Leonardo Chiariglione – Co-founder of MPEG

Patents, MPEG, and Innovation

• Many see MPEG’s patent-heavy regime (MP3, H.264, H.265, etc.) as having slowed innovation and adoption for decades, creating “patent minefields” that deter products and individual researchers.
• Multiple overlapping patent pools with incompatible terms are viewed as rent-seeking, raising costs and legal risk for even small incremental improvements.
• The notorious MPEG-LA AVC license text (non‑commercial only) illustrates how poorly licensing terms match modern real-world uses like videoconferencing.
• Some argue patent pools were originally meant to enable collaboration under existing IP law, but became bloated, political, and counter to technical simplicity.

Royalty-Free Codecs and AOM/Google

• AV1, VP9, Opus, Vorbis, Daala, etc. are cited as important royalty‑free alternatives, heavily driven by Google and allied companies.
• There’s debate over whether these are truly “free”: some link to industry FUD pieces about AV1; others note that several pools are now trying to claim AV1 royalties anyway.
• Critics of MPEG note that H.265’s fractured patent pools and HEVC/VVC licensing mess are precisely what pushed large players toward AV1.
• Others counter that royalty‑free codecs only gained traction once they matched or beat MPEG quality, and that expecting permanent corporate subsidization is fragile.

Economics and Who Funds Codecs

• Codec R&D is described as expensive and specialized: hundreds of highly paid engineers per standard, huge compute clusters, complex hardware considerations, and extensive testing.
• Supporters of patents say this scale requires strong financial incentives; without IP, they predict under‑investment and stagnation.
• Opponents respond with analogies to Linux, web servers, GCC, FLAC, LAME, Xiph, and hobby formats (QOI/QOA) as evidence that non‑patent and mixed‑funding models can work.
• Academia and government funding are proposed as alternatives; critics note universities often patent aggressively and typically stop at non‑productized proofs of concept, though MP3 is cited as a successful publicly funded example.

MPEG Governance, Collapse, and MPAI

• Several comments say “obscure forces” are simply large patent owners gaming MPEG’s FRAND regime: pushing many overlapping patents into standards, blocking royalty‑free efforts, and then splintering pools.
• Some hold the MPEG leadership responsible for enabling a system where any minor patent holder could hold an entire standard hostage.
• Others emphasize that the founder later declared MPEG’s business model “broken,” criticized patent‑pool dysfunction, and left to found MPAI with stricter pre‑agreed licensing frameworks, though its practical impact is unclear.

AI, Compression, and Future Codecs

• Multiple commenters note the close relationship between compression and prediction/AI: “every predictor is a compressor, every compressor is a predictor.”
• Deep‑learning codecs (e.g., Microsoft’s DCVC claiming better-than‑H.266 performance) are mentioned as promising, though likely to raise their own IP issues.
• Some speculate on edge‑AI and auto‑generated codecs for specific hardware; others remain skeptical that this will escape the current patent dynamics.

Broader IP Critiques and Proposals

• Strong antipathy toward software and codec patents is common; some advocate banning algorithm/software patents entirely, or even all IP.
• A proposed reform: tax patents based on self‑declared value and allow forced purchase at that price, to discourage hoarding and trolling.
• One commenter notes that, in practice, improving networks and storage often outpaces codec gains, further weakening economic incentives for ever‑more-complex, patent‑encumbered standards.