The formation of Moving Picture, Audio, and Data Coding by Artificial Intelligence (MPAI) was announced in July 2020 and we reported on it in Mid-July. It is planned to be set up as a non-profit organization by the end of September 2020. It will develop technical specifications of data coding, especially using artificial Intelligence, and their integration in Information and Communication Technology systems, bridging the gap between its technical specifications and their practical use through Intellectual Property Rights Guidelines, such as Framework Licenses.
MPAI has just announced its first use case—Context-based Audio Enhancement (MPAI-CAE)—has reached sufficient maturity to warrant the start of the next stage where detailed functional requirements are identified.
MPAI-CAE addresses a variety of consumer-oriented use cases, e.g., entertainment, voice communication, audio conferencing, gaming, etc., relevant to different contexts—e.g., at home, in the car, and on the go—that may greatly influence the audio experience. MPAI-CAE also addresses professional applications such as audio (post-)production and restoration.
The MPAI-CAE standard will specify:
- Input and output interfaces for a set of contexts
- Interfaces of updatable and extensible processing modules, both traditional and AI-based, to create processing pipelines for possibly partly local and partly on-the-cloud execution
- Interfaces of the processed audio signals to a variety of delivery protocols.
|Functional block diagram of MPAI. (Source: MPAI)|
MPAI envisages that technology providers will benefit from a wider usage of their technologies beyond their specific domains; application vendors adopting the emerging MPAI-CAE standard will be able to tap from the common set of technologies to support their specific needs; service providers will benefit from an accelerated delivery by being able to third parties’ components from both a technical and licensing perspective and end-users will be able to tap from a competitive market providing constantly improved user experiences and AI-based audio endpoints.
MPAI is investigating several other draft projects that are expected to become standard development projects as they mature.
MPAI is a non-profit, unaffiliated association whose goal is to establish a set of standards for advanced audio, video, and data coding using artificial intelligence and to establish procedures that facilitate the timely and effective use of the standards it develops.
Any entity, such as corporation and individual firm, partnership, university, governmental body or international organization supporting the mission of MPAI may apply for membership, provided that it is able to contribute to the development of technical specifications for the efficient use of data.