Low Light Lens

Then I sat down at my desk, pulled the clips into DaVinci Resolve, and just stared at this flat, washed-out, kind of grey-green mess on my monitor. That was my wake-up call to the whole color grading thing. Turns out, shooting in D-Log M on the Mini 4 Pro and then doing nothing about it in post is not a workflow. It's just... sad files on a hard drive.

I want to talk about what I've learned over the past few months trying to actually grade my drone footage, because honestly this part of the process completely blindsided me. The flying stuff you can kind of feel out. Color grading is a whole separate skill that I didn't respect at all until now.

Why I Even Shoot in D-Log M

Quick context for anyone who's newer than me: D-Log M is a flat color profile on DJI drones that compresses the highlights and shadows to retain more dynamic range. The idea is you get more information in your file that you can then shape in post. The sky doesn't blow out as fast, dark areas hold more detail. It sounds great in theory.

The tradeoff is that the footage straight out of the drone looks genuinely awful. Like, your parents would watch that and think something's wrong with your camera. You have to do work in post to make it look like anything. For a while I was just slapping a LUT on it and calling it done, which, okay, that's better than nothing, but I was getting inconsistent results and I didn't really understand why.

The LUT Problem

LUTs — lookup tables — are basically color presets that try to interpret your flat log footage and transform it into something that looks cinematic or at least normal. There are free ones everywhere and paid packs that cost anywhere from ten bucks to ridiculous amounts. I bought a pack early on because someone on a Facebook drone group swore by them.

Here's what I didn't get: LUTs are not magic fixes. They're a starting point. And they're designed with specific exposure levels in mind. If your footage is slightly overexposed or your white balance was a bit off when you flew, the LUT is going to fight you. I kept getting these clips where the sky turned this unnatural teal color or the greens on trees looked almost neon. I blamed the LUT. The problem was actually that my exposure was inconsistent clip to clip.

I spent a lot of time thinking I just needed better LUTs. What I actually needed was to be more careful with my histogram while flying. On the Mini 4 Pro you can enable the histogram overlay in the camera settings and I basically never looked at it. Now I won't fly without it. Keeping my exposure consistent in camera made my LUT results about three times more predictable.

Learning to Use Nodes in DaVinci

I resisted DaVinci Resolve for a long time. I was editing in CapCut like a complete gremlin because it was easy. But once I started trying to do real color work, CapCut just couldn't do what I needed. Resolve has this node-based color system that looked terrifying at first.

The way it clicked for me: think of each node as a separate layer of color adjustment that feeds into the next one. I now have a pretty simple setup. First node is my CST — color space transform — which converts the D-Log M footage into a workable color space. Second node is my exposure and contrast corrections, where I balance clips so they match each other. Third node is where I apply my LUT at a reduced opacity, usually around 60-70% so it's not overwhelming. Fourth node is where I do any final tweaks, maybe a slight saturation pull or a vignette.

That's it for most clips. It's not fancy but it's consistent and I understand every step, which means I can fix problems instead of just undoing everything and starting over.

Matching Clips Shot at Different Times

This one is still kinda hard. When I fly a location over a couple of hours, the light changes. A shot I did at golden hour looks completely different from one I grabbed twenty minutes later when a cloud rolled in. Cutting between them without any color work looks jarring, even to people who don't think about this stuff.

What I've started doing is picking one clip as my "hero" clip — usually the one with the best light — and then manually pulling the other clips toward it. In Resolve there's a scopes panel with a waveform monitor and a vectorscope and I've slowly learned to read them instead of just trusting my eyes. Your eyes lie to you, especially if you've been staring at the screen for an hour. The scopes don't lie.

I'm not great at this yet. Sometimes I'll export a video and watch it on my TV instead of my monitor and something will look way too warm or the shadows will be crushed. But I'm getting more consistent.

The Part About Monitors Actually Mattering

Speaking of TVs — I did not have a calibrated monitor when I started this. I was grading on a cheap gaming monitor that made everything look more saturated than it actually was. So I'd think my colors looked great and then send a clip to my cousin and he'd say it looked washed out on his phone. That was frustrating until I figured out what was happening.

I'm not saying you need to drop a thousand dollars on a professional monitor. I'm saying you should at least be aware that your monitor's color output might not be accurate. I got a halfway decent IPS panel and calibrated it with a Spyder device I borrowed from a friend. Big difference in how predictable my results are now.

Where I'm At Now

Honestly my footage still doesn't look like the stuff you see from experienced aerial videographers on YouTube. But it looks like something I made intentionally, which is a huge step up from six months ago. The color grading process went from being this intimidating mystery to being a part of my workflow I actually look forward to.

My next thing to tackle is audio — as in, actually building proper soundscapes under my videos instead of just slapping a royalty free track on and calling it done. The music choices I've been making are embarrassing in hindsight. But that's probably another post.

If you're in the same place I was, shooting D-Log and wondering why your footage looks rough — just know it's a learnable thing. Took me longer than it should have because I kept assuming the problem was my gear. It was my process. Always the process.

I spent last weekend doing a proper planned shoot for the first time. Same Mini 4 Pro I've been flying since October, same location I've been to probably fifteen times. But this time I actually sat down the night before and thought about what I wanted to walk away with. Big difference. Huge, actually.

The "Storyboard" Thing (Sort Of)

I'm not an animator or a filmmaker so when people say storyboard I get a little intimidated. But what I actually did was just open my notes app and write down a list of shots. Like literally: wide reveal of the reservoir from the east side, slow push-in toward the dam, low orbit around the pine cluster near the parking lot, that kind of thing. Maybe twelve shots total. Nothing fancy, just a rough sequence I could visualize.

I also looked at Google Earth the night before and figured out where the sun was going to be during my shoot window. I used an app called PhotoPills for this — I've seen photographers mention it a ton and finally downloaded it. The AR feature where you can hold your phone up and see where the sun will rise or set is genuinely useful. I knew I wanted backlit shots of the treeline and without checking I probably would have shown up on the wrong side and gotten flat front lighting like I did last time.

Honestly the planning part took maybe forty minutes total and it changed the whole shoot. When I landed I had eleven of my twelve planned shots. The one I missed was a low pass under some overhanging branches that I chickened out of, which is fine, that's a confidence thing I'll work on separately.

Smooth Movement Actually Requires Slower Everything

One thing I've been fighting since I started is jerky movement. I watch videos from people like Potato Jet or some of the other creators on YouTube and their footage just glides. Mine looked like I was flying through turbulence even on calm days.

I finally figured out a big part of my problem. I had my expo settings barely touched and my sensitivity was basically at defaults. Once I bumped my expo up on the sticks — I'm running about 0.35 expo on both pitch/roll and yaw now — the inputs felt a lot more gradual. Small stick movements don't immediately translate into big lurchy movements. This sounds like obvious beginner stuff but nobody told me and I didn't think to dig into it until someone in the comments here pointed me toward it.

The other thing that helped was just slowing way down physically. I used to think you needed speed to make aerial footage look cinematic. No. A six-second slow orbit that covers maybe thirty degrees of arc looks way better than me whipping around a subject in three seconds. I set a max speed in the DJI Fly app settings and kept myself capped. Felt painfully slow in the air but on the screen it looked smooth for the first time.

Shooting for the Edit, Not Just Shooting

This is the part that I think was missing most from my first attempt. I went out and just grabbed stuff and hoped editing would save it. Spoiler: it didn't. Editing can't fix bad raw footage, it can only arrange it.

This time I made sure I had a beginning shot, some middle detail shots, and something that could work as an ending. For me that ending was a slow backward reveal from the dam until the whole reservoir came into frame with the hills behind it. I shot it three times to make sure I had a clean take.

I also started recording a few seconds before and after each move. So if I'm doing a push-in, I let the drone sit still for three seconds, do the push, then let it sit still again. That gives me handles in the edit so I'm not scrambling to find a clean cut point. This is such a small thing but it makes the editing process so much less frustrating.

My Edit Workflow Right Now

I'm using DaVinci Resolve for editing, free version. Took me a few weeks to get comfortable with it but I'm glad I pushed through because the color tools are genuinely great even at no cost. My basic flow right now is: import the D-Log M footage, create a new timeline at 4K 30fps, apply a LUT to get a starting point (I bought a small pack from a creator whose work I liked for like twelve dollars), then do my cuts, then go back and do per-clip color adjustments if something looks off.

I'm not doing anything fancy with audio right now. Just finding royalty-free music on Artlist — I have the basic subscription — and cutting to beats loosely. I'd like to get better at music sync but one thing at a time.

The video I made from this shoot is about two and a half minutes long. It's the first thing I've made that I actually feel okay sharing. I put it up on my personal Instagram and got some genuinely nice feedback from people I know, which felt good after weeks of thinking I was terrible at this.

What I'm Working On Next

A few things I want to get better at before winter really locks me out of flying:

The Dronie shot — where you start close to a subject and fly backward while ascending. I've tried this twice and both times the movement felt uneven because I was riding two sticks at once without much coordination. I think this is just practice reps.

Better horizon leveling. I noticed in a couple of my shots the horizon had a slight tilt that I didn't catch in the air. Going to be more deliberate about checking this before I start a move.

Shooting in lower light. I know this is risky on the Mini 4 Pro because the sensor isn't huge but I want to experiment with golden hour and maybe some dusk shots. That's kind of why I named this blog what I named it, so I feel like I have to eventually make it relevant.

If anyone has tips on multi-stick coordination for complex moves I'd genuinely love to hear them in the comments. I've watched tutorials but I think I just need someone to explain the mental approach to it because the physical part isn't clicking yet. Thanks for reading and hopefully the next one I post will have actual footage attached.

I picked up a DJI Mini 4 Pro back in the spring and for the first couple months I was just kind of flying around my neighborhood, checking out local parks, that kind of thing. Nothing serious. But I kept watching these YouTube videos of people doing these gorgeous slow reveals over mountain ridges and smooth push-ins toward buildings and I kept thinking — okay I want to do that. How hard can it be.

Turns out, pretty hard. At least for me.

The Plan That Wasn't Really a Plan

So I decided my first "real" cinematic project would be a short video of this old grain elevator about twenty minutes outside of town. It's got great textures, peeling paint, some rusted machinery around the base, and the surrounding fields are flat which I thought would be good for big wide shots. I'd driven past it a hundred times and always thought it looked cinematic. Good instinct, bad execution on my part.

My biggest mistake was not storyboarding anything. I just showed up with my drone, looked at the structure, and thought I'd figure it out on the fly. Ha. What I ended up with after three battery cycles was about forty minutes of footage that was mostly me hovering in one spot trying to decide what to do next, a bunch of shots where I started moving too fast, and honestly some clips where I just forgot to adjust my exposure before taking off so the sky was completely blown out.

The golden hour light was genuinely beautiful that evening. And I wasted most of it being indecisive.

What I've Learned About Planning a Shot List

After that frustrating edit session — or attempted edit session, I should say — I went back and actually studied some of the videos I'd been watching. Not just watching them for the pretty pictures but actually pausing and thinking about what move the pilot was doing and why it worked.

What I noticed is that good cinematic drone footage usually has a clear beginning, middle, and end to each individual clip. Like the shot starts with some kind of reveal or entry point, moves through space in a deliberate way, and ends somewhere meaningful. It's not just the drone going from point A to point B. There's intention behind it.

So now before I go out I actually write down shot ideas. Nothing fancy — I use the Notes app on my phone. I'll list maybe six to eight shots I want to get and what the movement should be. Something like: slow reverse pull-back starting low at ground level, elevator centered, rise to about 80 feet while pulling back 150 feet. That kind of thing. It forces me to think about the composition before I'm standing in a field squinting at my phone screen.

The Cinematic Mode Situation

One thing I did do right early on was switching to Cine mode on the Mini 4 Pro. If you haven't used it, basically it slows down the stick response so your movements aren't so jerky. It's like flying through honey compared to Normal mode. My initial footage before I found that setting looked like I was trying to swat a fly. Everything was twitchy and overreactive.

Even in Cine mode though I had to learn to actually slow down my inputs. The temptation is to push the stick and hold it and let the drone carry the shot. But what really works better — at least for me — is gentle presses, tiny corrections, letting the momentum do the work. I've been practicing on boring subjects just to train my thumbs to be more patient.

I also started using the ND filters I bought and kind of ignored for two months. ND16 in bright daylight has made a real difference. The motion blur in the footage feels more natural and the image just looks smoother overall. I was running with a 1/1000 shutter before because I was afraid of overexposure and the footage looked weirdly sharp and video-gamey. 180 degree shutter rule has become my religion now.

The Editing Side of Things

I edit in DaVinci Resolve because it's free and because I kept seeing people online say it's what the pros use for color grading. The learning curve is real. I'm still very much figuring it out.

For that grain elevator project I tried to salvage what I could. I ended up with maybe five usable clips out of everything I shot, and I cut them together with some music from Artlist that I subscribed to specifically for this. The final video is like 45 seconds long, which is embarrassing considering I was out there for almost two hours, but those 45 seconds actually look pretty decent. My wife said it looked "really professional" which okay she's biased but I'll take it.

One thing that helped the edit a lot was just being brutal about cutting anything that had even a tiny wobble or drift. If a clip isn't smooth it's gone, doesn't matter how cool the composition was. Smooth is everything in this style of video. People can forgive a lot but shaky aerial footage just looks amateurish and there's no fixing it in post.

Going Back Out This Weekend

I'm planning to go back to the grain elevator this Saturday morning — aiming for the blue hour before sunrise. I've got a shot list this time. Seven specific shots, each one with notes on the movement, approximate altitude, and direction relative to the structure. I've also been messing around in Google Earth to get a sense of where the sun will be rising relative to the building so I can plan for side lighting on the textured walls.

I know the second attempt won't be perfect either. There's a lot I still don't know about this stuff. But I feel like I at least understand now why my first attempt failed and that's something.

If anyone else is at this same beginner stage of trying to figure out cinematic flying, I'd really love to hear what's been working for you. Especially curious if other people are using any specific practice routines to improve their stick control, because that's still my biggest weak point. Drop something in the comments if you've got tips. I'll take all the help I can get.