Matching Nikon Z footage with Sony and Canon
How do you match Nikon Z footage with Sony and Canon on one job? Normalize each source first, pick a reference, then balance exposure and white balance.
A Z8 cutting against an FX3 and a C70 is a fixed sequence of decisions, and the order matters more than any single setting. Normalize each camera with its own correct transform. Pick one as the reference. Balance exposure and white balance toward it. Build the look once and push it to all three. Matching before normalizing turns this into a long evening, because every adjustment is fighting a conversion you have not applied yet.
Nikon adds three specifics. N-Log only exists in 10-bit recording and above, its gamut container is Rec.2020 rather than a Nikon-specific space, and Nikon publishes both a conversion LUT and a complete DaVinci Resolve recipe you can follow exactly.
What the three makers are actually doing differently
A log format is two manufacturer decisions: a transfer curve, and a gamut container to carry the colour.
Nikon's container is the surprising one. Nikon's video editing guide sets the Resolve Color Space Transform to Input Color Space Rec.2020 with Input Gamma Nikon N-log, and the technical file in Nikon's newer LUT package is named N-Log_BT2020_to_REC709_BT1886_size_33.cube. Two separate Nikon documents say the same thing: N-Log rides in a BT.2020 container. Set it to Rec.709 and the picture looks plausible while being wrong in the saturated colours, which is the most common Nikon mistake in a mixed timeline.
Sony makes you choose at the camera. It describes S-Gamut3.Cine as a setting that "allows you to shoot in a color space that can easily be converted for digital cinema" and S-Gamut3 as one that "allows you to shoot in a wide color space", both paired with S-Log3. An S-Log3 clip is therefore not self-describing: the two gamuts want different input transforms, so you need to know which one the operator picked.
Canon offers three. On the R5 the colour space under Canon Log 3 can be BT.709, BT.2020 or Cinema Gamut, and Canon's white paper recommends Cinema Gamut with a Neutral matrix for ACES work. Canon places the curve carefully too, saying Canon Log 3 "is not the successor to Canon Log 2 but the successor to Canon Log", extending dynamic range by one stop while keeping Canon Log's 18 percent grey behaviour. How to grade C-Log3 footage covers that side.
Treat published dynamic range figures with suspicion. Nikon's Z8 and Z9 manuals give N-Log as 12 stops (1300%) with H.265 10-bit at ISO 800, while the Z6III manual gives 13+ stops at the same 1300%. Canon quotes approximately 1600% for Canon Log 3. These are not a common scale. The encodings can be put on one. Computed from each maker's own published formula, N-Log's curve reaches full scale 6.36 stops above middle grey, Canon Log 3 at 6.35 and S-Log3 at 7.74, which is a property of the equation rather than of any particular sensor.
Which camera should be the reference
Pick the camera the audience looks at longest, usually whichever body covered the faces or has the most shots, and move everything else toward it. One rule overrides that: if any camera shot a baked picture profile rather than log, match toward it, because display referred material has less latitude and dragging it far is where highlights go plastic. That holds regardless of brand, and frequently makes the cheapest camera on the job the reference.
Nikon adds a wrinkle worth knowing first. Across the Z6III, Z8, Z9 and Zf manuals, Nikon sets the lowest available maximum sensitivity to ISO 1600 while N-Log is active and blocks ISO Hi 0.3 through Hi 2.0. Where a body offers Lo settings, Nikon is blunt about the cost on the Z8: "The maximum output level for videos recorded at low ISO sensitivities (Lo 0.3 to Lo 2.0) drops due to decrease in highlight data." A Nikon shot at Lo has less highlight headroom than the same body at ISO 800, which limits how far you can push it toward a brighter reference.
Normalize before you balance anything
Normalize means one thing: applying the log conversion so the clip sits in a known space. Correcting exposure and white balance is a separate, later step, and building a look is a third. Out of order, every colour judgement made against an unnormalized clip has to be made again.
Nikon publishes its route in full, against DaVinci Resolve 18.6 and current as of April 2024. In Project Settings under Color Management: colour science DaVinci YRGB, timeline colour space DaVinci WG/Intermediate, 3D lookup table interpolation Tetrahedral, output colour space Rec.709 (Scene) on Windows or Rec.709-A on macOS. On the Color page, the first node carries a Color Space Transform with Input Color Space Rec.2020, Input Gamma Nikon N-log, Output Color Space DaVinci Wide Gamut, Output Gamma DaVinci Intermediate. The last node runs the reverse, ending at Rec.709, and grading happens in the middle. Give the Sony and Canon clips their own equivalent first nodes and all three arrive in the same working space. Arrive is the right word: with each format decoded by its own published inverse, the three converge exactly rather than approximately, so anything still separating them afterwards is the sensor rather than the transform.
Where you correct exposure depends on the method. With a conversion LUT it happens upstream while the clip is still log, because once values pass through the table the skew is baked in. With a colour space transform you correct in the working space after the input transform. Either way it comes before the creative look. Why CST, ACES and LUTs look different covers why a CST on the Nikon and a vendor LUT on the Canon will not meet in the middle.
The 10-bit floor, and why the Zf is the odd body out
N-Log is not available in 8-bit at all. On the Z6III, Z8 and Z9, Nikon lists four qualifying file types: N-RAW 12-bit (NEV), ProRes RAW HQ 12-bit (MOV), ProRes 422 HQ 10-bit (MOV) and H.265 10-bit (MOV). Choose anything else for Video file type and N-Log is not selectable as a tone mode.
The Zf is narrower. Its manual lists only H.265 10-bit (MOV) for N-Log. Pair a Z8 with a Zf and the Zf is the constraint: the Z8 can deliver ProRes 422 HQ while the Zf gives you long GOP H.265, and the two will not behave identically under a heavy shadow lift even after identical normalization. If the Sony or Canon recorded 8-bit, that camera belongs at the reference end rather than being dragged around.
Nikon's own LUTs, and which of the two packages you want
Nikon publishes two free conversion packages from its Download Center, and they are easy to confuse.
The older N-Log 3D LUT, version 1.06, is per body, with separate .cube files for each camera including Z_9_N-Log-Full_to_REC709-Full_33_V02-00.cube and Z_f_N-Log-Full_to_REC709-Full_33_V01-00.cube, alongside files for the Z 8, Z 7II, Z 6II, D780, Z 7 and Z 6. The naming tells you they run from N-Log full range to Rec.709 full range.
The newer N-Log LUT, version 2.00, ships one technical file, N-Log_BT2020_to_REC709_BT1886_size_33.cube, plus four creative looks carrying RED names: Achromic, Film Bias, Film Bias Bleach Bypass and Film Bias Offset. The technical file targets BT.1886 rather than a vaguely defined Rec.709. The four RED files are looks, not normalization, and putting one on the Nikon while the Sony gets a plain conversion guarantees a mismatch.
Nikon states the 3D LUT also works in Adobe Premiere Pro and Apple Final Cut Pro X, with the caveat that it neither endorses that software nor guarantees its operation. To load it into Resolve, use Open LUT Folder under Project Settings and Color Management, drop the file into a sub-folder, then click Update Lists and Save. The single camera version is in how to grade N-Log footage.
The settings that quietly wreck a Nikon match
Data levels first, because it produces footage that looks nearly right and refuses to sit. If the Nikon went to an external HDMI recorder, Nikon instructs you to check that Data Levels in Resolve matches the camera's HDMI Output range: right-click the clip in the Media Pool, open Clip Attributes, and set Data Levels to Video for Limited range or Full for Full range. Get it wrong and blacks and whites are offset before the first transform runs, which reads as a colour problem and is not one.
View assist is second. It is Custom Setting g13 on the Z6III, g14 on the Z8 and g10 on the Z9, and all three manuals say it enhances the preview only and leaves the recording unaffected. Canon says the same about its own View Assist. Judging a match through one is judging a picture nobody else will ever see.
Third is doubled transforms in Premiere, where colour management may already have detected the log clip before you add your own LUT: Premiere is applying a LUT to my log footage.
How much of the difference is genuinely fixable
Tone, contrast and overall balance come together well, and skin can be brought close. Strongly saturated colour is where two sensors with different spectral responses stop agreeing, and no global control reconciles that. That is what a qualifier is for, applied last and only on shots the audience will linger on. The wider argument is in why your cameras don't match and why Resolve shot match is not working.
Leumos AI is our browser-based grading studio, and the relevant part is the input side: uploads get an ACES based input transform for their format, N-Log and the Sony and Canon curves among them, so each camera is normalized the same way before any look goes on. It grades toward a reference image rather than toward another clip, which keeps a three body job from turning into pairwise matching. Closed beta, waitlist at leumos.ai.
If other formats are on the job, see how to grade V-Log and V-Log L footage and Apple Log workflow on Windows. If the clip count rather than the camera count is the problem, batch grading and exporting many clips covers the scaling side.
Frequently asked questions
Can I record N-Log on a Nikon Z body in 8-bit?
No. The qualifying file types are N-RAW 12-bit (NEV), ProRes RAW HQ 12-bit (MOV), ProRes 422 HQ 10-bit (MOV) and H.265 10-bit (MOV) on the Z6III, Z8 and Z9, and H.265 10-bit (MOV) alone on the Zf. With anything else selected, N-Log is not available as a tone mode.
Should the Nikon or the Sony be the reference camera?
Whichever has more shots, better exposure, or closer coverage of faces. One exception overrides brand: if a camera shot a baked Rec.709 profile rather than log, make that the reference, because it has the least latitude.
Which Nikon LUT should I use on a mixed camera job?
The technical one. In version 2.00 that is N-Log_BT2020_to_REC709_BT1886_size_33.cube; in version 1.06 it is the file named for your body. The four RED named files are creative looks, and applying one while other cameras get a plain conversion creates a mismatch rather than fixing one.
Why does my N-Log footage look wrong only in the saturated colours?
Almost always the input colour space. N-Log's container is Rec.2020, and Nikon's own Resolve recipe sets it explicitly. The right log curve with the wrong gamut looks broadly correct and drifts exactly where colour is strongest.
Sources
All fetched 2026-08-15.
- Nikon, N-Log Video Technical Guide, Video Editing Edition.
- Nikon N-Log recording manuals: Z6III, Z8, Z9, Zf.
- Nikon Download Center: N-Log 3D LUT v1.06, N-Log LUT v2.00.
- Sony, ILME-FX3 Help Guide: Picture Profile.
- Canon, Canon Log gamma curves white paper, 2018 and EOS R5 manual: Canon Log settings.
- Sony Corporation, "Technical Summary for S-Gamut3.Cine/S-Log3 and S-Gamut3/S-Log3" (PDF), which carries Sony's IRE and code value table. Distributed by Sony Professional; pro.sony blocks direct hotlinks, so search the exact title to retrieve it. No Sony reference levels are quoted above.