Context-Aware Generative Audio Engine

You are the audio systems director for [INSERT GAME TITLE]; design a generative audio engine creating adaptive music and sound effects from live scene context. Outputs must support [INSERT AUDIO MIDDLEWARE], avoid rights conflicts, and perform within strict CPU and latency budgets. Module 1 - Gameplay design: define an emotional music taxonomy, interaction motifs, stinger rules, vertical remixing and horizontal sequencing principles, and silence policy. Module 2 - Generation pipeline: specify tension extraction, adaptive stem composition, real-time SFX synthesis from material/action descriptors, and optional voice style transfer for generated NPC lines. Include deterministic seeds, an authored-foundation plus generated-variations hierarchy, and licensed-sample provenance enforcement. Define latency tiers: pre-rendered stems, sub-80ms reactive synthesis, and offline batch jobs for UGC scene soundtracks. Explain accessibility: visual cue pairing, ducking, and hearing-impaired signal layers. Define age- and region-appropriate rules for sung or spoken generation, with explicit voice-consent records. Specify how UGC creators control the adaptive envelope of their own scenes. Module 3 - Data structures: provide schemas for audio contexts, stem graphs, synth recipes, voice consent records, and licensing metadata. Module 4 - Testing: define loudness and true-peak compliance, repetition-fatigue studies, emotional-intent matching, profiling, and rights-filter tests. Module 5 - Operations: cover package size, per-user caching, opt-outs, creator marketplace terms, monitoring, and takedown workflows. Provide a text architecture diagram, three context-to-audio scenarios, risk register, and KPIs for immersion and fatigue. List assumptions about platforms, middleware, and whether voice cloning is in scope. Begin with a one-page audio vision summary, then deliver each module one at a time for approval. ================