Post-processingSingle documentFrom the Lightroom Classic editing guide
Lightroom Classic 15.x · April 2026
Editing Fujifilm GFX 100S II
RAW Files
High-Bit-Depth RAF
Order-Dependent Workflow
Six Stream Profiles
Per-Lens Defaults
Pixel Shift Pipeline
Companion to the Fujifilm GFX 100S II guide hub and Settings Overview.
Companion navigation
How to use this document
Use this during edit sessions after capture to maintain stream-consistent rendering decisions.
Show companion links
00
RAW → Lightroom Baseline
The camera guide establishes the capture philosophy: the camera's role is to deliver the richest possible RAW file and a preview accurate enough for field decisions. Lightroom Classic's role begins where the camera's ends. This is the second half of a single continuous workflow — not a standalone Lightroom guide.
The workflow has two sides. The camera avoids settings that damage the RAW file; Lightroom handles profile, white balance, exposure, tone, detail, and noise in a stable order. If either side is skipped, later corrections become harder to judge.
Camera must NOT
Apply sharpening to RAW data · Compress DR via D-Range Priority · Impose Contrast via Tone Curve on the sensor output · Apply in-camera NR to the RAW file · Process JPEG parameters into the RAW data.
Camera must
Deliver the highest-quality RAF mode appropriate to the assignment · Record with a neutral preview simulation · Capture at native ISO whenever possible · Apply GF lens correction profiles · Provide an accurate live histogram for ETTR discipline.
LRC must NOT
Apply Contrast before recovering Highlights/Shadows · Apply NR before Clarity is set · Apply Sharpening before NR · Change Camera Profile after tonal work · Apply unnecessary Luminance NR at base ISO.
LRC must
Set Profile and Lens Corrections first · Set WB before any tonal decision · Establish Exposure before Highlights/Shadows · Recover Highlights before lifting Shadows · Evaluate NR at final brightness and Clarity settings.
The central insight: The GFX 100S II's broad dynamic range and high-bit-depth RAW options exist to give Lightroom headroom. Every setting in the camera guide is designed to maximise that headroom. Every setting in this guide is designed to spend it wisely — not waste it through wrong-order processing.
01
What Makes GFX RAW Files Different
The GFX 100S II produces Bayer-pattern RAW files — not X-Trans. LRC's demosaicing is clean and free from the maze artefacts that require careful handling on older X-Trans sensors. The challenges here are different: resolution, dynamic range headroom, and the precision demands that high-bit-depth capture at 102 MP imposes on every editing decision.
Bayer CFA
Standard colour filter array. Default Detail 25 is correct — you do not need to lower the Detail slider as required with X-Trans files.
~14 Stops DR
3–4 EV of blown-looking highlights often recoverable. The histogram's clip warning fires 1–2 stops before the RAW data is exhausted. Never discard a frame for highlights without testing Highlights −100 first.
102 MP · High Bit Depth
Sharpening demands are modest — the sensor does the work. Aggressive sharpening creates unnatural micro-texture visible at print scale. Masking is essential on every file.
Low Noise Floor
At base ISO, Luminance NR should start at zero. Heavy NR values carried over from smaller-sensor workflows destroy micro-detail on GFX files. Start at 0, add only what you can see at 1:1.
AI Denoise in current Lightroom Classic — processed as a new DNG
In current Lightroom Classic, AI Denoise is available from the Detail panel but is processed as a new enhanced DNG file. Adjust the amount in the Denoise flow, run Enhance, and continue editing from the generated DNG. Treat it as a derivative RAW master rather than a live toggle attached to the original capture.
AI Denoise vs Luminance NR: AI Denoise operates at the raw data level before demosaicing and often preserves more micro-detail than slider-based Luminance NR at equivalent noise reduction. For ISO 1600+ GFX files, evaluate AI Denoise case by case. At base ISO, Luminance NR 0–10 is typically sufficient and AI Denoise often adds processing time for marginal gain.
02
The Rendering Pipeline
LRC processes adjustments in a fixed internal rendering order independent of the UI panel layout. The sequence below reflects both the engine's computation order and the optimal working order for GFX files. Each step operates on the output of the previous — order is not preference, it is physics.
Camera Profile & Lens CorrectionFoundation
First in the engine. Sets colour rendering baseline and corrects geometry before any tonal work. Changing profile after tonal work shifts colour and invalidates all downstream decisions.
White BalanceFoundation
Temperature and tint define the colour cast. Must be set before tonal judgements — warm casts make images appear brighter; cool casts darker. Wrong WB makes every Exposure decision unreliable.
ExposureTonal
Master luminance scale. Every downstream slider operates within the window Exposure defines. Target the midtones. The single most consequential adjustment in the workflow.
Highlights & WhitesTonal
Recover overexposed regions using raw data beyond the visible histogram clip. Do before Shadows. GFX files recover 3–4 EV of blown-looking highlights routinely.
Shadows & BlacksTonal
Lift shadow detail after highlights are resolved. Lifting shadows exposes the noise floor — this informs NR decisions. Blacks anchors deep shadow. Do not lift Shadows before Highlights are set.
Contrast & Tone CurveTonal
Shape tonal relationships after the full range is established. Contrast before Shadow/Highlight recovery crushes data before you can access it. Always last among tonal sliders.
Texture & ClarityLocal Contrast
Applied to the tonally-corrected image. Clarity amplifies noise — NR must come after Clarity, never before. Never assess noise before Clarity is finalised.
AI Denoise · Noise ReductionDetail
Assess noise at final brightness and Clarity settings. Shadow lifting and Clarity both amplify noise — NR must reflect the actual state after all tonal work. NR before tonal work means under- or over-correction.
SharpeningDetail
Capture sharpening after NR. NR reduces fine detail that Sharpening then recovers. Always Alt-drag Masking before setting Amount. Over-sharpening at 102 MP is the most visible processing error at print scale.
HSL · Colour Grading · CalibrationColour
Colour relationships shaped last, after luminosity is fully resolved. HSL hue shifts affect perceived brightness — colour work before Exposure creates a moving target.
Vignette · Grain · EffectsEffects
Applied last as compositional decisions on the finished image. Post-crop vignette modifies final pixel values — setting it during tonal work creates false feedback.
03
Why Order Matters: The Interdependency Map
Every major LRC adjustment has downstream consequences. At 102 MP with this much tonal headroom, those consequences are amplified — the latitude to correct mistakes is large, but the mistakes are also more visible at print scale.
Exposure ↑
clipsHighlights may need recovery — higher Exposure means more Highlights must be pulled. These are always paired operations.
revealsShadow noise floor — brightens shadows, exposes noise. Calibrate NR only after final Exposure.
Shadows ↑
amplifiesLuminance noise — primary cause of visible noise in GFX files at higher ISO. Directly governs NR setting.
requiresBlacks re-anchored — after Shadows +40–60, pull Blacks −15–25 to restore depth. These always travel together.
Contrast ↑ early
crushesShadow data before recovery — the most common workflow error. Compresses the tonal headroom recovery sliders need to access.
clipsHighlights before recovery window — Contrast boosts highlights simultaneously. Always set after both Highlights and Shadows are resolved.
Clarity ↑
amplifiesNoise visibility — increases midtone contrast, making luminance noise more apparent. Always finalise Clarity before setting NR.
reducesNeed for high Sharpening Amount — strong Clarity adds perceived sharpness. Do not compound with aggressive Amount after high Clarity.
Luminance NR ↑
destroysMicro-detail — the GFX's core asset — excessive NR eliminates fine detail that justifies 102 MP. Use minimum effective amount.
requiresSharpening Amount to compensate — NR smooths the image; Sharpening must then recover acuity. Adjust Sharpening after NR.
Most dangerous order: Contrast → Shadows → Highlights. Applying Contrast first compresses both ends. Recovery sliders then fight against compressed tonal data with reduced headroom. The result is muddy shadows or gradation-less highlights.
Correct order: Exposure → Highlights → Shadows → Blacks → Contrast/Curve → Clarity → NR → Sharpening. Each step operates on the cleanest possible version of the tonal range.
04
Foundation: Profile, Lens, White Balance
These operate at the raw data level before any tonal processing. Changing any of them after tonal work invalidates all downstream decisions and requires re-evaluation of every subsequent slider.
Camera ProfileCamera Calibration panel → Profile
Default starting profile
Adobe Color (general) · Camera profiles per stream
| Profile | Characteristic | Best For |
|---|---|---|
| Adobe Color | Default, balanced | Versatile starting point, most subjects |
| Adobe Standard | Neutral baseline | Maximum headroom, colour grading work |
| Camera Classic Chrome | Muted, filmic | Landscape, documentary, street (colour) |
| Camera Astia/Soft | Soft, accurate skin | Portrait, controlled light |
| Camera Classic Neg. | Low contrast, warm | Street, film-look starting point |
| Camera Provia/Standard | Accurate, neutral | Architecture — material colour fidelity required |
| Camera Velvia/Vivid | High saturation | Use cautiously — compresses highlight headroom |
Lens CorrectionLens Corrections panel → Profile tab
Enable Profile Corrections
ON
Remove Chromatic Aberration
ON
Build both into your Import Preset. All GF lenses have embedded profiles that LRC auto-detects. Without corrections active, Exposure, Shadows, and Vignette sliders compensate for optical characteristics that should not be there.
Vignetting correction — GF 80mm F1.7
75 (preserve natural optical vignette)
The 80mm's natural vignette at f/1.7–f/2 contributes to subject isolation and the characteristic optical rendering. Setting 75 rather than 100 preserves that quality as a compositional asset. Use 100 on the 20–35mm, 45–100mm, and GF 100-200mm F5.6.
Per-Lens LRC Import Defaults
| Lens | Profile | Sharp. Amount | Masking | Lum. NR (base ISO) | Vignette |
|---|---|---|---|---|---|
| GF 20–35mm F4 | Classic Chrome | 30–40 | 20–35 | 0 | 100 |
| GF 45–100mm F4 | Adobe Color | 30–40 | 25–40 | 0 | 100 |
| GF 80mm F1.7 | Astia/Soft | 25–35 | 55–70 | 0 | 75 |
| GF 100-200mm F5.6 | Adobe Color | 30–40 | 30–45 | 0 | 100 |
GF 80mm Masking at import: Alt-drag Masking to 55–70. The overlay should show sharpening applied only to clearly defined edges — iris, lashes, lips, structural lines. Smooth skin and out-of-focus backgrounds must be black in the mask. Set Masking before adjusting Amount; the Amount value is only meaningful in the context of where the mask places it.
White BalanceBasic panel → WB · Temperature · Tint
Starting point
As Shot → correct from there
If the camera was set to a fixed Kelvin (as recommended in the camera guide for landscape and studio work), As Shot gives you that deliberate field value. Correct with the WB selector on a known neutral, or set Temperature manually.
Tint at GFX resolution
Check every file carefully
Tint matters more at 102 MP. Even ±3 on Tint is perceptible at print scale on skin tones and neutral surfaces. This is not a set-and-forget slider at this resolution — check it on every portrait and every image with a visible neutral surface.
05
Exposure & Tonal Range
The Exposure slider performs a linear luminance scaling of the entire tonal range. It is the engine's first tonal operation. Every subsequent tonal slider operates within the window it defines. Target midtone brightness before touching any other tonal control.
ExposureBasic panel · −5.00 to +5.00 EV
ETTR files on import
Expect near-clip — do not reduce Exposure
ETTR files arrive in LRC with a near-clipped histogram. This is correct. Your first action is to pull Highlights down — not to reduce Exposure. Reducing Exposure to bring highlights under control wastes the shadow quality ETTR captured. The sensor filled the tonal bucket; your job is to shape it, not compress it.
Target midtone placement
Dominant grey ≈ 50% luminance
LRC Histogram + J key: Press J to toggle the clipping overlay. When highlights clip, pull Highlights to −40 before deciding data is genuinely lost. At −40, most GFX landscape files reveal complete cloud and sky structure that appeared blown in the preview.
Highlights & WhitesRecover before lifting shadows
GFX starting points
−40 to −80 depending on scene
| Scene | Highlights | Notes |
|---|---|---|
| Outdoor bright sky | −50 to −80 | GFX recovers full cloud structure; check Alt-drag |
| Studio / controlled | −20 to −40 | Watch specular highlights separately |
| Interior window light | −60 to −100 | Use Tone Curve Lights node for upper midtone |
| Overcast / flat | 0 to −20 | Focus on contrast shape, not recovery |
| Portrait, direct sun | −40 to −70 | Highlights slider pulls upper midtones — watch skin |
Whites — set with Alt-drag
First specular clip point
Alt/Option-drag Whites shows only clipping areas on a black canvas. Pull up until first clipping appears in specular regions, then back off slightly. If you find yourself at Whites ±60, you've confused it with Highlights — they address different tonal zones.
Shadows & Blacks — always travel togetherSet after Highlights are resolved
Core principle
Shadows ↑ · Blacks ↓
Lifting Shadows without re-anchoring Blacks produces flat, washed-out images. Shadow lift removes depth from the lower zone; Blacks restores the true black point. Rule of thumb: if Shadows is +50, consider Blacks −15 to −25. Adjust the ratio by subject — portraits need less Blacks compression than landscape to preserve skin shadow gradation.
| ISO | Shadows | Blacks offset | Luminance NR |
|---|---|---|---|
| ISO 100–200 | +40 to +60 | −15 | 0 |
| ISO 400–800 | +30 to +50 | −10 | 10–20 |
| ISO 1600–3200 | +20 to +40 | −10 | AI Denoise → 20–30 |
| ISO 6400+ | +15 to +25 | −5 | AI Denoise → 30–50 |
06
Contrast, Tone Curve & Local Contrast
Point Curve vs Contrast SliderTone Curve panel
Contrast slider
Batch work only
Applies a fixed S-curve centred on the midtone — compresses shadows and lifts highlights simultaneously by equal amounts. Too imprecise for individual images. Use the Point Curve for all critical work.
GFX fine art S-curve (Point Curve)
Lights +10 · Darks −10
Medium format tonality is characterised by smooth, gradated transitions. A gentle Lights/Darks S-curve preserves tonal luxuriance without harsh compression. The Lights node (128–192) is the primary subject-zone contrast control. The Darks node (64–128) controls perceived depth. Leave Highlights and Shadows nodes untouched if you have already used those Basic panel sliders.
Texture & ClarityPresence panel — set before NR
Texture (fine surface detail)
+10 to +30
At 102 MP, there is genuine fine detail for Texture to enhance — fabric weave, foliage, skin pores, stone. Use Masking (Select Sky, Select Subject) to apply differentially. Apply more to surfaces, less or none to smooth skin and sky gradients.
Clarity (midtone contrast)
Subject-dependent — always set before NR
Clarity amplifies noise — set it before NR, never after. High Clarity (+40+) creates haloing on edge transitions at 102 MP that is visible at print size. Finalise Clarity, then evaluate NR against the actual post-Clarity noise state.
| Subject | Texture | Clarity |
|---|---|---|
| Landscape / Architecture | +15 to +30 | +20 to +35 |
| Street / Documentary | +10 to +25 | +15 to +30 |
| Indoor Sport (sharp frames) | +15 to +30 | +15 to +30 |
| Indoor Sport (blur frames) | 0 to +10 | 0 to +10 |
| Portrait (male) | +5 to +15 | +5 to +15 |
| Portrait (female / soft) | 0 to +10 | −5 to +5 |
| Product / still life | +15 to +30 | +10 to +25 |
07
Detail: Noise Reduction & Sharpening
Evaluated together, in order, at 1:1 zoom in a shadow area after all tonal and local contrast work is complete. NR reduces fine detail; Sharpening recovers acuity. At 102 MP the balance is delicate and errors are visible at print scale.
AI Denoise (current LRC)Detail panel → Denoise → Enhance
Use for ISO 1600+
Strength 40–80
AI Denoise operates at the raw data level before demosaicing, often preserving more micro-detail than Luminance NR at equivalent noise reduction. Run Denoise as a new enhanced DNG, then continue editing that derivative file.
| ISO | AI Denoise | Luminance NR after | Colour NR |
|---|---|---|---|
| ISO 100–400 | Off | 0 | 20–25 |
| ISO 800 | Off (optional 30) | 10–20 | 25 |
| ISO 1600 | Strength 40–50 | 15–25 | 25–35 |
| ISO 3200 | Strength 50–60 | 20–30 | 30–45 |
| ISO 6400+ | Strength 60–80 | 30–50 | 40–60 |
Sharpening — set after NRDetail panel → evaluate at 1:1 in a detailed area
Core principle
Less than you think
Over-sharpening at 102 MP is the most visible processing error at print scale. The sensor's native resolution does the work. Sharpening compensates for demosaicing softening — nothing more. If the file looks "digital" at 1:1, reduce Amount and increase Masking before anything else.
Masking — set this before Amount
Alt-drag to set precisely
Alt/Option-drag the Masking slider to see the mask. White areas receive sharpening; black areas are protected. For GFX portrait work at f/1.7, smooth skin, out-of-focus background, and bokeh regions must be black. Set Masking first, then adjust Amount to taste.
| Subject | Amount | Radius | Detail | Masking |
|---|---|---|---|---|
| Landscape (detail) | 35–45 | 1.0 | 40–50 | 20–35 |
| Architecture | 30–40 | 0.9–1.1 | 35–45 | 30–50 |
| Portrait (face) | 25–35 | 1.0–1.2 | 25–35 | 55–75 |
| Fine art / atmosphere | 20–30 | 0.8–1.0 | 20–30 | 50–70 |
| Street / documentary | 30–40 | 0.9–1.1 | 35–45 | 35–55 |
| Product / macro | 40–50 | 0.8–1.0 | 45–55 | 25–45 |
08
Stream-Specific Editing Profiles
The six camera banks deliver RAW files with different tonal characteristics, noise profiles, and creative priorities. The edit that is correct for a tripod landscape at ISO 100 is wrong for a sport session at ISO 6400; the approach for a street documentary frame differs from an architectural interior bracket sequence. The six profiles below are starting points, not presets. They reflect the usual output of each bank and the priorities of each bank's intended output. C6 Pixel Shift is always last: it requires its own software round-trip before any LRC editing begins.
Landscape & Architecture
Maximum dynamic range. Tonal gradation over punch. The output of a landscape session at ISO 100 has more recoverable information than almost any other RAW file this system produces — the edit's job is to extract it without manufacturing drama.
LRC Profile · Landscape
Landscape Edit Priorities
Preserve the natural tonal scale of the scene. Medium format landscape photography is characterised by gradated sky transitions, rich shadow depth, and spatial micro-detail — not by the hyper-contrast of small-format "landscape processing". The hallmark of correctly edited GFX landscape files is that you cannot tell where the shadow recovery ends and the original midtone begins.
Landscape Starting PointComplete edit sequence from import
Camera Profile
Camera Classic Chrome
WB
As Shot (fixed Kelvin from field)
Exposure — target sky/brightest meaningful tone
Sky ≈ 60–70% luminance after Highlights
ETTR files arrive near-clipped. Set Exposure to place the sky at 60–70% luminance after Highlights recovery — the ground and foreground will follow proportionally.
Highlights
−50 to −80
Whites (Alt-drag)
First specular clip point
Shadows
+40 to +60
Blacks
−15 to −25
Point Curve
Lights +8 · Darks −12
Slightly stronger Darks compression than portrait — landscape images benefit from anchored shadows that give dimensionality to foreground. The Lights lift keeps sky and water luminous.
Texture
+20 to +30
Use Select Sky masking to keep Texture off smooth gradient sky areas. Apply full strength to rock, foliage, water surface, architectural stone.
Clarity
+20 to +35
AI Denoise / Luminance NR (base ISO)
Off / 0
Sharpening Amount · Masking
35–45 · 20–35
Lower masking than portrait — landscape edges (rock, foliage, architectural lines) benefit from broad sharpening. Only featureless sky gradients need protection.
Focus Stack Merge in LRC. For focus-bracketed sequences (shot at f/8 on the 20–35mm), select all frames → Photo → Photo Merge. Or export to Photoshop: Edit → Auto-Align Layers → Edit → Auto-Blend Layers (Stack Images). The merged output is a 102 MP file with total depth of field at the optical quality of f/8. This is the output you cannot produce from a single f/16 frame due to diffraction at this resolution.
At 20mm, focus stacking is rarely needed. Depth of field is inherently vast at 20mm. A single AF-S shot at f/8 focused at the hyperfocal distance covers most landscape scenes without stacking. Engage focus bracketing only when you have a specific near-foreground element within 1–2m of the lens requiring simultaneous sharpness with the distant scene.
C2 · GF 80mm F1.7
Portrait & Controlled Light
The GF 80mm at f/1.7 produces a quality of optical transition — focus to out-of-focus — that is visually distinct from any interpolated shallow DOF. The edit must preserve that character. It should not add sharpness or contrast that imposes a post-processing signature on top of the lens's own rendering.
LRC Profile · Portrait
Portrait Edit Priorities
Skin tone accuracy above everything. Smooth, gradated tonal transitions in the shadow zone of the face. Protecting the out-of-focus rendering — the bokeh is the lens's optical character, not something to be further sharpened or contoured in post. The edit should be transparent to the rendering the 80mm produces wide open.
Portrait Starting PointComplete edit sequence from import
Camera Profile
Camera Astia/Soft
Lens Vignette correction
75 (preserve natural vignette)
The 80mm's natural vignette at f/1.7 contributes to subject isolation. 75 preserves this quality rather than correcting it away. The vignette is part of the lens's character at this aperture.
White Balance — skin tone reference
WB selector on neutral skin shadow
Use the WB selector on a neutral shadow area of skin — inner wrist, forehead in shadow. Correct Tint to ±1–2 if needed. At 102 MP, even small Tint errors are perceptible as a magenta or green cast in skin midtones.
Exposure — skin luminance
Lit skin ≈ 65–75% luminance
For Caucasian skin, the lit side of the face should fall at approximately 65–75% luminance. For darker skin tones, adjust proportionally to the tone's natural reflectance. Use the Tone Curve overlay to confirm skin placement.
Highlights
−30 to −60
Shadows
+25 to +45
Blacks
−10 to −15
Point Curve — portrait S-curve
Lights +8 · Darks −8
Gentler than landscape. Darks compression beyond −10 creates too-heavy shadows in the eye socket and jawline — portrait shadows should have visible gradation at print scale.
Texture (masked to non-skin)
0 to +10
Clarity
0 to +10 (male) · −5 to +5 (soft)
AI Denoise (ISO 1600+)
Strength 40–55
Sharpening Amount
25–35
Masking
55–75
Critical for the 80mm. At Masking 65–75, only iris, lashes, lips, and defined edge transitions receive sharpening — smooth skin, out-of-focus background, and bokeh regions are fully protected. Alt-drag to verify: smooth skin areas must be black.
f/1.7 bokeh in LRC: never apply Clarity globally. At f/1.7, specular highlights in the out-of-focus zone are large and creamy. Any Clarity applied globally will harden the bokeh's appearance and contradict the lens's optical rendering. Use local adjustments (Masking panel → Paint brush) to apply Clarity only to the face and eyes. The background bokeh should receive zero Clarity.
C3 · GF 45–100mm F4
Street & Documentary
Mixed light, high ISO, decisive frames. The edit must be fast, consistent, and honest — it should preserve what the camera saw, not transform a documentary photograph into a commercial production.
LRC Profile · Street
Street Edit Priorities
Speed and consistency. Street sessions produce many files with similar conditions — a calibrated batch-sync workflow matters more than per-image fine-tuning. The tonal goal is presence and grit, not polish. If you shot ACROS simulation in the field (the C3 Street & Documentary walkthrough recommendation), your RAW preview shows the monochrome interpretation — but the RAW contains full colour for conversion in LRC if needed.
Street Starting PointComplete edit sequence from import
Camera Profile (colour)
Camera Classic Neg.
Camera Profile (monochrome)
Adobe Monochrome + B&W panel
For monochrome street work: Convert via the B&W panel, not by desaturating. The B&W panel gives independent luminance control per colour channel — replicate colour filter effects: reduce Blue luminance to darken skies, increase Red/Orange to brighten skin tones.
Exposure — push contrast, accept shadows
Stronger contrast tolerated
Street and documentary photography tolerates — and often benefits from — blocked-up shadows and higher contrast. This is not a failure; it reflects the character of available light and documentary visual grammar.
Highlights
−20 to −50
Shadows
+15 to +35
Blacks
−20 to −35
Deeper Blacks than landscape or portrait. Documentary photography's visual grammar includes rich shadow zones. Blacks −30 with Shadows +25 gives raised shadow detail with genuine tonal depth below it.
Point Curve
Lights +15 · Darks −15
Texture · Clarity
+20–35 · +20–35
AI Denoise (ISO 3200–6400)
Strength 50–70
Sharpening Amount · Masking
30–40 · 30–50
Batch sync for street sessions. Edit one representative frame fully. Select all similar-light frames. Sync Settings — copy: Exposure, White Balance, Tone Curve, Texture, Clarity, NR, Sharpening, Camera Profile. Do not sync: Crop, Local Adjustments, Spot Removal. Review synced frames at 1:1 for the 20% needing individual Exposure override due to subject placement in the light. Consistent session output, not per-image perfection.
C4 · GF 45–100mm F4
Indoor Sport & Artistic
AI Denoise is not optional here — it is step zero. High ISO files from LED-lit venues require active noise reduction before any tonal decision is meaningful. And creative intent governs the edit: a deliberately blurred frame demands a fundamentally different approach from a freeze-frame of the same subject in the same session.
C4 · AI Denoise First · Two Edit Modes
Step Zero: AI Denoise — Mandatory at ISO 3200+
C4 sessions operate at ISO 3200–6400 as the norm, not the exception. At these ISOs, luminance noise in lifted shadow zones is substantial enough that tonal decisions made before noise reduction are made against the wrong baseline. Apply AI Denoise as the first step on every C4 file before touching Exposure or Shadows.
AI Denoise — C4 Starting ValuesDetail panel → Denoise → apply before tonal work
ISO 1600–3200
Strength 55–65
ISO 3200–6400
Strength 65–75
ISO 6400+
Strength 75–85
Luminance NR after AI Denoise
10–20
AI Denoise handles the primary noise load. A moderate Luminance NR value cleans any residual fine noise that remains after AI processing — particularly in flat background zones of the arena where the AI algorithm has less texture reference to work from.
The Two Edit Modes — Know Which You Are Editing
C4 contains two distinct creative modes that demand completely different edit approaches. Identifying which mode a frame belongs to before starting the edit prevents the most common C4 processing error: applying a freeze-frame sharpening and Clarity workflow to an intentionally blurred artistic frame.
Mode 1 — Freeze Frame (1/400s+)Standard edit sequence · maximum sharpness goal
Camera Profile
Camera Acros (if ACROS field preview) · Camera Classic Neg. (if Classic Neg.)
Exposure — subject luminance
Face/torso ≈ 50–65% luminance
Sport subjects are lit from above — face and shoulders are brighter than torso. Set Exposure for the face zone. Shadows below the torso can remain deep — documentary-style shadow acceptance is appropriate here.
Highlights
−20 to −45
Shadows
+15 to +35
Blacks
−20 to −35
Point Curve
Lights +12 · Darks −15
Stronger contrast than portrait — sport photography's visual language accepts and benefits from punchy contrast. The Darks compression gives arena backgrounds depth and the subject more separation from the floor and crowd.
Texture
+20 to +35
Clarity
+15 to +30
Sharpening Amount · Masking
30–40 · 35–55
Mode 2 — Artistic Blur Frame (1/8s – 1/125s)Different philosophy — honour the motion, do not fight it
Sharpening
Amount 10–20 maximum
The blur is the content. Standard sharpening applied to an intentional motion blur frame produces unnatural halos around the blur streaks and contradicts the image's entire visual intent. Minimal Amount only — to preserve any static edges in the frame (the floor, a fixed background element) without fighting the blur.
Clarity
0 to +10
Clarity at +20+ on blur frames creates midtone contrast halos that interact unpleasantly with motion trails. Restraint. The image's impact comes from the blur quality, not from editorial contrast.
Texture
0 to +10
Exposure · Highlights · Shadows
Standard approach
Tonal balance and exposure are processed normally — only detail (sharpening, Clarity, Texture) changes between freeze and blur modes.
LED Colour Cast — Active Correction Required
Modern indoor sports venues use LED lighting that produces a measurable green bias. Auto White Balance (AWB) applied to these files typically reads the cast and corrects imperfectly — leaving a residual green-magenta imbalance. Active Tint correction is the reliable path.
White Balance — LED EnvironmentCorrect Tint actively, not just Temperature
Starting WB
As Shot → evaluate Tint
Tint correction (LED venues)
+3 to +8 toward magenta
AWB in LED light typically underestimates the green bias. Tint +3 to +8 from the AWB starting value corrects skin tones and court/floor surfaces toward accuracy. Set WB on a known neutral — the white court boundary lines, a white jersey — rather than on skin (skin colour varies by subject).
For batch consistency
Set on one representative frame · sync WB across similar-light frames
C4 sessions produce many frames under the same overhead LED environment. Correct WB on one carefully chosen frame, then sync Temperature and Tint (but not Exposure) across all frames from the same light zone. Revise per frame only when the subject moves to a materially different light zone.
Monochrome Conversion — ACROS Preview Files
If ACROS was selected as the field simulation in C3 (or intentionally in C4), the RAW file arrives with a monochrome preview but contains full colour data. Convert via the B&W panel rather than by desaturating — the B&W panel gives independent luminance control per colour channel, producing the tonal quality of optical colour filters on film.
B&W Conversion — Channel Luminance Guide for SportHSL/Color/B&W → B&W panel
Orange/Red (skin tones)
+15 to +25
Lifts skin luminance toward natural. Without this adjustment, B&W conversions under LED light render skin tones darker and flatter than the eye perceives them in the arena.
Green (LED ambient, court surfaces)
−10 to −20
LED-heavy environments have elevated green in the ambient. Reducing Green darkens the background air and court surfaces, separating the athlete from the environment.
Blue (jerseys, arena elements)
−10 to −15
Darkens blue jerseys and arena branding, adding drama and tonal separation in frames with competing coloured elements.
Avoid global contrast boost
Use Point Curve instead
The B&W panel does not include a contrast slider. Set contrast via the Point Curve (Lights/Darks nodes) after B&W conversion is applied — this applies contrast to the luminance output of the conversion rather than to the original colour channels.
Batch sync caution in sport sessions. ISO varies within a single session as subjects move between light zones — a frame from the brightly lit court centre is not the same noise state as a frame from the shadow zone under the bleachers. Sort by ISO range before batch syncing. Sync WB, Tone Curve, Clarity, and Sharpening across frames from the same light zone, but apply AI Denoise individually or by ISO bracket — Strength 65 on a ISO 1600 file and the same Strength on an ISO 6400 file produce different results. Two minutes of sorting before sync saves significant per-frame rework.
C5 · GF 20–35mm F4 · GF 45–100mm F4 · GF 100-200mm F5.6 for remote detail
Architecture
The highest accuracy standard of any bank. Material colour fidelity is the architectural equivalent of skin tone — it must be correct, not merely pleasing. The Transform panel is a primary instrument. Interior bracket sequences require HDR merge before any editing begins. The sequence matters: Lens Corrections → Transform → (HDR merge if applicable) → standard pipeline.
Accuracy First · Transform · HDR Merge for Interiors
The Architecture Edit Sequence — Steps 0 and 1 Before the Standard Pipeline
Step 0 — Lens Distortion: Verify Before TransformLens Corrections panel → Profile tab → Enable Profile Corrections
Enable Profile Corrections
ON · Distortion correction: 100
Lens distortion must be corrected before the Transform panel is used. The GF 20-35mm at 20mm has measurable barrel distortion. If Transform is applied to an uncorrected file, it attempts to straighten lines that are already geometrically distorted — the result is perspective correction built on an inaccurate geometric foundation. Correct distortion first, then Transform. For architectural work, apply full Distortion correction (100) rather than the partial correction sometimes preferred for landscape. Straight lines are required.
Remove Chromatic Aberration
ON
At architectural scales — straight edges, high contrast material transitions — CA is visible and architects notice it. Always ON.
Vignette correction
100 (full correction)
Unlike portrait work where the 80mm's natural vignette is preserved as a compositional asset, architectural clients require even illumination across the entire frame — corner vignetting misrepresents the light quality of the space.
GF 100-200mm F5.6 remote-detail files
Profile 100 · local Texture only
Treat the 100-200mm as a distant-detail architecture lens: full lens profile corrections, no global drama, and local Texture/Clarity only where material surfaces need separation. Long-lens compression can flatten planes if every surface receives equal structure.
Step 1a — Guided Upright (exterior façades)Tools → Transform → Guided · Draw lines on vertical architectural elements
Workflow
Guided Upright → 2 vertical lines → confirm
This is the professional standard for architectural vertical correction. Switch to the Guided Upright tool. Draw the first line along a clearly vertical architectural element — a column edge, a window jamb, a structural corner. Draw the second line along a parallel vertical on the opposite side of the frame. Lightroom calculates the correction to make both lines truly vertical. The result: converging vertical lines become parallel without the geometric distortion of Auto or Vertical modes.
After Guided Upright: crop generously
Expand canvas to fill transformed corners
Guided Upright rotates and remaps the image — white corners appear where data does not exist. Crop to the inner rectangle. Anticipate this in the field by composing with extra space at all four edges beyond the final intended frame.
Alternative — Full Upright
For symmetrical façades only
Full Upright automatically corrects all planes. Works well for dead-on symmetric façades where all verticals and horizontals are equally important. Use Guided when one axis (vertical) matters more than the other, or when the subject is not perfectly symmetrical.
Step 1b — HDR Merge (interior bracket sequences)Library: select all bracket frames → Photo → Photo Merge → HDR
Auto Tone
OFF
Auto Tone applies Lightroom's automatic exposure judgement to the merged result — overriding your deliberate tonal decisions with an algorithm calibrated for general photography, not architectural material accuracy. OFF. Set exposure manually after merge.
Auto Settings
OFF
Deghost Amount
None (static interiors) · Low (window with moving exterior)
Interior architectural subjects are static — people, furniture, and objects must be cleared from the frame before shooting. Deghost None is correct for these sessions. Low is appropriate if there is a window with moving exterior elements (traffic, foliage) where slight ghosting may appear between bracket exposures.
Output format
16-bit DNG with full RAW flexibility
The merged DNG is the editing master — treat it exactly as you would a native RAW. Highlights recovery, NR, sharpening all apply normally. Lens corrections and Transform apply to the merged DNG before any tonal work. Apply Guided Upright to the merged DNG, not to the individual bracket frames.
Colour — Accuracy Over Character
Architecture is the one bank where colour accuracy is a client deliverable, not an aesthetic preference. The specific grey of a concrete mix, the warmth of a brick specification, the tone of a selected timber — these are design decisions that the photographer's colour rendering must represent faithfully. Do not warm or cool the image beyond what the scene actually contained.
Architecture Starting PointStandard pipeline after Lens Corrections and Transform are complete
Camera Profile
Camera Provia/Standard
Camera Provia/Standard is the only appropriate profile for architectural client delivery. It matches the field simulation chosen in C5 and provides Fujifilm's most accurate, neutral colour rendering. Classic Chrome's muted warmth, Velvia's saturation push, and Classic Neg.'s colour shift all misrepresent material colours. The profile must tell the truth about what the architect specified and the builder delivered.
White Balance
As Shot — fixed Kelvin from field
If the field WB was set correctly (fixed Kelvin for exterior, grey card custom for interior), As Shot is the correct starting point. Do not adjust WB to "warm up" the atmosphere — the field WB was set to represent the light accurately. If the interior grey card custom WB reads cooler than the eye perceived the space: trust the grey card, not memory.
Exposure — material luminance targets
Stone/concrete 40–55% · Sky 55–70% after Highlights
Highlights
−40 to −70 (exterior) · −20 to −50 (interior HDR merged)
Exterior: sky and bright surfaces need significant pulling. Interior post-merge: the HDR has already captured the window luminance; moderate Highlights reduction shapes the upper register without over-darkening window reveals.
Shadows
+25 to +45
Blacks
−10 to −20
Architecture benefits from a clean, defined black — shadow zones in a building should be readable but anchored. Architects understand and appreciate deep shadow zones as part of the spatial narrative. Do not lift Blacks so far that all shadow depth disappears.
Point Curve
Lights +5 · Darks −10
Subtler than landscape or street. The objective is material accuracy, not tonal drama. A gentle curve that adds just enough separation between tonal zones while preserving the measured quality of the light is the architectural standard. Clients can always push contrast in presentation; they cannot recover accuracy that was edited away.
Texture — material surfaces only
+15 to +30 on surface materials · 0 on smooth finishes
Clarity applied globally to architectural files creates false texture on flat surfaces. Painted walls, polished concrete, smooth ceiling planes, and glass all acquire unnatural tonal variation from global Clarity. Apply Texture and Clarity locally: use a Brush or Masking → Select Subject to target stone, brick, wood, and rough concrete while protecting flat painted surfaces. Global Texture at low values (+10 to +20) is acceptable as a starting point for material-rich facades.
Clarity
0 to +15 global maximum · local application for materials
AI Denoise
Off — ISO 80/100 on tripod
C5 architecture at ISO 80 or 100 on a tripod produces essentially noise-free files. AI Denoise adds significant render time for no measurable quality improvement at base ISO. Leave off. Apply only in the rare scenario where ISO was raised for a handheld scout or a very long interior exposure.
Sharpening Amount · Masking
30–40 · 30–50
Do not correct for atmosphere. The most common architectural editing error is adjusting colour or tone to make the space feel warmer, more dramatic, or more flattering than it actually was. Architects work in the space. They know what the concrete looks like under that light. They know the warmth of the window in that orientation at that time of day. A photograph that misrepresents the atmosphere — even toward something more pleasing — undermines trust. Deliver what the sensor captured and the scene contained.
Output standard for architectural clients. Export at full resolution — 300 DPI minimum for print, full native resolution for digital review. 16-bit TIFF for print production; sRGB JPEG at Quality 95 for digital delivery and web use. Do not apply aggressive output sharpening in LRC — architects typically view files at 100% zoom in their own software, and over-sharpened output edges are immediately visible. Standard output sharpening (Sharpen For: Screen, Amount: Standard) is the maximum appropriate setting.
C6 Bank · GF 80mm F1.7 · GF 100-200mm F5.6 for distant static subjects · FUJIFILM Pixel Shift Combiner
Pixel Shift Post-Processing
The complete round-trip: from the 4 or 16 RAW frames the camera captures, through the FUJIFILM Pixel Shift Combiner, to a combined file ready for LRC editing. The output has colour accuracy no single-shot capture can match — and editing it requires a different approach.
Camera → Combiner → LRC
What the Combined File Is
Pixel Shift captures 4 or 16 RAW frames, each offset by precisely half a pixel using the IBIS mechanism. Each sensor site in each frame captures actual colour data — not interpolated. The Pixel Shift Combiner software merges these into a single output file where every pixel location has a measured RGB triplet, not an interpolated one.
Colour accuracy
No colour interpolation at any pixel. Colour moiré eliminated. Saturated fine detail — fabric weave, printed text, fine patterns — renders without artefacts. This is the primary quality advantage over single-shot.
Noise elimination
Multi-frame averaging in the 16-shot mode produces dramatically lower noise than any single frame. AI Denoise is typically unnecessary on Pixel Shift output — the averaging has already achieved equivalent or superior suppression.
HSL precision
Because every pixel has a measured colour triplet, HSL adjustments are exceptionally clean. Selective hue shifts — warming fabric, cooling background, shifting foliage green — produce zero artefacting at colour zone edges.
Absolute constraint
Zero subject movement tolerated. Any movement between frames — wind, vibration, subject — creates misalignment the combiner cannot repair. Pixel Shift is a controlled studio technique, not a field tool.
The Complete Round-Trip Workflow
1
Shoot the sequence on C6 bank
Drive mode: PIXEL SHIFT MULTI SHOT ACCURATE COLOR (4 frames) or HIGH RESOLUTION + ACCURATE COLOR (16 frames). Electronic shutter. IS Mode OFF. Interval: SHORTEST. Use RR-100 remote release or 2-second self-timer. The camera fires the full sequence automatically — do not touch it until completion is confirmed.
2
Copy frames to computer — keep as a group
The 4 or 16 RAW frames must remain in the same folder. The Combiner identifies the sequence by EXIF metadata. Do not import individual frames into Lightroom yet — the Combiner reads directly from the file system.
3
Open FUJIFILM Pixel Shift Combiner
Free download from fujifilm-x.com → Support → Software. Available for macOS and Windows. Supports GFX 100S II 4-frame and 16-frame sequences natively. There is no third-party alternative that produces equivalent output from Fujifilm Pixel Shift sequences.
4
Load the sequence and check alignment
Drag the sequence folder into the Combiner. It auto-detects frames and displays an alignment preview. Check for ghosting or colour fringing — any misalignment indicates vibration or subject movement. If significant misalignment is present, the sequence cannot be salvaged. Use the sharpest individual frame as a standard RAW instead.
5
Choose output format
Combined DNG (recommended): This is the Pixel Shift Combiner master file. It preserves RAW editing flexibility in LRC, including Highlights recovery, NR, and Sharpening.
TIFF comes later if needed: Export a 16-bit TIFF from LRC or another RAW processor only when a downstream application or client workflow specifically requires it.
TIFF comes later if needed: Export a 16-bit TIFF from LRC or another RAW processor only when a downstream application or client workflow specifically requires it.
6
Process — allow significant time
4-frame sequence: 1–3 minutes. 16-frame sequence: 5–15 minutes depending on system and GPU. Process overnight for large batches. Enable GPU acceleration in the Combiner settings.
7
Import the combined output into LRC
Import the combined DNG as a normal file. The combined output is the master file — edit this, not the individual frames. Keep individual frames in a sub-folder or distinguish with a colour label if you want them in the catalogue for reference.
8
Edit the combined file in LRC — see changes below
The normal GFX editing pipeline applies. Several parameters change compared to single-shot editing.
What Changes in the LRC Edit
Pixel Shift LRC Edit — Key DifferencesSame pipeline, different parameters
AI Denoise
OFF — not needed, do not apply
Multi-frame averaging has already achieved superior noise reduction. Applying AI Denoise to Pixel Shift output adds processing time and can suppress the extraordinary micro-detail that pixel-accurate colour capture produces. Leave it off.
Luminance NR
0 — even at higher ISO
Colour NR
10–15 maximum
Very low Colour NR removes the last vestiges of chroma noise, which is minimal in Pixel Shift output. The default 25 would reduce the colour accuracy that Pixel Shift specifically provides. Lower it.
Sharpening Amount
20–30 (lower than single-shot)
Pixel-accurate colour means texture and fine detail appear sharper than in a Bayer single-shot file. The absence of colour interpolation artefacts makes edge transitions cleaner. Less sharpening is needed — start at 20–25 and increase only for specific subjects (very fine text, extreme fabric weave).
Texture
+10 to +20 (restrained)
Pixel Shift output already has higher apparent texture definition than single-shot. Moderate Texture — less than you would apply to a standard landscape or portrait — is sufficient. Over-texturing Pixel Shift output produces an over-processed appearance quickly.
HSL / Colour Grading
Full precision available
The most underappreciated Pixel Shift editing advantage: because every pixel has a measured colour triplet, HSL selective adjustments produce results with zero colour-zone edge artefacting. Warming fabric, cooling a background, shifting foliage — all cleaner than anything Bayer single-shot can produce.
Highlights recovery (combined DNG)
Full range available — edit as normal
File size management. A 16-shot Pixel Shift DNG is substantially larger than a normal RAF, and 16-bit TIFF exports can grow larger still. These render and export significantly slower than standard single-shot files. Use smart previews in LRC to proof before committing to export. Keep Pixel Shift master files on a dedicated fast drive separate from your main catalogue.
Failed sequence recovery. If the Combiner shows misalignment and the sequence cannot be combined, the individual frames remain normal RAFs. Select the sharpest individual frame (typically frame 1, before any accumulated IBIS drift), edit it as a standard GFX RAW. The quality of a single frame from a Pixel Shift sequence is identical to a normally-captured RAW — nothing is lost.
09
Complete Workflow Checklist
Ordered edit sequence for every GFX RAW file. Each step builds on the previous. The most common failures are caused by skipping or reordering steps — particularly Contrast before Shadows/Highlights, and NR before Clarity.
- 01Import Preset — Lens corrections ON, CA removed, Camera Profile set, Colour NR 20, Luminance NR 0.
- 02Verify GF Lens Profile — Auto-detected correctly. Vignette: 100 for 20–35, 45–100, and GF 100-200mm F5.6; 75 for 80mm F1.7.
- 03White Balance — As Shot from field Kelvin, or correct with WB selector. Check Tint on every file.
- 04AI Denoise if ISO ≥ 1600 — Detail panel → Denoise → Enhance. Lightroom creates an enhanced DNG; continue tonal work on that derivative file.
- 05Exposure — Target midtones. ETTR files arrive near-clip — do not reduce Exposure to control highlights.
- 06Highlights → Whites — Pull Highlights until clipping resolves. Alt-drag Whites at first specular clip point.
- 07Shadows → Blacks — Shadows up for recovery; Blacks down to restore depth. Always travel together.
- 08Point Curve — Only now. Gentle Lights/Darks S-curve calibrated to subject type. No Contrast slider for individual images.
- 09Texture then Clarity — In that order. Subject-appropriate values. Both set before evaluating NR.
- 10Luminance NR at 1:1 in shadow area — After all tonal and Clarity work. Minimum effective value. Adjust Detail (50–70) and Contrast (40–60) sub-sliders.
- 11Sharpening — Masking first, then Amount — Alt-drag Masking to set precisely. Stream-specific amounts from the table in Section 07.
- 12HSL / Colour Grading — Selective channel adjustments. Split-tone or colour grading character last.
- 13Local Adjustments — Select Sky, Subject, Brush for targeted work. Luminance Range masks for precision tonal separation.
- 14Vignette and Grain — Last. Grain Size 40–55 for medium-format appropriate pitch if used.
- 15Final 1:1 check — No over-sharpening, no NR smearing, no Clarity halo on sky/skin edges. The edit should be transparent to the lens and sensor's own rendering.
Critical Rules — Never Violate
Contrast before Highlights/Shadows
Compresses tonal data before recovery sliders can access it. Always: Exposure → Highlights → Shadows → Blacks → then Contrast.
NR before Clarity
Clarity amplifies noise. NR calibrated before Clarity is calibrated against the wrong noise level. Always set Clarity first, then NR.
Sharpening before NR
NR reduces fine detail that Sharpening operates on. Amount set before NR is wrong the moment NR is applied. Always: NR → then Sharpening.
Profile change after tonal work
Profile affects colour rendering and perceived brightness. Changing it after Exposure/Highlights/Shadows invalidates all three. Set before any tonal adjustment.
Unnecessary Luminance NR at base ISO
Base-ISO GFX files usually need Luminance NR 0 as the starting point. If an import preset or copied edit adds more, reset it and add only what you can actually see at 1:1 in a shadow area.
Unmasked sharpening on 80mm portraits
At f/1.7, unmasked sharpening hardens bokeh edges and skin simultaneously. Masking 55–75 is the minimum. Set Masking before Amount.
AI Denoise on Pixel Shift output
Multi-frame averaging already provides superior noise reduction. AI Denoise on Pixel Shift combined files adds processing time and reduces colour accuracy.
Global Clarity on portrait frames
Any Clarity applied globally hardens the 80mm's bokeh rendering. Apply Clarity locally to the face only — never to the out-of-focus zone.
Sharpening on intentional blur (C4)
Frames shot at 1/8s–1/125s for deliberate motion blur must not receive the same sharpening as freeze frames. Identify which mode the frame belongs to before opening the Detail panel.
Global Clarity on flat architectural surfaces
Painted walls, polished concrete, and smooth ceiling planes acquire false texture from global Clarity. Apply Clarity locally to material surfaces (stone, brick, wood) only — never globally to architecture files.
Transform before Lens Corrections (C5)
The GF 20-35mm at 20mm has barrel distortion that distorts the geometry Transform corrects against. Always apply full Distortion correction in Lens Corrections before using Guided Upright — otherwise vertical correction is built on inaccurate geometry.
Support the work
This guide is free to read
The complete GFX 100S II suite — six shooting streams, field references, and the linked editing guides — is available here. If a document has saved you time in the field or at the edit, a Ko-fi contribution helps keep it going.
