality assessor with configurable sample rate.

QualityMetrics

Quality metrics for audio assessment.