From R3D to Delivery: My Nikon ZR Post-Production Workflow

My first-impressions piece on the Nikon ZR covered the camera. This one covers what happens after you stop rolling, because a camera that shoots 6K RAW is only half the story if the workflow to actually deliver a finished film is a mess.

R3D NE looks incredible in a grade. It's also a genuinely heavier file to move, store and scrub through than anything I was shooting before. Here's the workflow I've settled on after several months and a few projects' worth of trial and error.

1. Ingest and Backup: Before Anything Else Touches the Footage

Cards don't get formatted and nothing gets graded until footage exists in at least two places outside the camera. For a day's shoot that's usually a copy onto my working drive and a second copy onto a separate backup drive, done as two independent transfers rather than one copy fanned out afterwards: if the first copy fails partway through, I want to know before the card gets wiped, not after.

I check file counts and spot-check a handful of clips on both copies before the card goes back in rotation. It's not exciting work, but a corrupted card halfway through a client edit is a much worse day than the ten minutes this takes.

2. Proxies First: Working with R3D NE on an M1 Pro

I'll say this plainly: my main machine is a MacBook Pro 14" with an M1 Pro chip, not a Mac Studio or a desktop workstation. It handles most things I throw at it well, but native 6K R3D playback at full resolution, especially with multiple grading nodes active, is not one of them.

So proxies aren't optional for me: they're step one. I generate optimized media in DaVinci Resolve straight after backup, at a resolution that keeps the timeline responsive (I don't need to scrub 6K to make an edit decision) while still holding enough detail to judge focus and framing accurately. Cut the whole edit against proxies, then relink to the full R3D NE camera originals only for the final grade and export pass.

It's an extra step that costs render time up front. It's a much smaller cost than fighting a stuttering timeline for the entire edit.

3. Grading: RED IPP2 in DaVinci Resolve

R3D footage is graded in DaVinci Resolve, not Final Cut Pro: Resolve's native support for RED's IPP2 image pipeline and R3D metadata is more direct, and its node-based colour page is built for exactly this kind of work. Log3G10 and REDWideGamutRGB give a genuinely flat, wide starting point, which is the whole point of shooting Log: the 15+ stops of dynamic range only pay off if you don't crush it back down in-camera.

My node structure is deliberately simple: a base RED IPP2 conversion node, a primary correction node (exposure, white balance, contrast), a secondary node for anything that needs isolating (usually skin tones on an interview setup), and a look node last, where any LUT or stylistic grade sits on top of a technically correct image rather than replacing the correction step. Grading on top of a proper base means the same look node can move between shots from different lighting setups without needing to be rebuilt each time.

4. Audio: Handling 32-Bit Float in Post

The ZR's internal 32-bit float audio genuinely simplifies this part of the workflow. Because clipping essentially isn't possible within the recorded range, I don't spend time during the shoot second-guessing gain levels: I set a sensible level and worry about it in post if at all. In the edit, that usually just means a normalization pass and, for interview audio, a light noise reduction and EQ pass in Resolve's Fairlight page before anything gets exported.

5. Finishing: Why Final Cut Pro Still Has a Place

Not everything shot on the ZR goes through the full Resolve grading pipeline above. For interview and podcast-style work recorded in H.265 10-bit or ProRes rather than R3D NE (where the priority is fast turnaround over maximum grading latitude), I'll cut and finish directly in Final Cut Pro. It's quicker for straightforward multicam interview cuts, and there's no meaningful grading benefit to routing footage that was never shot Log through a colour-managed RAW pipeline it doesn't need.

In short: R3D NE, grading-heavy work → DaVinci Resolve start to finish. H.265/ProRes, turnaround-focused work → straight into Final Cut Pro. Matching the tool to what the footage actually needs, rather than forcing every project through one pipeline, is most of what "workflow" means in practice.

6. Delivery and Archiving

Client deliverables export as ProRes or H.264/H.265 depending on where they're headed: a broadcast or agency handoff gets ProRes, anything going straight to social or web gets a compressed H.264/H.265 pass sized for that platform. Once a project is signed off, the graded master and the original camera files both go to long-term storage; the working proxies get deleted, since they're trivial to regenerate from the originals if a project is ever reopened.

Quick Tips: The Short Version

If you only take away six things from this:

• Two copies before you format the card. Two independent transfers, not one copy fanned out: check both before the card goes back into rotation.
• Cut on proxies, grade on originals. Don't fight a stuttering timeline on a laptop that isn't built for native 6K RAW: generate optimized media first, relink for the grade and export only.
• Correct before you stylise. A LUT or look node sits on top of a technically correct base grade, not instead of one: get exposure and white balance right first.
• Don't over-manage 32-bit float audio. Set a sensible level on the day and fix it in post if needed: the format is designed so you don't need to chase levels while shooting.
• Match the tool to the codec, not the other way round. R3D NE and grading-heavy work belongs in DaVinci Resolve; H.265/ProRes turnaround work is often faster straight into Final Cut Pro.
• Keep the master and the originals, delete the proxies. Proxies are trivial to regenerate: don't pay to store them long-term.

Final Thoughts

None of this is complicated once it's a routine rather than a decision made fresh on every project. The camera side of the ZR gets most of the attention (understandably, it's the more exciting half), but the workflow around it is what actually determines whether a shoot day turns into a finished film on time or a week spent fighting a slow timeline.