Why do AI video grading comparisons start with a LOG-looking "before"?
@AI__TSUBAKI4.8K followers
13,918 views
11 Comments
Comments from the original post on X (Twitter)

@artimindArt
Fair point. Generated footage has no sensor data to recover, so a flat 'before' is just staging. Grading from the native output is the honest comparison.
2 likes

@AI__TSUBAKI
Yeah, you're totally right. I actually started this one with a LOG-style look too, since it makes the "after" pop way more lol
2 likes

@artimindArt
Smart, as long as the post says so. An honest before/after builds more trust than a dramatic one.
1 like

@slumdogETH
Its very precise love how you prompt it. Its very natural to the point you wont noticed its an AI video
1 like

@3ihctiAyv087986
こんな演出AIで出来るんですね、リップシンクしながらDanceと教師、別の動き これは知らなかったです、リップシンクしながらやっと編み物出来るようになりましたが
1 like

@portskillahee
lut The concept has been completely stripped of its pedestal, and it no longer has the same level of dynamic material adjustability as it used to. It's just something to fool the public now.
1 like

@kosx_unedited
Lol, you're already able to generate log from Rec.709 footage with Higgsfield. It’s extremely useful. I’ll always choose log over normalised. And if you’re generating and stacking a few videos together, it’s easier to match them in a large color space.
1 like

@AI__TSUBAKI
Where on Higgs can I do that?

@Awesome_O_AI
The graded version looks way cleaner, that LUT really pulled the colors together
1 like

@AI_VideoLab
Fair question — with generated footage there's no sensor data to recover, so a flat 'before' is really just staging. Your LUT still earns its place here: the classroom warm tones and the autumn light outside finally feel like one scene.
1 like

@im_shahid7
It creates a more dramatic and impressive final reveal
1 like