The previous document Domain and Control demonstrated true digital-first multimedia coordination where computational domains interact through shared timing references, data transformation pipelines, and temporal coordination mechanisms. Rather than separating audio, visual, and temporal processing into isolated systems, MayaFlux treats them as unified aspects of a single computational-creative expression.

The Engine's default automation provides optimal performance for common workflows, while explicit control mechanisms enable precise customization when specific computational patterns are required. This default automation with expressive override philosophy ensures that creative flow remains uninterrupted while advanced techniques remain accessible when needed.

This document explores replacing Engine defaults with custom managers, handlers, subsystems, backends, and the default Engine itself—enabling complete architectural customization while maintaining the digital-first computational paradigm.

Note Many currently non existent subsystems will me mentioned to showcase possible actions, workflows and overrides. All referenced methods are available in the ISubsystem interface, but the specific concrete class may not be part of the default library.

Note This document offers less textural or verbose explanations as most low level c++ code itself contains explanations within its structure and via comments. Textual context is added only where it is necessary to describe intent of operation beyond whats made obvious by the code. As the title suggests, this is an advanced section.

Processing Handles: Token-Scoped System Access
- BufferProcessingHandle
  - Direct handle creation
- NodeProcessingHandle
- TaskSchedulerHandle
- SubsystemProcessingHandle
  - Handle composition and custom contexts
- Why Processing Handles Instead of Direct Manager Access?
  - Handle Performance Characteristics
SubsystemManager: Computational Domain Orchestration
- Subsystem Registration and Lifecycle
  - Cross-Subsystem Data Access Control
  - Processing Hooks and Custom Integration
  - Direct SubsystemManager Control
Subsystems: Computational Domain Implementation
- ISubsystem Interface and Implementation Patterns
- Specialized Subsystem Examples
  - AudioSubsystem
  - GraphicsSubsystem
  - Subsystem Communication and Coordination
- Direct Subsystem Management
Backends: Hardware and Platform Abstraction
- Audio Backend Architecture
  - IAudioBackend Interface
  - Backend Factory Expansion
- Graphics Backend Architecture
  - IGraphicsBackend Interface
  - Custom Rendering Pipeline
- Network Backend for Distributed Processing%

Processing Handles: Token-Scoped System Access

Processing handles provide token-scoped access to the core processing systems (Nodes, Buffers, Coroutines), ensuring proper domain isolation and thread safety. Rather than direct access to managers, handles act as computational interfaces that enforce processing boundaries while enabling cross-domain coordination.

Each handle represents a specific computational context defined by processing tokens, providing controlled access to the underlying systems while maintaining the engine's coordination patterns.

BufferProcessingHandle

Provides token-scoped access to buffer operations within a specific processing domain:

// Engine-managed buffer handle access
auto audio_subsystem = get_context()->get_subsystem_manager()->get_audio_subsystem();
auto buffer_handle = audio_subsystem->get_processing_context_handle()->buffers;
 
// Token-scoped buffer operations
auto audio_buffer = buffer_handle.create_buffer<AudioBuffer>(0, 1024);
buffer_handle.process(512);  // Process 512 samples for all channels
 
// Handle respects token boundaries
auto buffer_data = buffer_handle.get_buffer_data(0);  // Only AUDIO_BACKEND token buffers

The handle automatically routes operations through the appropriate token-specific processing paths, ensuring that buffer operations remain within their designated computational domain while enabling optimal performance.

Direct handle creation

Handles can be created independently for custom processing scenarios:

// Create custom buffer handle with specific token
auto custom_handle = BufferProcessingHandle(
    custom_buffer_manager,
    Buffers::ProcessingToken::CUSTOM | Buffers::ProcessingToken::PARALLEL
);
 
// Use handle for domain-specific operations
custom_handle.add_buffer(gpu_accelerated_buffer);
custom_handle.process(1024);
 
// Handle maintains token constraints
auto gpu_buffer_data = custom_handle.read_channel_data(0);  // Only GPU_PROCESS token buffers

NodeProcessingHandle

Provides token-scoped access to node graph operations within specific processing domains:

// Subsystem-managed node handle
auto graphics_subsystem = get_context()->get_subsystem_manager()->get_graphics_subsystem();
auto node_handle = graphics_subsystem->get_processing_context_handle()->nodes;
 
// Token-specific node operations
auto visual_output = node_handle.process_channel(0, 1024);  // VISUAL_RATE processing
auto sample_output = node_handle.process_sample(0);         // Single sample processing
 
// Handle coordinates with appropriate token processor
node_handle.register_node(visual_processor_node, 0);  // Register to VISUAL_RATE channel 0

Node handles ensure that processing operations respect the temporal characteristics of their assigned domain while providing unified access patterns across different processing contexts.

TaskSchedulerHandle

Provides token-scoped access to coroutine scheduling within specific temporal domains:

// Temporal domain handle access
auto sync_subsystem = get_engine()->get_subsystem_manager()->get_sync_subsystem();
auto task_handle = sync_subsystem->get_processing_context_handle()->tasks;
 
// Token-specific temporal processing
task_handle.process(1024);  // Process 1024 temporal units for this token
 
// Handle manages domain-specific scheduling
auto routine = create_sync_routine();
task_handle.add_task(routine, "cross_domain_sync");

SubsystemProcessingHandle

Combines all three processing handles into a unified computational context for subsystem operations:

// Unified subsystem processing context
class CustomMultimediaSubsystem : public ISubsystem {
private:
    SubsystemProcessingHandle* m_handle;
 
public:
    void initialize(SubsystemProcessingHandle& handle) override {
        m_handle = &handle;
 
        // Access all processing systems through unified handle
        auto audio_buffer = m_handle->buffers.create_buffer<AudioBuffer>(0, 512);
        auto visual_node = create_visual_processor();
        m_handle->nodes.register_node(visual_node, 0);
 
        // Coordinate temporal processing
        auto sync_routine = create_coordination_routine();
        m_handle->tasks.add_task(sync_routine, "multimedia_sync");
    }
 
    void process_multimedia_frame() {
        // Coordinated processing across all domains
        m_handle->buffers.process_all_buffers(512);
        auto visual_data = m_handle->nodes.process_channel(0, 512);
        m_handle->tasks.process(512);
 
        // Cross-domain data coordination
        coordinate_audio_visual_data(visual_data);
    }
};

Handle composition and custom contexts

Processing handles can be composed for specialized computational requirements:

// Custom handle composition for advanced workflows
auto multimedia_tokens = SubsystemTokens{
    .Buffer = Buffers::ProcessingToken::AUDIO_BACKEND | Buffers::ProcessingToken::GRAPHICS_BACKEND,
    .Node = Nodes::ProcessingToken::AUDIO_RATE | Nodes::ProcessingToken::VISUAL_RATE,
    .Task = Vruta::ProcessingToken::MULTI_RATE
};
 
auto multimedia_handle = SubsystemProcessingHandle(
    buffer_manager, node_manager, task_scheduler, multimedia_tokens
);
 
// Use composed handle for cross-domain operations
multimedia_handle.buffers.transfer_data_between_domains(audio_buffer, graphics_buffer);
multimedia_handle.nodes.coordinate_processing_rates(audio_nodes, visual_nodes);
multimedia_handle.tasks.sync_temporal_domains();

Why Processing Handles Instead of Direct Manager Access?

Processing handles enforce computational boundaries that prevent:

Cross-domain memory corruption
Token privilege escalation
Uncontrolled resource access

// This enforces token boundaries automatically

auto buffer_data = buffer_handle.get_buffer_data(0); // Only AUDIO_BACKEND token buffers

Direct manager access would require manual boundary checking in every subsystem.

Handle Performance Characteristics

Token validation: O(1) bitfield operations
Cross-domain access: Single atomic read with permission check
Memory overhead: ~64 bytes per handle (pointer + token state)

// Zero-copy when tokens match
auto data = handle.get_buffer_data(0);  // Direct pointer return
 
// Copy required when crossing domains
auto cross_data = handle.read_cross_subsystem_buffer(...);  // Memcpy + validation

SubsystemManager: Computational Domain Orchestration

The SubsystemManager serves as the central coordinator for all computational domains in MayaFlux, managing subsystem lifecycle, processing handle distribution, and cross-domain data access. Rather than a simple container, it acts as a computational orchestrator that ensures proper isolation while enabling coordinated operation across different processing contexts.

Subsystem Registration and Lifecycle

// Engine creates SubsystemManager with core processing systems
auto subsystem_manager = std::make_shared<SubsystemManager>(
    node_graph_manager, buffer_manager, task_scheduler
);
 
// Automatic subsystem creation with proper handle assignment
subsystem_manager->create_audio_subsystem(stream_info, AudioBackendType::PORTAUDIO);
subsystem_manager->create_graphics_subsystem(display_config, GraphicsBackendType::VULKAN);
subsystem_manager->create_sync_subsystem(sync_config);
 
// Coordinated startup sequence
subsystem_manager->start_all_subsystems();

The manager automatically creates appropriate processing handles for each subsystem based on their token requirements, ensuring that each computational domain receives properly scoped access to the core processing systems.

Cross-Subsystem Data Access Control

// Configure cross-domain data access permissions
subsystem_manager->allow_cross_access(SubsystemType::AUDIO, SubsystemType::GRAPHICS);
subsystem_manager->allow_cross_access(SubsystemType::GRAPHICS, SubsystemType::AUDIO);
 
// Safe cross-domain data access
auto spectral_data = subsystem_manager->read_cross_subsystem_buffer(
    SubsystemType::GRAPHICS,  // Requesting subsystem
    SubsystemType::AUDIO,     // Target subsystem
    0                         // Channel
);
 
if (spectral_data) {
    apply_spectral_visualization(*spectral_data);
}

Cross-access permissions ensure that data flow between computational domains remains controlled and explicit, preventing unintended interference while enabling intentional coordination.

Processing Hooks and Custom Integration

// Register custom processing hooks for specialized workflows
subsystem_manager->register_process_hook(
    SubsystemType::AUDIO,
    "spectral_analysis",
    [](SubsystemProcessingHandle& handle) {
        // Custom pre-processing logic
        auto audio_data = handle.buffers.get_buffer_data(0);
        perform_spectral_analysis(audio_data);
    },
    HookPosition::PRE_PROCESS
);
 
subsystem_manager->register_process_hook(
    SubsystemType::GRAPHICS,
    "visual_coordination",
    [](SubsystemProcessingHandle& handle) {
        // Custom post-processing coordination
        coordinate_visual_elements();
        update_display_parameters();
    },
    HookPosition::POST_PROCESS
);

Processing hooks enable custom integration points without modifying subsystem internals, supporting the expressive override philosophy by allowing precise customization when specific computational patterns are required.

Direct SubsystemManager Control

// Custom subsystem manager configuration
auto custom_manager = std::make_shared<SubsystemManager>(
    custom_node_manager, custom_buffer_manager, custom_scheduler
);
 
// Manual subsystem registration with custom tokens
auto custom_subsystem = std::make_shared<CustomProcessingSubsystem>();
custom_manager->add_subsystem(SubsystemType::CUSTOM, custom_subsystem);
 
// Direct subsystem access and control
auto subsystem = custom_manager->get_subsystem(SubsystemType::CUSTOM);
auto processing_handle = subsystem->get_processing_context_handle();
 
// Manual processing coordination
processing_handle->buffers.process_all_buffers(1024);
processing_handle->nodes.process_channel(0, 1024);
processing_handle->tasks.process(1024);

Subsystems: Computational Domain Implementation

Subsystems represent complete computational domains that encapsulate specific processing characteristics, resource management, and coordination patterns. Each subsystem implements the ISubsystem interface, defining how it integrates with the broader MayaFlux processing architecture.

ISubsystem Interface and Implementation Patterns

class CustomAudioVisualSubsystem : public ISubsystem {
private:
    SubsystemProcessingHandle* m_handle;
    std::atomic<bool> m_running{false};
    std::atomic<bool> m_ready{false};
 
    // Domain-specific processing state
    std::unique_ptr<SpectralAnalyzer> m_spectral_analyzer;
    std::unique_ptr<VisualRenderer> m_visual_renderer;
 
public:
    void initialize(SubsystemProcessingHandle& handle) override {
        m_handle = &handle;
 
        // Initialize domain-specific components
        m_spectral_analyzer = std::make_unique<SpectralAnalyzer>();
        m_visual_renderer = std::make_unique<VisualRenderer>();
 
        // Set up processing resources
        setup_audio_visual_coordination();
        m_ready = true;
    }
 
    void start() override {
        if (!m_ready) return;
 
        m_running = true;
 
        // Start domain-specific processing
        m_spectral_analyzer->start();
        m_visual_renderer->start();
 
        // Register processing callbacks
        register_audio_callback();
        register_visual_callback();
    }
 
    SubsystemTokens get_tokens() const override {
        return {
            .Buffer = Buffers::ProcessingToken::AUDIO_BACKEND |
                     Buffers::ProcessingToken::GRAPHICS_BACKEND,
            .Node = Nodes::ProcessingToken::AUDIO_RATE |
                   Nodes::ProcessingToken::VISUAL_RATE,
            .Task = Vruta::ProcessingToken::MULTI_RATE
        };
    }
 
private:
    void setup_audio_visual_coordination() {
        // Create coordination buffers
        auto audio_buffer = m_handle->buffers.create_buffer<AudioBuffer>(0, 1024);
        auto texture_buffer = m_handle->buffers.create_buffer<TextureBuffer>(0, 1024);
 
        // Set up processing nodes
        auto fft_node = create_spectral_analyzer_node();
        auto visual_node = create_visual_processor_node();
 
        m_handle->nodes.register_node(fft_node, 0);
        m_handle->nodes.register_node(visual_node, 1);
 
        // Create coordination coroutine
        auto sync_routine = create_audio_visual_sync_routine();
        m_handle->tasks.add_task(sync_routine, "audio_visual_sync");
    }
};

Specialized Subsystem Examples

AudioSubsystem

class AudioSubsystem : public ISubsystem {
private:
    Utils::AudioBackendType m_backend_type;
    std::unique_ptr<IAudioBackend> m_backend;
    std::unique_ptr<AudioDevice> m_audio_device;
    std::unique_ptr<AudioStream> m_audio_stream;
    GlobalStreamInfo m_stream_info;
 
public:
    AudioSubsystem(GlobalStreamInfo& stream_info, Utils::AudioBackendType backend_type)
        : m_stream_info(stream_info), m_backend_type(backend_type) {
        // Create appropriate audio backend
        m_backend = AudioBackendFactory::create_backend(backend_type);
        m_audio_device = m_backend->create_device_manager();
    }
 
    void initialize(SubsystemProcessingHandle& handle) override {
        m_handle = &handle;
 
        // Initialize audio-specific processing
        setup_audio_buffers(handle);
        setup_audio_nodes(handle);
        setup_sample_accurate_timing(handle);
 
        // Create audio stream
        m_audio_stream = m_backend->create_stream(
            m_audio_device->get_default_output_device(),
            m_audio_device->get_default_input_device(),
            m_stream_info,
            this
        );
 
        m_ready = true;
    }
 
    SubsystemTokens get_tokens() const override {
        return {
            .Buffer = Buffers::ProcessingToken::AUDIO_BACKEND,
            .Node = Nodes::ProcessingToken::AUDIO_RATE,
            .Task = Vruta::ProcessingToken::SAMPLE_ACCURATE
        };
    }
 
private:
    void setup_audio_buffers(SubsystemProcessingHandle& handle) {
        // Create input/output buffers for hardware interface
        for (uint32_t channel = 0; channel < m_stream_info.output.channels; ++channel) {
            auto input_buffer = handle.buffers.create_buffer<AudioInputBuffer>(channel, 512);
            auto output_buffer = handle.buffers.create_buffer<AudioOutputBuffer>(channel, 512);
        }
    }
};

GraphicsSubsystem

class GraphicsSubsystem : public ISubsystem {
private:
    Utils::GraphicsBackendType m_backend_type;
    std::unique_ptr<IGraphicsBackend> m_backend;
    std::unique_ptr<VulkanRenderer> m_renderer;
    std::unique_ptr<DisplayManager> m_display_manager;
 
public:
    GraphicsSubsystem(DisplayConfig& display_config, Utils::GraphicsBackendType backend_type)
        : m_display_config(display_config), m_backend_type(backend_type) {
        // Create graphics backend (Vulkan, OpenGL, etc.)
        m_backend = GraphicsBackendFactory::create_backend(backend_type);
        m_display_manager = m_backend->create_display_manager();
    }
 
    void initialize(SubsystemProcessingHandle& handle) override {
        m_handle = &handle;
 
        // Initialize graphics-specific processing
        setup_texture_buffers(handle);
        setup_visual_nodes(handle);
        setup_frame_timing(handle);
 
        // Create renderer
        m_renderer = m_backend->create_renderer(m_display_config);
 
        m_ready = true;
    }
 
    SubsystemTokens get_tokens() const override {
        return {
            .Buffer = Buffers::ProcessingToken::GRAPHICS_BACKEND,
            .Node = Nodes::ProcessingToken::VISUAL_RATE,
            .Task = Vruta::ProcessingToken::FRAME_ACCURATE
        };
    }
 
private:
    void setup_texture_buffers(SubsystemProcessingHandle& handle) {
        // Create texture buffers for GPU processing
        auto color_buffer = handle.buffers.create_buffer<ColorBuffer>(0, 1920*1080);
        auto depth_buffer = handle.buffers.create_buffer<DepthBuffer>(0, 1920*1080);
        auto compute_buffer = handle.buffers.create_buffer<ComputeBuffer>(0, 1024*1024);
    }
};

Subsystem Communication and Coordination

// Cross-subsystem coordination patterns
class SynchronizationSubsystem : public ISubsystem {
private:
    SubsystemManager* m_manager;
 
public:
    void initialize(SubsystemProcessingHandle& handle) override {
        m_handle = &handle;
 
        // Set up cross-domain synchronization
        setup_audio_visual_sync(handle);
        setup_temporal_coordination(handle);
 
        m_ready = true;
    }
 
    SubsystemTokens get_tokens() const override {
        return {
            .Buffer = Buffers::ProcessingToken::CUSTOM,
            .Node = Nodes::ProcessingToken::CUSTOM_RATE,
            .Task = Vruta::ProcessingToken::MULTI_RATE
        };
    }
 
private:
    void setup_audio_visual_sync(SubsystemProcessingHandle& handle) {
        // Create sync routine that coordinates between domains
        auto sync_routine = [this, &handle](Vruta::TaskScheduler& scheduler) -> Vruta::SoundRoutine {
            while (true) {
                // Wait for audio frame completion
                co_await Kriya::ProcessingGate{scheduler,
                    [this]() { return m_manager->get_audio_subsystem()->frame_complete(); }};
 
                // Trigger visual frame processing
                auto visual_subsystem = m_manager->get_graphics_subsystem();
                visual_subsystem->trigger_frame_render();
 
                // Coordinate timing for next cycle
                co_await Kriya::SampleDelay{scheduler.seconds_to_samples(1.0/60.0)};
            }
        };
 
        handle.tasks.add_task(sync_routine(scheduler), "cross_domain_sync");
    }
};

Direct Subsystem Management

Subsystems can operate independently of the SubsystemManager when specific computational patterns require direct control:

// Direct subsystem instantiation and management
auto custom_subsystem = std::make_shared<CustomProcessingSubsystem>();
 
// Create custom processing handle
auto custom_tokens = SubsystemTokens{
    .Buffer = Buffers::ProcessingToken::CUSTOM,
    .Node = Nodes::ProcessingToken::CUSTOM_RATE,
    .Task = Vruta::ProcessingToken::CUSTOM
};
 
auto custom_handle = SubsystemProcessingHandle(
    buffer_manager, node_manager, task_scheduler, custom_tokens
);
 
// Manual subsystem lifecycle management
custom_subsystem->initialize(custom_handle);
custom_subsystem->start();
 
// Direct processing control
auto processing_handle = custom_subsystem->get_processing_context_handle();
MayaFlux::schedule_metro(0.02, [=]() {
    processing_handle->buffers.process_all_buffers(512);
    processing_handle->nodes.process_channel(0, 512);
    processing_handle->tasks.process(512);
});

Backends: Hardware and Platform Abstraction

Backends provide the abstraction layer between MayaFlux's computational architecture and platform-specific hardware or software interfaces. Each backend type handles a specific category of external interface: audio hardware, graphics systems, file I/O, networking, etc.

Audio Backend Architecture

The audio backend system abstracts different audio APIs (RtAudio, JACK, ASIO, CoreAudio, etc.) behind a unified interface:

IAudioBackend Interface

class CustomAudioBackend : public IAudioBackend {
private:
    std::unique_ptr<CustomAudioContext> m_context;
    std::unique_ptr<CustomDeviceManager> m_device_manager;
 
public:
    CustomAudioBackend() {
        m_context = std::make_unique<CustomAudioContext>();
        m_device_manager = std::make_unique<CustomDeviceManager>(m_context.get());
    }
 
    std::unique_ptr<AudioDevice> create_device_manager() override {
        return std::make_unique<CustomAudioDevice>(m_context.get());
    }
 
    std::unique_ptr<AudioStream> create_stream(
        unsigned int output_device_id,
        unsigned int input_device_id,
        const GlobalStreamInfo& stream_info,
        void* user_data) override {
 
        return std::make_unique<CustomAudioStream>(
            m_context.get(), output_device_id, input_device_id, stream_info, user_data
        );
    }
 
    std::string get_version_string() const override {
        return "CustomAudioBackend v1.0.0";
    }
 
    int get_api_type() const override {
        return static_cast<int>(Utils::AudioBackendType::CUSTOM);
    }
 
    void cleanup() override {
        if (m_context) {
            m_context->shutdown();
        }
    }
};

Backend Factory Expansion

// Extended AudioBackendFactory for custom backends
class AudioBackendFactory {
public:
    static std::unique_ptr<IAudioBackend> create_backend(Utils::AudioBackendType type) {
        switch (type) {
        case Utils::AudioBackendType::RTAUDIO:
            return std::make_unique<RtAudioBackend>();
        case Utils::AudioBackendType::JACK:
            return std::make_unique<JackAudioBackend>();
        case Utils::AudioBackendType::ASIO:
            return std::make_unique<AsioAudioBackend>();
        case Utils::AudioBackendType::CUSTOM:
            return std::make_unique<CustomAudioBackend>();
        default:
            throw std::runtime_error("Unsupported audio backend type");
        }
    }
 
    // Register custom backend at runtime
    static void register_custom_backend(
        Utils::AudioBackendType type,
        std::function<std::unique_ptr<IAudioBackend>()> factory) {
 
        custom_factories[type] = factory;
    }
 
private:
    static std::map<Utils::AudioBackendType,
                   std::function<std::unique_ptr<IAudioBackend>()>> custom_factories;
};

Graphics Backend Architecture

Similar to audio backends, graphics backends abstract different rendering APIs:

IGraphicsBackend Interface

class VulkanGraphicsBackend : public IGraphicsBackend {
private:
    std::unique_ptr<VulkanContext> m_vulkan_context;
    std::unique_ptr<VulkanRenderer> m_renderer;
 
public:
    VulkanGraphicsBackend() {
        m_vulkan_context = std::make_unique<VulkanContext>();
    }
 
    std::unique_ptr<DisplayManager> create_display_manager() override {
        return std::make_unique<VulkanDisplayManager>(m_vulkan_context.get());
    }
 
    std::unique_ptr<IRenderer> create_renderer(const DisplayConfig& config) override {
        return std::make_unique<VulkanRenderer>(m_vulkan_context.get(), config);
    }
 
    std::unique_ptr<ComputeManager> create_compute_manager() override {
        return std::make_unique<VulkanComputeManager>(m_vulkan_context.get());
    }
 
    std::string get_version_string() const override {
        return "Vulkan Backend v1.3.0";
    }
 
    void cleanup() override {
        if (m_vulkan_context) {
            m_vulkan_context->shutdown();
        }
    }
};

Custom Rendering Pipeline

// Custom renderer for specialized visual processing
class SpectralVisualizationRenderer : public IRenderer {
private:
    std::unique_ptr<VulkanComputeShader> m_fft_shader;
    std::unique_ptr<VulkanRenderPass> m_spectral_pass;
 
public:
    void initialize(const DisplayConfig& config) override {
        // Set up specialized shaders for spectral visualization
        m_fft_shader = create_fft_compute_shader();
        m_spectral_pass = create_spectral_render_pass();
    }
 
    void render_frame(const RenderData& data) override {
        // Custom rendering pipeline for audio-visual coordination
        m_fft_shader->dispatch(data.audio_buffer);
        m_spectral_pass->render(data.spectral_data);
    }
 
    void coordinate_with_audio(const AudioFrameData& audio_data) {
        // Real-time coordination with audio processing
        update_spectral_parameters(audio_data.frequency_data);
        trigger_visual_events(audio_data.onset_detection);
    }
};

Network Backend for Distributed Processing

class NetworkAudioBackend : public IAudioBackend {
private:
    std::unique_ptr<NetworkManager> m_network;
    std::unique_ptr<AudioCodec> m_codec;
 
public:
    NetworkAudioBackend(const NetworkConfig& config) {
        m_network = std::make_unique<NetworkManager>(config);
        m_codec = std::make_unique<AudioCodec>(CodecType::OPUS);
    }
 
    std::unique_ptr<AudioStream> create_stream(
        unsigned int output_device_id,
        unsigned int input_device_id,
        const GlobalStreamInfo& stream_info,
        void* user_data) override {
 
        return std::make_unique<NetworkAudioStream>(
            m_network.get(), m_codec.get(), stream_info, user_data
        );
    }
 
    // Network-specific methods
    void connect_to_peer(const std::string& address) {
        m_network->connect(address);
    }
 
    void start_collaboration_session() {
        m_network->enable_multicast();
    }
};

This comprehensive architecture enables complete computational paradigm customization while maintaining the digital-first creative philosophy. Every layer—from processing handles through backends to the engine itself—can be replaced with custom implementations that serve specific creative or technical requirements.

The three-tier architecture (Processing Handles → SubsystemManager → Subsystems) combined with replaceable backends and engine contexts provides ultimate flexibility for specialized computational workflows. Whether implementing quantum-inspired processing, neural adaptation, biomimetic algorithms, or entirely novel paradigms, the architecture supports seamless integration while preserving the unified data transformation approach that defines MayaFlux's digital-first creative environment.

The expressive override philosophy extends through every architectural level, ensuring that creative expression never encounters artificial limitations while computational precision remains accessible at every scale of interaction.