The Google Yamnet model runs on the ONNX Runtime accelerated by DirectML; ready-to-drop into a project for 64-bit Windows, providing fast speech-to-text with minimal setup, no extra downloads.This is an audio tool that is not affected by SRP compatiblity.This package currently supports 64-bit Windows for the Unity Editor and Standalone.Yamnet (OnnxRuntime/DirectML) is a turnkey Windows x64 Unity package that ships a single native DLL containing Google’s YAMNet audio event classifier accelerated by ONNX Runtime + DirectML. It works offline in the Editor and Windows Standalone builds, enabling fast, local audio-scene understanding from clips or microphone input. Drop it into your project to detect more than 500 everyday sound classes (speech, music, sirens, footsteps, barks, ambience, and more) with minimal setup, plus ready-made example scenes and UI.Summary- Single native DLL with YAMNet on ONNX Runtime + DirectML (offline, Windows x64).- Audio event classification over 521 classes; returns top-K labels and scores.- Works with audio clips (StreamingAssets) and live microphone input.- 16 kHz mono PCM pipeline with lightweight C# wrappers for inference.- Example scenes for clip-based and live/continuous classification.User scenarios- Add contextual sound awareness to gameplay (react to music, alerts, footsteps, animal sounds).- Build accessibility and safety features that respond to ambient cues without cloud latency.- Batch-score audio archives locally to tag or filter content on air-gapped machines.- Used AI to draft conversion scripts and ONNX Runtime wiring for YAMNet + DirectML.- Assisted with Unity scripting to hook UI buttons, microphone capture, and example scenes.- Helped outline testing; all code was manually reviewed and tested.- AI aided marketing copy and image ideation; final assets were hand-edited for quality.
