Monthly Video Meetup

Please join us online tomorrow Tuesday September 24th at 1pm ET for a double feature of Kitchener Waterloo Video-Tech Meetup talks
- Hojat Yeganeh PhD, Senior Manager at IMAX Labs, discusses how he and his team developed a Film Grain Similarity objective metric which measures film grain fidelity
- Major Jiang PhD, Research Scientist at IMAX Labs, discusses how his team developed a Large Multi-modal Model (LMM) framework for per-title video encoding optimization

Speakers

Hojat Yeganeh, (IMAX) - Quantitative Assessment of Film Grain Similarity: An Objective Mode

Although digital cinematography has advanced, many artists still favor film grain for its unique texture, but streaming platforms struggle to compress this high-entropy signal. Our studies show that film grain synthesis methods are technically robust, they fail to fully replicate the original grain. To address this, we developed the IMAX Film Grain Similarity (FGS) Metric, which measures film grain fidelity and optimizes each component of the film grain synthesis process.

Major Jiang, (IMAX) - Large Multimodal Model-based (LMM) Video Encoding Optimization

In this talk, we introduce a Large Multi-modal Model (LMM) framework for per-title video encoding optimization, reducing computational complexity while maintaining bitrate savings. The approach offers a more efficient alternative to traditional methods, paving the way for adaptive video encoding strategies.