DJI D-Log color correction: fix flat drone footage the right way

You flew the perfect line at golden hour, landed, and the footage looks like dishwater. Nothing went wrong. Here is why D-Log and D-Log M look like that, the correction order that fixes them, and the drone-specific traps that ruin more aerial grades than any LUT choice.

Why D-Log looks flat

D-Log and D-Log M are logarithmic encodings: they compress the sensor’s full dynamic range into the file so a bright sky and a shadowed valley can survive in the same shot. Your display expects Rec.709, a much smaller container, so until you apply a color space transform or conversion LUT the image renders low-contrast and desaturated. The detail is all there; it has not been mapped for your screen yet.

One DJI-specific wrinkle: D-Log and D-Log M are different curves, not one profile with two names. The larger drones and cameras shoot full D-Log; the compact drones shoot the milder D-Log M. Each needs its own conversion. Put a D-Log LUT on D-Log M footage and you get crunched contrast and skewed color that no amount of tweaking fully undoes.

The metadata trap

Some DJI files label themselves Rec.709 in their metadata even when the image is log. Editors and tools that trust the tag will treat one clip as already-converted and its twin as log, and suddenly half your timeline grades differently from the other half for no visible reason. If two clips from the same flight refuse to match, check what the software thinks each clip is before you touch a single wheel. Judge the image, not the label.

The correction order that works

1. White balance first. Neutralize while the footage is still log, using a gray reference or the scopes. Auto white balance drifts between takes as the drone yaws from sunlit ground to open sky, and every later step inherits whatever cast you leave here.

2. Transform with the right LUT. Apply the conversion that matches the actual profile: D-Log or D-Log M, not whichever LUT came with a preset pack. A conversion on top of a color cast bakes the cast in; that is why white balance comes first.

3. Set exposure with scopes. Aerials are dominated by sky, so your eye reads them brighter than they are. Trust the waveform: protect the highlights, lift the ground detail, and stop before the haze starts to glow.

4. Match the drone to the ground cameras last. Most drone footage lives in a timeline next to a Sony, Canon, or iPhone. Match the corrected clips to each other before any creative look goes on. A look applied to mismatched sources just gives you mismatched clips wearing the same look.

The three mistakes that ruin most drone grades

Grading the sky by eye. Skies fool adapted eyes within seconds. By clip ten your blues have drifted teal and the sequence flickers. Scopes do not adapt.

One LUT for two drones. A shoot with a full-size drone and a compact one is a two-profile shoot. Detect and convert per clip, or one drone always looks slightly wrong.

Skipping the ground match. The aerial can be gorgeous in isolation and still yank the viewer out of the edit if its greens and skin tones disagree with the shot before it.

The boring half is the automatable half

Notice that everything above is measurement, not taste: detect the profile, neutralize, transform, hit exposure targets, match sources. That is exactly the work Basecut automates. It detects the camera and log profile per clip (DJI D-Log among 16 profiles), reads the image instead of trusting mislabeled metadata, runs a closed-loop white balance and exposure correction with real scopes, matches color across cameras, and hands per-clip grades to DaVinci Resolve, where the creative work starts. Locally on your machine, nothing uploads.

Culling and color are free on your own footage, so you can judge the correction on your own flights before spending anything. See it on real footage.