Salon Light

An original design for the warm society portrait: a deep black under a bright white, skin and silk carrying the whole palette, and not one pixel of green, cyan or violet in the reference. Derivation in the notes.

Painterly reference: John Singer Sargent, measured from fifteen public-domain reproductions against a twelve-painter control group. Named for the room and the light rather than the painter, which is this bank's standing rule.

Provenance. Built 2026-10-05 from fifteen Sargent oils, public domain, cropped to the canvas and verified by eye. One head study on paper passed the medium filter and was removed by hand after it measured 78 per cent bare support.

The palette has nothing cool in it at all

Three warm families account for 99.7% of coloured pixels, and the entire cool half of the wheel measures zero. That is the most completely one-sided palette in the twelve-painter control group, past even Caravaggio, who reaches 98.8% across two families but still registers a trace of cyan.

The tone is the other half of the picture.

A black point of 7.0 under a white point of 71.6 is a wide range, and a median of 32.1 puts the mass of the picture low without the frame going to the near-black that Raking Light targets. This is a dark room with a well-lit sitter in it, rather than a cellar with a shaft of light.

Why Pro Neg. Hi

Fujifilm's description is the argument: it "boosts contrast, creating shadows even in flat light", and this bank's Dynamic Range & Tone Curve, the contrast controls tone table records it as the base to reach for because the light is flat.

That is historically exact. Sargent worked in a north-lit studio, which is large, soft and directionless, and the drama in these pictures is built rather than found. A base whose published purpose is manufacturing contrast under flat light is doing the same job the painter did.

The rest follows. It is a portrait simulation, and fifteen of fifteen reference works are portraits. Its contrast reaches a black of 7.0 without the hard shadow end that would crush the dark silks. And because the palette has no cool content, nothing is lost to a base with no particular opinion about blue.

How it differs from Fashion Editorial, which shares the base

Fashion Editorial is this bank's existing Pro Neg. Hi portrait recipe and the obvious objection. The two are opposites on colour, which is the axis that matters.

One recipe cools and desaturates a studio portrait toward a magazine's house neutral. This one warms and slightly saturates it toward a lamplit drawing room. They share a base and disagree about every colour decision made on top of it, which is the same relationship Raking Light has with Dark Ground.

The rest

• White Balance +2 Red and −1 Blue, aimed at a highlight b* of +14.33 over a shadow b* of +4.32. Both ends are warm, so this is an uniform shift rather than a split, and it is smaller than Raking Light's +3 Red because the lit tone here measures a* +2.07 against that corpus's +10.06. Sargent's light is warm. Caravaggio's is hot.
• DR200 rather than DR400. One stop of highlight protection covers a white dress or a lit shoulder, which is the recurring highlight in this subject. The second stop would lift everything below the highlights and cost the 7.0 black, which is the half of the look the tone curve is defending.
• Shadow +1 and Highlight −1. The shadow value hardens an already-contrasty base to reach the black. The highlight value holds a white point of 71.6 short of clipping so that silk and linen keep their weave.
• Color +1 with Color Chrome Effect Strong. Chroma measures 13.1, which is moderate, and Pro Neg. Hi is restrained by design. Color lifts the warm passages toward the target, and Color Chrome Effect then puts gradation inside them, which is what stops a red curtain or a copper-lit wall becoming a flat shape.
• Color Chrome FX Blue Off. Cyan-blue and blue-violet both measure 0.0%. There is nothing for it to act on, and this is the clearest case in the bank for the setting being irrelevant rather than merely unwanted.
• Sharpness +1 with Clarity −1. The subject is fabric against skin. Sharpness holds the weave of silk, lace and brocade, which is what Sargent's brush is famous for. Clarity softens the broad modelling across a face, which is what portrait retouching has always done. They act at different scales, the same pairing Pink Chrome uses for a different century.
• Exposure −1/3 to −2/3. The median measures 32.1, so the picture sits below middle grey and a reflective meter will try to correct that. Not as far down as Raking Light's −1, because this room is lit rather than dark.

Variants

Movie mode

Movie mode drops Grain, Color Chrome Effect and Clarity. Grain is near-neutral and Clarity is one notch, so the real loss is Color Chrome Effect Strong, which means the warm passages read flatter. Pro Neg. Hi, the white balance, DR200 and the tone curve all carry, and those produce the deep black and the warm cast. See Video / Movie Mode, settings, differences, and photo→video translation.

Validation tier: Creative design, field-test pending. Grounded in Fujifilm's published design statements for Pro Neg. Hi and each base rejected above, in this bank's Dynamic Range & Tone Curve, the contrast controls tone-curve table, and in a CIELAB measurement with hue census across fifteen verified reproductions against a twelve-painter control group.

No frames have been shot with it.

Test priority: in order

1. A sitter in a dark interior with one broad soft light, metered at −1/3. The whole recipe in one frame. Measure the black and the median: the targets are L* 7 and L* 32, which is a dark picture with a legible face in it rather than a dark picture.
2. Check the frame for anything cool. The reference measures 0.0% green, cyan and violet across fifteen paintings, which is the most one-sided palette in the seam. A blue cushion or a window with sky in it will sit outside the look rather than inside it, and that is a styling decision rather than a settings one.
3. White silk or linen against skin. The Highlight −1 and Sharpness +1 check together. The fabric should hold weave at a white point near 71 while the face stays smooth at Clarity −1.
4. Skin at +2 Red. The most likely setting to have overshot, and the variants table carries the retreat for fair complexions.
5. Shoot the same sitter with this and with Fashion Editorial. The pair is instructive: same base, opposite colour decisions. If they look alike, the white balance and Color are not reaching far enough in either direction.

What would count as the recipe being wrong rather than needing a tweak: if a 7.0 black cannot be reached at DR200 without the mid-tones going with it, then Pro Neg. Hi's contrast is landing in the wrong part of the curve for this subject, and the look needs Classic Chrome's hard shadow end with the white balance doing more undos its cool cast.

Pro tip: The palette has no cool half. Fifteen paintings and not one pixel of green, cyan or violet above the chroma threshold. That is a styling instruction before it is a camera one: dress the frame in skin, silk, wood, brass and lamplight, and keep the blue cushion out of shot.

Pro tip: This is Pro Neg. Hi used for its stated purpose. Fujifilm says the base exists to create shadows in flat light, and Sargent worked in a big soft north-lit studio where the drama had to be built rather than found. Most recipes reach for this base despite its contrast. This one reaches for it because of it.

Pro tip: It shares a base with Fashion Editorial and agrees with it about nothing else. That recipe goes cool and desaturated toward a magazine neutral. This one goes warm and slightly saturated toward a drawing room. If you have both in custom slots, the white balance is the tell.

Pro tip: Meter down, but only a little. A median of L* 32 is below middle grey and the camera will fight it, but this is a lit room rather than Raking Light's cellar. One third of a stop usually does it. Two thirds when the sitter is in dark clothing.

Shoot JPEG. HEIF silently drops Clarity. See Grain & Detail. Grain Effect, Sharpness, Clarity, High ISO NR.

Creative design (field-test pending). See the recipe and the Knowledge base.

Part 0: method

Fifth recipe from the painting seam, measured with the instrument and control group described in the notes Open Air's Part 0: twelve painters, one instrument, only differences between them counted as findings.

Part 1: the corpus, and one removal by hand

Fifteen Sargent oils, public domain, cropped to the canvas, verified by eye against the manifest.

Sixteen were harvested. One, a head study of Rosina Ferrara, passed the Commons medium filter because it is catalogued under painting categories, and is a drawing. It measured a median L* of 93.9 at a mean chroma of 4.3, with 78 per cent of its area bright and near-neutral, which is bare pale paper rather than a painted surface. It was removed by hand.

That is worth recording because it is the second time a category filter has admitted a work on paper, after the Turner failure in September. The filter catches most of them and should not be trusted to catch all of them. Looking at the pictures remains mandatory, and a per-image bare-support check is an useful second pass once a corpus is otherwise verified.

The images were fetched into a scratch workspace and are not in this collection.

Part 2: the measurement

Finding 1: the palette has no cool half

Three warm families account for 99.7% of coloured pixels across fifteen paintings, and the cool half of the wheel measures zero to one decimal place. That is the most one-sided palette in the control group, past Caravaggio's 98.8%, who at least registers a trace of cyan.

The figure deserves one caution. Sargent painted a great deal outdoors, in watercolour and in oil, and that work is full of blue. The open-access holdings that pass a painting filter are dominated by the commissioned studio portraits, which is a real body of work rather than an accident of sampling, but it is not the whole painter. the recipe carries a variant row for the plein-air reading and says plainly that it measures nothing like this.

Finding 2: a wide range with the mass low

A black point of 7.0 under a white point of 71.6 is the second-widest range among the dark painters, and a median of 32.1 puts the bulk of the picture well below middle grey without approaching the near-black of Caravaggio at 15.5 or Rembrandt at 13.5.

That gap is the recipe's subject. Sargent is a dark room with a well-lit sitter in it. Caravaggio is a dark room with a shaft of light in it. The settings that follow differ accordingly: this recipe exposes at −1/3 to −2/3 where Raking Light goes to −1 and beyond.

Finding 3: warm at both ends, mildly

Shadow b* +4.32 against highlight b* +14.33 is a shadow-to-highlight split of −10.01, which is middling for the group. Both ends are warm. The lit end is warmer. Highlight a* is +2.07, which is modest next to Caravaggio's +10.06.

So the light in these pictures is warm rather than hot, and the shadows keep some of it. That settles the white balance at two notches of Red rather than three, and it is the measured difference between a north-lit studio with warm reflected fill and a single hard warm source.

Part 3: the base

Fujifilm publishes Pro Neg. Hi as a base that "boosts contrast, creating shadows even in flat light", and this bank's Dynamic Range & Tone Curve, the contrast controls tone table records the pair rule plainly: reach for Hi because the light is flat, and for Std when the light is already good.

That is historically exact for this corpus. Sargent worked in a large north-lit studio, which is soft, broad and directionless. The contrast in these pictures was built rather than found, and a base whose published purpose is building contrast under flat light is doing the same job the painter did. Most recipes in this bank reach for Pro Neg. Hi despite its contrast. This one reaches for it because of it.

Three supporting requirements follow:

1. A portrait simulation, for a corpus that is fifteen portraits out of fifteen.
2. Enough contrast to reach a 7.0 black without the hard shadow end that would crush dark silk and brocade, which is most of what a Sargent sitter is wearing.
3. Indifference to the cool half of the wheel, which costs nothing when the palette contains none of it.

Part 4: the separation from Fashion Editorial

Fashion Editorial is the bank's existing Pro Neg. Hi portrait recipe, and sharing a base with an existing look is the objection this section exists to answer.

The two agree about the base and disagree about every colour decision made on it.

Three notches of white balance and two of Color, in opposite directions, on top of a different dynamic-range decision. One recipe pulls a studio portrait toward a magazine's cool house neutral. The other pulls it toward a lamplit drawing room.

This is the same relationship Raking Light has with Dark Ground: a shared genre, a measured inversion on colour, and two recipes that answer different questions about the same room.

Part 5: honest limits

• No frames have been shot with it.
• The corpus is the commissioned studio portraits, which is what passes a painting filter on open-access sources. Sargent's plein-air work measures nothing like this and the variants table says so.
• One drawing passed the automated medium filter and was removed by hand. Any future corpus from this source needs the same check.
• These are reproductions, so the endpoints belong to a photograph of a painting rather than to pigment. Varnish on late-nineteenth-century work is a smaller factor than on the old masters in the group, but it is not zero and it pushes in the warm direction.
• The 0.0% cool figure is a property of the reference and becomes a styling instruction rather than a camera setting. No recipe removes a blue cushion from the frame.
• A reference tells you what a look is and never whether a recipe reproduces it.

Part 6: where it sits in the bank

Second Pro Neg. Hi original after Fashion Editorial, and the fifth from the painting seam.

The seam now spans the full range of the shadow-temperature axis. Signac at b* −8.55 and Caravaggio with highlights at +24.83 over shadows at +0.49 are the two extremes. Monet, Van Gogh and Sargent sit between them. Five recipes from twelve painters, with seven rejected because the bank already covered them or because the corpus could not be measured.

Last updated: 2026-10-05