Crate voirs_acoustic 

VoiRS Acoustic Models

Neural acoustic models for converting phonemes to mel spectrograms. Supports VITS, FastSpeech2, and other state-of-the-art architectures.

Re-exports

pub use backends::Backend;

pub use backends::BackendManager;

pub use batch_processor::BatchProcessingStats;

pub use batch_processor::BatchProcessor;

pub use batch_processor::BatchProcessorConfig;

pub use batch_processor::BatchProcessorTrait;

pub use batch_processor::BatchRequest;

pub use batch_processor::ErrorStats;

pub use batch_processor::MemoryStats;

pub use batch_processor::QueueStats;

pub use batch_processor::RequestPriority;

pub use memory::lazy::ComponentRegistry;

pub use memory::lazy::LazyComponent;

pub use memory::lazy::MemmapFile;

pub use memory::lazy::MemoryPressureHandler;

pub use memory::lazy::MemoryPressureLevel;

pub use memory::lazy::MemoryPressureStatus;

pub use memory::lazy::ProgressiveLoader;

pub use memory::AdvancedPerformanceProfiler;

pub use memory::MemoryOptimizer;

pub use memory::OperationTimer;

pub use memory::PerformanceMetrics;

pub use memory::PerformanceMonitor;

pub use memory::PerformanceReport;

pub use memory::PerformanceSnapshot;

pub use memory::PerformanceThresholds;

pub use memory::PoolStats;

pub use memory::ResultCache;

pub use memory::SystemInfo;

pub use memory::SystemMemoryInfo;

pub use memory::TensorMemoryPool;

pub use metrics::EvaluationConfig;

pub use metrics::EvaluationPreset;

pub use metrics::MetricStatistics;

pub use metrics::ObjectiveEvaluator;

pub use metrics::ObjectiveMetrics;

pub use metrics::PerceptualEvaluator;

pub use metrics::PerceptualMetrics;

pub use metrics::ProsodyEvaluator;

pub use metrics::ProsodyFeatures;

pub use metrics::ProsodyMetrics;

pub use metrics::QualityEvaluator;

pub use metrics::QualityMetrics;

pub use metrics::QualityStatistics;

pub use metrics::RhythmFeatures;

pub use metrics::WindowType;

pub use models::DummyAcousticConfig;

pub use models::DummyAcousticModel;

pub use models::ModelLoader;

pub use optimization::DistillationConfig;

pub use optimization::DistillationStrategy;

pub use optimization::HardwareOptimization;

pub use optimization::HardwareTarget;

pub use optimization::ModelOptimizer;

pub use optimization::OptimizationConfig;

pub use optimization::OptimizationMetrics;

pub use optimization::OptimizationReport;

pub use optimization::OptimizationTargets;

pub use optimization::PruningConfig;

pub use optimization::PruningStrategy;

pub use optimization::PruningType;

pub use optimization::QuantizationConfig as OptQuantizationConfig;

pub use optimization::QuantizationMethod as OptQuantizationMethod;

pub use optimization::QuantizationPrecision as OptQuantizationPrecision;

pub use prosody::DurationConfig;

pub use prosody::EnergyConfig;

pub use prosody::EnergyContourPattern;

pub use prosody::IntonationPattern;

pub use prosody::PauseDurations;

pub use prosody::PitchConfig;

pub use prosody::ProsodyAdjustment;

pub use prosody::ProsodyConfig;

pub use prosody::ProsodyController;

pub use prosody::RhythmPattern;

pub use prosody::VibratoConfig;

pub use prosody::VoiceQualityConfig;

pub use quantization::ModelQuantizer;

pub use quantization::QuantizationBenchmark;

pub use quantization::QuantizationConfig;

pub use quantization::QuantizationMethod;

pub use quantization::QuantizationParams;

pub use quantization::QuantizationPrecision;

pub use quantization::QuantizationStats;

pub use quantization::QuantizedTensor;

pub use simd::Complex;

pub use simd::FftWindow;

pub use simd::SimdAudioEffects;

pub use simd::SimdAudioProcessor;

pub use simd::SimdCapabilities;

pub use simd::SimdDispatcher;

pub use simd::SimdFft;

pub use simd::SimdLinearLayer;

pub use simd::SimdMatrix;

pub use simd::SimdMelComputer;

pub use simd::SimdStft;

pub use simd::StftWindow;

pub use simd::WindowFunction;

pub use singing::ArticulationMarking;

pub use singing::BreathControlConfig;

pub use singing::DynamicsMarking;

pub use singing::FormantAdjustment;

pub use singing::KeySignature;

pub use singing::MusicalNote;

pub use singing::MusicalPhrase;

pub use singing::ResonanceConfig;

pub use singing::SingingConfig;

pub use singing::SingingTechnique;

pub use singing::SingingVibratoConfig;

pub use singing::SingingVoiceSynthesizer;

pub use singing::VocalRegister;

pub use singing::VoiceType;

pub use speaker::Accent;

pub use speaker::AgeGroup;

pub use speaker::AudioFeatures;

pub use speaker::AudioReference;

pub use speaker::CloningQualityMetrics;

pub use speaker::CrossLanguageSpeakerAdapter;

pub use speaker::EmotionConfig;

pub use speaker::EmotionModel;

pub use speaker::EmotionType;

pub use speaker::FewShotSpeakerAdaptation;

pub use speaker::Gender;

pub use speaker::MultiSpeakerConfig;

pub use speaker::MultiSpeakerModel;

pub use speaker::PersonalityTrait;

pub use speaker::SpeakerEmbedding;

pub use speaker::SpeakerId;

pub use speaker::SpeakerMetadata;

pub use speaker::SpeakerRegistry;

pub use speaker::SpeakerVerificationResult;

pub use speaker::SpeakerVerifier;

pub use speaker::VoiceCharacteristics;

pub use speaker::VoiceCloningConfig;

pub use speaker::VoiceCloningQualityAssessor;

pub use speaker::VoiceQuality;

pub use streaming::LatencyOptimizer;

pub use streaming::LatencyOptimizerConfig;

pub use streaming::LatencyStats;

pub use streaming::LatencyStrategy;

pub use streaming::PerformanceMeasurement;

pub use streaming::PerformancePredictor;

pub use streaming::StreamingConfig;

pub use streaming::StreamingMetrics;

pub use streaming::StreamingState;

pub use streaming::StreamingSynthesizer;

pub use traits::AcousticModel;

pub use traits::AcousticModelFeature;

pub use traits::AcousticModelMetadata;

pub use vits::TextEncoder;

pub use vits::TextEncoderConfig;

pub use vits::VitsConfig;

pub use vits::VitsModel;

pub use vits::VitsStreamingState;

pub use latency_optimizer::ChunkStrategy;

pub use latency_optimizer::LatencyBudget;

pub use latency_optimizer::LatencyMeasurement;

pub use latency_optimizer::LatencyOptimizer as AdvancedLatencyOptimizer;

pub use latency_optimizer::LatencyStatistics;

pub use latency_optimizer::ProcessingPriority;

pub use neural_codec::CodecQualityMetrics;

pub use neural_codec::CodecType;

pub use neural_codec::NeuralCodec;

pub use neural_codec::NeuralCodecConfig;

pub use vad::VadConfig;

pub use vad::VadSegment;

pub use vad::VoiceActivity;

pub use vad::VoiceActivityDetector;

pub use config::*;

pub use mel::*;

Modules

acoustic_utils

Acoustic Processing Utilities

backends

Backend implementations for acoustic models

batch_processor

Advanced Batch Processing System for VoiRS Acoustic Models

batching

Dynamic batching system for variable-length sequences

cache

Advanced caching strategies for acoustic synthesis

conditioning

Conditional layers for feature-controlled synthesis.

config

Configuration management for acoustic models

diagnostics

Advanced diagnostics and analysis tools for acoustic modeling

error

Enhanced error handling with diagnostic context and recovery suggestions

fastspeech

FastSpeech2 implementation.

fastspeech2_trainer

FastSpeech2 model training infrastructure

fusion

Kernel Fusion Optimization Module

latency_optimizer

Advanced Latency Optimization for Real-Time TTS

mel

Mel spectrogram computation and processing

memory

Advanced memory management utilities for efficient acoustic model inference

metrics

Audio quality metrics for TTS evaluation

model_manager

Auto-generated module structure

model_warmup

Model warmup and preloading utilities

models

Acoustic model definitions and management.

neural_codec

Neural Audio Codec Integration

optimization

Model optimization techniques for efficient inference

parallel_attention

Parallel attention computation for improved performance

performance_targets

Performance targets monitoring and validation system

prelude

Prelude for convenient imports

production

Production hardening features for reliable acoustic synthesis

production_monitoring

Advanced production monitoring and observability

profiling

Advanced Performance Profiling and Tracing System

profiling_integration

Integration between Performance Profiling and Production Monitoring

prosody

Prosody control for natural speech synthesis.

quantization

Model quantization utilities for compression and optimization

scirs2_ops

SciRS2-Optimized Acoustic Operations

simd

SIMD-accelerated operations for acoustic processing

singing

Singing voice synthesis functionality for acoustic models.

speaker

Speaker control and voice characteristics management.

streaming

Streaming synthesis infrastructure for real-time text-to-speech

synthesis_cache

Advanced synthesis result caching system

traits

Core traits for acoustic models

unified_conditioning

Unified conditioning interface for all synthesis features.

utils

Utility functions for acoustic modeling.

vad

Voice Activity Detection (VAD) Integration

vits

VITS (Variational Inference Text-to-Speech) model implementation

Structs

AcousticModelManager

Acoustic model manager with multiple architecture support

MelSpectrogram

Mel spectrogram representation

Phoneme

A phoneme with its symbol and optional features

SynthesisConfig

Simple synthesis configuration for basic operations

Enums

AcousticError

Acoustic model specific error types with enhanced diagnostic information

LanguageCode

Language codes supported by VoiRS

Type Aliases

Result

Result type for acoustic model operations