What these two bodies have in common is not a sensor. It is a transfer curve and a gamut, and that is the part that decides your grade.
One curve, one matrix
Both record S-Log3 in S-Gamut3.Cine. In the transform library this site runs on, that is a single entry: one decode curve and one 3×3 matrix into Rec.709, with 18% grey landing at 0.4106 normalised — which matches Sony's published 420/1023.
So there is no per-body Sony LUT to find, because there is no per-body Sony curve. One Color Space Transform setting, or one S-Log3 → Rec.709 conversion, is correct for every body in the rig recording S-Log3 / S-Gamut3.Cine — which is why the same conversion normalises an FX5 A-cam and an FX3 B-cam onto one timeline. If two Sony angles do not match after it, the conversion is not the thing that differs between them.
Where they actually differ
| FX3 | FX5 | |
|---|---|---|
| Sensor | 12.1MP BSI CMOS | 16.6MP fully stacked CMOS |
| Base ISOs (as published) | Dual base: 800 / 12,800 | Triple base: 800 / 4,000 / 12,800 |
| Capture | 16:9 4K UHD | 5K open gate, full frame |
| Internal RAW | No — external over HDMI only | 16-bit X-OCN, internal |
The FX5 is the newer body, announced July 2026 and positioned between the FX3 and the FX6.
Those rows are manufacturer-published figures, not anything measured here — check the base-ISO row against your own camera's menu before you plan around it. The table stops at two columns on purpose. These are the bodies this site documents in detail; a spec column filled in from memory is how a page starts being wrong. The one-curve argument above is not a claim about a model number, though — it is a claim about a format. Any body recording S-Log3 in S-Gamut3.Cine takes the same conversion.
What that means on a two-camera night
The base-ISO ladder is where the published specs differ, not the colour. The FX5 is listed with a third base at 4,000, between the 800 and the 12,800 the FX3 also carries; on paper the FX3 has the two ends of that ladder and not the middle, which is the whole of the low-light argument between the bodies. It changes nothing about the grade: the S-Log3 curve carries no ISO term, so 18% grey encodes to the same place whatever the camera is set to.
Treat every Sony angle as one family in the grade. The curve and the gamut are the same across bodies, so normalise every angle with the same conversion, match exposure between them, then put one look on footage that is already in the same space.
Expose by the numbers, because the monitor lies in a dark room. Middle grey sits at 41% IRE in S-Log3, and skin lands in the 55–62% band once you are exposing to the right by about a stop — which you want anyway, to keep noise out of the shadows. Zebra 1 at 94% catches laser spikes before the sensor clips.
Next
- How to expose S-Log3 on the FX5 — the full method.
- Grading FX5 X-OCN RAW in Resolve — what changes when the file is RAW.
- FX3 vs FX5 for low-light event coverage — picking the A-cam.
- Sony S-Log3 conversions — what the transform is actually doing.
Sources
The specification table, and the fact that one Sony log preset normalises both bodies' S-Gamut3.Cine response onto a single timeline, come from this site's own FX3 vs FX5 comparison guide. The curve, the matrix and the 0.4106 figure come from the transform library this site ships. That the S-Log3 curve carries no ISO term comes from the site's FX5 S-Log3 exposure guide. The exposure numbers come from the site's concert page and homepage tips, and the FX5 announcement date from the site's own engineering notes.
---