SUPER 16 as Practice: The Camera and Colour Pipeline Behind Agents of Control
The Bodies: Cousins, Not Twins
The C70 and the Fuji X-H2S aren't an obvious pairing. But they share the things that matter: Super 35 sensors, modest megapixel counts, and a bias toward documentary application over spectacle. Two years of testing showed me CLOG2 and FLOG2 cut together like cousins: both warm, both carrying a slight retro cast and flattering to skin tones.
Blackmagic was the other option on the table, but I passed. I needed hybrid bodies that could shoot stills as well as motion, and the C70 needed to double as a proper client camera on other jobs. That ruled out a body built for one purpose.
Colour Science: Picking a LOG
FLOG2C on the Fuji looked beautiful in isolation. However, it's practically its own colour world, closer to the GFX100 II, which I also tested, than to anything Canon. So I let it go, along with the idea of shooting RAW on the Fujis at all: Fuji's RAW colour science sits too far from Canon's to resolve cleanly.
The compromise was Canon RAW LT on the C70: the right balance of latitude, flexibility, and file size. I then chose ProRes HQ 422 at the highest bitrate on the X-H2S, and the compressed codec at highest bitrate on the X-S20. Everything at 4K DCI, 24fps. I bought Cinematch as insurance, and it earned its keep: some X-H2S footage got shot in FLOG2C by mistake.
On Set: White Balance and Exposure
White balance stayed locked at 5600K throughout. A fixed reference meant the two camera systems stayed consistent against each other, and it gave the grade a stable starting point.
Exposure was the bigger discipline. The Fuji bodies don't offer much in the way of proper monitoring with small screens, and no reliable false colour worth trusting in the field. The histogram became the actual tool. Reading it on set rather than eyeballing the LCD meant exposure was correct going into the edit. That mattered more in post than I expected: clean, correctly-exposed log footage meant the problems I found in the grade were actual colour and texture decisions, not damage control.
Glass: Selling the Sharp Stuff
Most of my Fuji glass went. It was too sharp, too photographic. The film leans on zooms at 2.8: the Canon 28-70 and Sigma 70-200, run through a speedbooster on the C70 and Fringer adapters on the Fujis. The Fuji 23mm f/2, the Canon 50mm STM, and the Sirui Nightwalker 35mm 1.2 covered the rest.
Most lenses carried an ND and a diffusion filter: 1/4 or 1/8 black mist from Neewer. I'd tried the TIDE Cinesoft filters first and found them too heavy, blooming further than the shots needed. Tiffen was the other option, priced for a budget this film didn't have. Neewer got most of the way there for half the cost. There's something in old Canon EF glass that meets Fuji halfway under diffusion: a soft vintage character that reads as documentary rather than romantic.
The Super 16 Problem
Matching true Super 16 digitally is its own issue. Full-frame sensors with shallow depth of field give you the wrong film entirely, especially when shooting anamorphic: too much creamy separation, not enough grit. The Super 35 crop helps. Super 16 is naturally a little flat without losing dynamism. I owned a first-generation Blackmagic Pocket Cinema Camera, and it remains the closest thing I've shot to a true Super 16 look: appalling in low light, correct in everything else.
The LUT search that followed was a small tour of the market. Triune Films' packs were well-built but never quite landed the film look. FilmConvert Nitrate and Dehancer both fell short on the same axis: technically competent, but bit exaggerated. Cineprint emulations got the colour right but leaned on Resolve's native grain plugin, which is too clean.
Filmbox Looks solved it. The Indie Scan simulation, part of the 16mm look family, paired an accurate Super 16 LUT with grain texture that actually breathed organically. At roughly £180 it was the one purchase in this whole chain that outperformed its price.
The Grade, Node by Node
Filmbox Looks: source set to Canon Log 2 / Cinema Gamut or Fuji Log 2, Looks Family 16mm, Look: Indie Scan. Contrast 0.4, Vibrance -0.1, Halation 0.0, Grain -0.4, Dust 0.0. The plugin's own grain is generous by default; pulling it back kept it from fighting the diffusion already baked into the footage.
Detail recovery: because the diffusion filters soften detail before the grade even starts, I rebuilt some of it back in Resolve: a white-recovery node and a mid-detail node (Mid Details pushed to 54), merged together downstream of the Filmbox node.
Log wheels: Pivot 0.435, Contrast 1.0, range 0.333–0.550. Highlights pulled to -0.11 across all three channels, Offset lifted to 19.40 across all three. The point wasn't crushing anything: just easing the highlight roll-off so white faces under diffusion didn't wash out, while the offset lift kept the blacks from reading too digital.
Red node: a custom red LUT on its own node, output gain at 0.4, pushing the whole image toward a faint conspiratorial red. Small move, but subtle effect.
I checked the result against a high-resolution pass of Moonrise Kingdom for tonal texture: not to copy it, but to have a fixed point of reference for what "restrained retro" actually looks like next to what I'd built.
What Actually Worked
Filmbox Pro's LUT export saved the Super 16 emulation as a standalone LUT, but stripped out grain and vibrance in the process: useless for my purposes, since the texture was the entire point. What did work was Resolve's PowerGrade save. Stacking the plugin and the PowerGrade together was the only combination that held onto the full emulation, grain and all.
None of this exists without the location and the costume doing their own share of the work: the Royal Berkshire Archives gave the film a texture no LUT can fabricate.