Crate oximedia_audio_analysis 

Advanced audio analysis and forensics for OxiMedia.

This crate provides comprehensive audio analysis capabilities for professional audio applications including forensics, voice analysis, music analysis, and more.

Features

Spectral Analysis

• Advanced frequency-domain analysis with multiple window functions
• Spectral centroid, flatness, crest factor, and bandwidth computation
• High-resolution spectral features for detailed audio characterization

Voice Analysis

• Voice characteristic analysis (F0, formants, jitter, shimmer, HNR)
• Gender detection using formant analysis and F0 range
• Age estimation based on voice characteristics
• Emotion detection (anger, joy, sadness, neutral)
• Speaker identification and verification

Music Analysis

• Harmonic analysis and chord progression detection
• Advanced rhythmic analysis extending MIR capabilities
• Timbral analysis for sound quality characterization
• Instrument identification using spectral and temporal features

Source Separation

• Vocal/instrumental separation using harmonic-percussive decomposition
• Drum track isolation
• Bass line extraction
• Multi-source separation

Echo and Reverb Analysis

• Echo and reverb detection
• Room acoustics analysis
• RT60 reverberation time measurement
• Early reflection pattern analysis

Distortion Analysis

• Distortion detection and quantification
• Total Harmonic Distortion (THD) measurement
• Clipping detection with threshold analysis
• Non-linear distortion characterization

Dynamic Range Analysis

• Detailed dynamic range computation
• Crest factor analysis
• RMS level tracking over time
• Loudness variation measurement

Transient Detection

• Transient and attack detection
• Envelope analysis with ADSR characterization
• Onset strength function computation

Pitch Analysis

• Pitch tracking using YIN algorithm (patent-free)
• Pitch contour analysis
• Vibrato detection and measurement
• F0 estimation with confidence scoring

Formant Analysis

• Formant frequency analysis (F1-F4)
• Formant tracking over time
• Vowel detection and classification
• Linear Predictive Coding (LPC) for formant extraction

Audio Forensics

• Audio authenticity verification
• Edit detection (cuts, splices, insertions)
• Compression history analysis
• Background noise consistency analysis
• ENF (Electrical Network Frequency) analysis

Noise Analysis

• Noise profiling and characterization
• Noise type classification (white, pink, environmental)
• Signal-to-noise ratio (SNR) computation
• Noise floor estimation

Usage Example

use oximedia_audio_analysis::{
    AudioAnalyzer, AnalysisConfig,
};

// Create analyzer with default configuration
let config = AnalysisConfig::default();
let analyzer = AudioAnalyzer::new(config);

// Analyze audio samples
let samples = vec![0.0_f32; 44100]; // 1 second of audio
let sample_rate = 44100.0;

let result = analyzer.analyze(&samples, sample_rate)?;

// Access spectral features
println!("Spectral centroid: {:.1} Hz", result.spectral.centroid);
println!("Spectral flatness: {:.3}", result.spectral.flatness);

// Access voice characteristics
if let Some(voice) = result.voice {
    println!("F0: {:.1} Hz", voice.f0);
    println!("Gender: {:?}", voice.gender);
}

Patent-Free Implementation

All algorithms are implemented using patent-free methods:

• YIN algorithm for pitch detection
• LPC for formant analysis
• Harmonic-percussive separation for source separation
• Autocorrelation-based methods

Real-Time Capable

Most analysis modules support frame-by-frame processing for real-time applications.

Modules

beat

Beat tracking and tempo estimation for audio analysis.

cepstral

Cepstral analysis for audio signals.

chroma

Chromagram and tonal centroid (tonnetz) computation.

compression_analysis

Audio compression analysis: bitrate estimation, codec artefact detection, spectral hole detection, and compression history scoring.

distortion

Distortion analysis module.

dynamics

Dynamic range analysis module.

echo

Echo and reverb detection module.

energy

Signal energy analysis — RMS energy, zero-crossing rate, energy envelope, and energy-based silence detection.

energy_contour

Energy contour analysis for audio signals.

forensics

Audio forensics module for authenticity verification.

formant

Formant analysis module using Linear Predictive Coding (LPC).

formant_track

Formant frequency tracking over time.

harmony

Harmonic analysis: pitch classes, musical keys, chords, and key detection.

loudness

Short-term and integrated loudness metering (inspired by EBU R128 / ITU-R BS.1770).

loudness_curve

Time-varying loudness curve analysis.

loudness_range

Loudness range (LRA) analysis conforming to EBU R 128 / ITU-R BS.1770.

mel_spectrogram

Mel spectrogram computation for ML-oriented audio feature extraction. Mel spectrogram computation for ML feature extraction.

music

Music analysis module (extends oximedia-mir).

noise

Noise analysis and profiling module.

onset

Onset detection for audio analysis.

pitch

Pitch analysis module using YIN algorithm.

pitch_detect

Pitch detection via autocorrelation and the YIN algorithm.

pitch_tracker

Pitch class tracking and pitch history analysis.

psychoacoustic

Psychoacoustic modeling for audio analysis.

rhythm

Rhythm analysis: tempo estimation, beat grid, time signature detection.

scene_classify

Audio scene classification: Indoor, Outdoor, Quiet, Noisy, Speech, Music, Mixed. Audio scene classification using spectral features.

separate

Source separation module using harmonic-percussive decomposition.

silence_detect

Silence detection and segmentation for audio streams.

spectral

Spectral analysis module for frequency-domain audio analysis.

spectral_contrast

Spectral contrast analysis for audio characterisation.

spectral_features

High-level spectral feature computation: centroid, flux, rolloff.

spectral_flux

Spectral flux onset detection.

stereo_field

Stereo field analysis.

tempo_analysis

Tempo analysis for audio signals.

timbre

Timbre analysis: MFCC coefficients, spectral centroid, brightness, roughness.

transient

Transient detection module.

voice

Voice analysis module.

Structs

AnalysisConfig

Configuration for audio analysis.

AnalysisResult

Complete analysis result.

AudioAnalyzer

Main audio analyzer that coordinates all analysis modules.

FrameAnalysis

Frame-level analysis result for real-time processing.

Enums

AnalysisError

Errors that can occur during audio analysis.

WindowType

Window function types for spectral analysis.

Functions

amplitude_to_db

Convert amplitude to decibels.

compute_rms

Compute RMS (Root Mean Square) level of audio samples.

db_to_amplitude

Convert decibels to amplitude.

generate_window

Generate window function of the specified type and size.

zero_crossing_rate

Compute zero-crossing rate.

Type Aliases

Result

Result type for audio analysis operations.