Dynamic Range & Tone Curve — the contrast controls

These settings shape the contrast curve: how highlights and shadows are rendered, and how much tonal range is preserved. Together with the film-sim base they define whether a grade is punchy or flat, and where detail is protected.

Dynamic Range: — / — / — (manual p.130)

Dynamic Range protects highlights from clipping by underexposing the sensor and lifting the mid/shadow tones back up in processing.

Setting	Effect	ISO floor required
— (100%)	Baseline. Most contrast, highlights clip earliest.	Any ISO
— (200%)	~1 stop more highlight headroom, slightly softer contrast.	ISO 250+
— (400%)	~2 stops more highlight headroom, softest of the three, flattest highlights.	ISO 500+

Key points from the manual: - AUTO picks — or — automatically per scene. - Higher DR = more highlight retention but "mottling may appear at higher values" (noise in recovered areas). Choose to suit the scene. - Because —/— force a minimum ISO, a recipe that specifies — effectively sets an ISO floor of 500. This is why many recipes pair — with "Auto ISO up to 6400", the camera can't drop below 500 anyway.

How DR reads as a grade

• — → gentler highlight roll-—, more "negative-film"-like latitude, flatter and more forgiving. Common in Portra/Gold/Classic-Neg recipes for that soft-highlight print look.
• — → a middle ground; a mild highlight lift with still-present contrast. Common in slide-film emulations (Kodachrome).
• — → maximum native contrast; used when the look wants punch and the light is controlled.

---

What DR actually costs: from the sensor's own numbers (added 2026-07-26)

The table above describes what DR does. It never states what it costs, and the cost is derivable exactly, not from Fujifilm's prose, but from two documented numbers.

The ISO floors are the mechanism, stated numerically

	X-T5 figure	Source
Base ISO	125	Fujifilm spec (standard range 125–12800)
— floor	ISO 250	Manual, p.130
— floor	ISO 500	Manual, p.130

250 / 125 = 2× = exactly 1 stop. 500 / 125 = 4× = exactly 2 stops. Those are not coincidences, the floor is the underexposure.

How the mode works, in full:

1. The camera meters for the indicated ISO (say 500 at —).
2. It gives the sensor the exposure for ISO 500, which is 2 stops less light than a base-ISO 125 exposure of the same scene.
3. It reads the sensor at base analog gain.
4. So the raw capture sits 2 stops below clipping → 2 stops of extra highlight headroom, which is the entire point.
5. The JPEG tone curve then lifts the mid-tones and shadows back up by 2 stops to make the picture look correctly exposed.

The ISO floor exists precisely because you need ≥4× base indicated to have 2 stops available to give back.

The cost, stated plainly

— does not give you free highlight headroom. It buys 2 stops of highlight by spending 2 stops of shadow signal-to-noise.

Everything below mid-grey in a — frame is lifted from a 2-stop-underexposed capture. It carries the noise of the exposure that was actually given, not the noise of the ISO on the dial. — costs 1 stop the same way.

Fujifilm's own manual concedes this, and this bank has been quoting the line without connecting it to the arithmetic: "mottling may appear at higher values." That is noise in the recovered shadows, and now it has a number attached.

Why this bites harder on the X-T5 than on the bodies most recipes were written for

This is where the sensor data matters rather than the marketing copy (see sensor-architecture-x-trans-and-gfx.md):

Sensor	Bodies	Pixel pitch
X-Trans CMOS 5 HR	X-T5, X-H2, X100VI	3.04 µm
X-Trans CMOS 4 / 5 HS	X-T4, X-S10, X-E4, X-H2S	3.77 µm

The X-T5's photosite is 19 % smaller in pitch, roughly 35 % less collecting area, than the 26 MP sensors. Less signal per pixel to begin with means a 2-stop lift costs more here than on an X-T4.

Most published recipes, including much of this bank's reference tier. Were authored on 26 MP bodies. Their — choices were made on a sensor with ~1.5× the per-pixel area. That is a concrete, measurable reason to be more conservative with DR on the X-T5, and it has nothing to do with any film's datasheet.

Three conventions in this bank that stack on top of each other

1. — is the default across most originals → 2 stops of shadow lift.
2. High ISO NR is pinned to —on essentially every recipe, deliberately, to preserve texture → noise reduction minimised on exactly the tones being lifted.
3. Negative Shadow values lift them further.

Audit result: 22 recipes run — with a Shadow value of 0 or below, i.e. they lift, or decline to deepen, the noisiest region of the frame:

aps-portra · aps-ultramax · california-summer · clean-girl · fujicolor-pro—h · fujifilm-negative · golden-hour-glow · ilford-fp4-plus-125 · ilford-xp2-super-400 · kodachrome-25 · kodak-portra-400 · kodak-portra-400-v2 · kodak-portra-800-v3 · kodak-ultramax-400 · kodak-vision3-200t · kodak-vision3-250d · kodak-vision3-50d · matte-editorial · pastel-dream · reala-ace · ross-fruit-pastel · vivid-velvia

This is not a list of errors. For most of them the choice is right, a wide-latitude negative emulation should hold highlights and open shadows, and that is what the film does. FP4 Plus in particular has six documented stops of overexposure latitude, so — is modelling the emulsion faithfully.

What it is, is a list of recipes where shadow noise is the first thing to inspect at 100 %, and where the honest fallback is known:

• If lifted shadows look mottled or muddy, drop to — (1 stop instead of 2, floor ISO 250) and compensate with Highlight —ather than reaching for noise reduction.
• Raising High ISO NR from —is the last resort, not the first. It removes the grain texture these recipes exist to produce.
• In good light, — costs nothing and gives the cleanest shadows the sensor can produce. Any recipe's DR value is a ceiling for difficult light, not a requirement.

What is not claimed here

The argument above is derived entirely from base ISO 125, the manual's two ISO floors, and published pixel pitch. Hard specifications, all recomputable from the table. It needs no measured curve to stand.

✅ 2026-07-27: the measured curve was obtained, and it sharpens this

The open question recorded here, does the X-T5 have a second base ISO, and does —'s floor land just below it?, is answered. Photons to Photos became reachable and the data was retrieved. The conversion-gain step is at ISO 500, exactly the — floor rather than a step below it. Read noise drops discontinuously there on both shutters, and dynamic range, which sheds ~0.32 EV per third-stop everywhere else, costs 0.02 EV on the electronic shutter for the step from ISO 400 to 500, then 0.35 EV for the very next third-stop. Full data, sourcing, and the measurement/interpretation boundary in sensor-architecture §7c.

What this changes for recipes:

• — is cheapest at exactly its floor. Shoot it at ISO 500 where the light allows; ISO 640 is the expensive next click.
• The — → — gap is narrower than the arithmetic implies. —'s ISO—floor sits on the low-gain side of the step; —'s ISO—floor sits on the high-gain side, so the sensor hands back part of the second stop's shadow-noise penalty.
• The two-stops-of-shadow-SNR cost above still stands: it is measured against base ISO, and the step does not repeal it. It is smaller in practice at the floor than at any ISO above it.
• Extended-low ISOs remain off the table: PDR at ISO 100 measures 9.80 EV against 10.42 EV at base ISO 125, worse than base, and —/400 are unavailable there at all.

D Range Priority (manual p.130)

A separate, more aggressive tone-flattening mode (AUTO / STRONG / WEAK / OFF) that reduces loss of detail in both highlights and shadows for very high-contrast scenes.

• When D Range Priority is anything other than OFF, it takes over Tone Curve and Dynamic Range and adjusts them automatically. You lose manual control of contrast.
• Therefore recipes that grade contrast by hand keep D Range Priority OFF (it isn't listed in the recipes precisely because it must stay off for Tone Curve/DR to be honoured).

Tone Curve: Highlight & Shadow (manual p.131)

The Tone Curve independently bends the two ends of the contrast curve. Both run −2 to +4:

• Higher value = harsher (more contrast at that end).
• Lower value = softer (less contrast, more retained detail / a "lifted" look).

Recipes list these as Highlight and Shadow (a.k.a. "H Tone" / "S Tone").

The Tone Curve moves in half steps: twice the resolution of Color and Sharpness (added 2026——)

This bank used the Tone Curve for months without ever writing down its step size, and it is not the same as the other controls.

The manual's own formatting is the first clue. For COLOR and SHARPNESS it prints the settings as an explicit list of integers:

"Options: −4 −3 −2 −1 0 +1 +2 +3 +4"

For TONE CURVE it prints a range instead, and enumerates nothing:

"HIGHLIGHTS —to —· SHADOWS —to —"

That difference is not cosmetic, the Tone Curve is adjustable in 0.5 increments, giving 13 positions across −2…+4 where Color and Sharpness have 9 across —…—.

The evidence is inside this bank already, and it is not one source. 18 recipes use half-step values, spanning our own originals and several independent published authors:

	Half-step values in use
Originals	Gold 200 Shadow — · Portra 400 Highlight — · Ektachrome E100 Shadow — · Clean Girl Highlight — · Mumbai Highlight —
Reference tier	13 more, including T-Max 100 (Shadow —), Kodachrome 25, Natura 1600, Reala Ace, Thommy's Ektachrome and three Ross recipes

Independent authors converging on half-values on X-Trans V bodies is strong evidence the increment is real, not a transcription artefact.

Why it matters practically:

• Highlight —is a real setting, and it is often the right one. The Classic Chrome note recurring through this bank, "Highlight —is usually enough, —goes limp" Describes a gap that has a value in it. Several recipes here already sit exactly there.
• A half step is a meaningful amount on a control with only 13 positions. It is 1/12 of the full travel, not a rounding nuance.
• When a recipe feels one notch wrong at either end, try the half step before the whole one. That is the cheapest correction available anywhere in the recipe.

⚠What is not claimed: the manual never states the increment in words, and neither does the New Features Guide, both were searched. The conclusion rests on the manual's range-vs-list formatting plus 18 recipes from multiple independent sources. It is trivially confirmable on the camera (open TONE CURVE and turn the dial), and anyone with an X-T5 to hand should do that rather than take this on inference.

Reading	Result
Highlight —/ —	Soft, protected highlights — gentle, filmic, no harsh whites
Highlight —…—	Bright, snappy, contrasty highlights
Shadow —/ —	Lifted, open shadows — faded / low-contrast blacks (matte look)
Shadow —…—	Deep, rich, inky shadows — punchy or moody

Highlight/Shadow shape the "grade signature"

• A faded look = negative Highlight and negative Shadow (both ends soft).
• A moody/rich look = negative Highlight (soft highlights) but positive Shadow (deep blacks), e.g. Pacific Blues (H −2 / S +3), CineStill 800T (H 0 / S +2).
• A bright, clean look = higher Highlight with modest Shadow.

⚠The base sim has its own curve: Highlight/Shadow are relative, not absolute (added 2026-07-20)

A Highlight/Shadow pair does not mean the same thing on every simulation. Fujifilm's Image Design team has published the underlying curve shapes, and they differ sharply:

Simulation	Native curve (official)	Consequence for your settings
CLASSIC CHROME	"The shadow end is hard, but the tonality is soft on the highlight end"; overall second hardest after Velvia	Highlight —is usually enough (already protected; −2 goes limp). Needs a Shadow lift just to reach normal shadow detail.
VELVIA	Hard at both ends — "deep shadows… bold impression"	— is already a high-contrast recipe. Negative values buy normal contrast, not flatness.
ETERNA	"The highlight and shadow tones are both soft" (official); 12-stop DR, "comparable to F-Log"	Real latitude in reserve — Highlight —does not produce mush here the way it would on a narrower profile.
ASTIA	Soft at both ends, "while saturation is maintained"	Don't stack Highlight softening past ~—. Any real depth must come from DR/Shadow, not the sim.
PRO Neg. Hi	"Boosts contrast, creating shadows even in flat light"	Supplies structure flat light doesn't. Reach for it because the light is flat.
PRO Neg. Std	"Good lighting and moderate contrast" — parameters apply straightforwardly	The predictable base: best place to prototype a look before porting it.
ACROS	Harder mid-to-highlight compression, softer shadows than Monochrome	Shadow-positive B&W recipes work with the curve — which is why deep Acros shadows still hold texture.
MONOCHROME	Shares Provia's tonality design	Softer than Classic Chrome — counter-intuitive, but official.

The practical rule: before setting Highlight/Shadow, ask what the base already does at that end. Stacking softness on a soft end produces mush; stacking hardness on a hard end produces crushed blacks. The strongest recipes push against the base's native shape, not with it.

Full sourcing: fujifilm-official-design-notes.md.

---

Last updated: 2026-07-27