Ok I understand. So they all settings that will affect rife would be fi_ settings like fi_adaptive, fi_precision or fi_masking? From a developers point of view, which would you choose as a least amount of artifacts workflow?

What rife model and best settings would everyone pick for a fast action 4k marvel movie for transcoding? Im trying 4.22 and 4.18 with mv_grid from 10 to 16. Mv_radius 6, mv_refine 5, mv_wide at 3. I am getting pretty good results, just trying to find the absolute best possible settings.

3

(1 replies, posted in Using SVP)

I keep getting crashes since I update nvidia driver when using rife models. Using Automatic settings at the moment, but i would like to fix using rife ai models. Anyone else have this problem, or any ideas to fix?

Thanks for the response. I went down the rabbit hole and now i end up extracting rpu.bin from original source, then generating .json from rpu.bin. then I use a python script to stretch .json to desired total number of frames. Reinject dolby vision with dovi_tool into raw hevc. Its a pain but it is effective.

Is it actually possible to keep hdr metadata or even better dolbyvision during a transcode? This is the best i can achieve.

Here is my log. and mediainfo

Thanks in advance!