Murakami Compressor Suite treats media compression as an optimization problem—compression as search—instead of selecting fixed presets.It uses a practical Probe→Final strategy: evaluate many candidates cheaply on a probe (downscaled images or the first seconds of audio/video), then perform the full encode only once using the best candidate. Quality is constrained by perceptual metrics: SSIM for images, and VMAF (or SSIM) for video probes when available via FFmpeg.Exploration is improved by (1) quantum-inspired phase interference sampling (amplitude accumulation and |amp|²-based selection) and (2) a QAOA(p=1) circuit simulation that converts an energy landscape C(z) over discrete codec parameters into a proposal distribution P(z) to prioritize top-K candidates. Important: this work does not claim quantum speedup. The QAOA component is a classical statevector simulation used to bias candidate selection, combined with heuristic/random exploration. Keywords: compression-as-search, Probe→Final, SSIM, VMAF, FFmpeg, quantum-inspired sampling, QAOA(p=1) classical simulation.
Hideyoshi Murakami (Thu,) studied this question.