InterDigital and Futuresource Consulting have published a new whitepaper looking at new opportunities for AI and Machine Learning in the video industry.
The report states that both AI and ML offer the potential to augment existing video encoding methods to reduce file sizes and bit rates whilst maintaining visual quality.
It suggests ML techniques are beginning to influence new solutions for the video industry, “though its benefits are not always universal or immediately tangible,” it adds.
Research shows machine learning will play a role in defining advanced video coding mechanisms, even though the significant complexity in encode and decode operations make traditional tools more efficient than AI-based alternatives, states the report, and cites how the Versatile Video Coding (VVC) standard offers a roughly 40 per cent compression improvement over HEVC but carries a tenfold increase in encoding complexity.
“In dynamic frame rate encoding, the AI aims to encode video at the minimum frame rates necessary to encapsulate the content without sacrificing quality,” states the report. “News broadcasts might be encoded at 30fps, whereas live sports content benefits from 60fps. Using machine learning to train AI to identify the type of content, it’s possible to significantly reduce the encode compute requirements, approaching a 30 per cent reduction overall for content with limited movement between frames.”
However, the research adds that while both AI and ML can help define advanced video coding mechanisms, traditional coding tools currently still outperform AI-based alternatives in most areas.
“As AI-based methods continue to evolve, it is abundantly clear that the technology will become deeply entrenched in video encoding and decoding solutions. As highlighted here, AI and machine learning are likely to become vital elements that enable a commercially viable 8K streaming or broadcast TV service,” states the whitepaper.
The full whitepaper is available to download here.