Returning Greatness To Film Grain

Returning Greatness to Film Grain: Introducing AV1-Compatible Film Grain Modeling for Existing
HEVC-Based Video Codecs. By Dan Grois, Alex Giladi, Thomas Guionnet, Thomas Burnichon,
Nikolay Tverdokhleb, and Mickael Raulet

Do the conclusions of this study, which presents an implementation of the existing AV1 film grain synthesis applied to HEVC-based video codecs, justify such a bombastic title? For sure, there are significant gains in efficiency, but what we’re primarily interested in is the ability to preserve the look of the grain, which does not appear to be all that successful: some coarse grain patterns proved to be challenging; the overall film grain appearance differed from the source; and in the case of very coarse grain, the denoiser struggled to fully remove the grain. The grain could be also be seen to “pump”, i.e., appearing and disappearing depending on the frame position within the group of pictures (GOP) and depending on the luminance characteristics.

“Returning Greatness to Film Grain” appeared in the Jan/Feb edition of the SMPTE Motion Imaging Journal several months after the publication of A Subjective Study of Film Grain Synthesis for the Preservation of Creative Intent (Sept. 2023), which found AV1 FGS to be suboptimal.