Case Study 27.2 — Shoot-Along: Grading a Neon Night Block From Accurate to Evocative

"It is more important to click with people than to click the shutter." — attributed to Alfred Eisenstaedt

The brief

Case Study 27.1 was analysis — taking a finished masterpiece apart. This one is production: we are going to shoot a scene specifically chosen for its color problems, develop it from the RAW, and walk the entire chapter's workflow on a real file — correction, HSL, grading, skin protection, and finally the two-grade decision. You will see every choice and what it changed.

We're going to the city block at night — one of the book's four recurring locations — because mixed artificial light is the toughest, most instructive color situation there is. A night street is lit by a chaos of sources: warm sodium streetlamps, cool-white LED storefronts, the colored glow of neon and signage, the green of an old fluorescent tube in a shop window, the blue of a phone screen on a face. Every one of those lights has a different color, and they all land in the same frame. There is no single "correct" white balance, which is exactly why it's the perfect place to learn the difference between correcting a cast and grading a mood. If you can develop a neon night street, you can develop anything.

The constraints (deliberate): - One location, one evening, after blue hour has faded to full dark. - A subject with a person in it, so we have to confront skin under colored light (the chapter's hardest problem). - We commit to the full two-grade deliverable: an accurate version and an evocative version, and a justified choice. - Everything achievable on a phone or a camera — the develop is identical either way.

🔗 Connection: This shoot leans on night exposure from Chapter 21 (long exposure, the city block at night) for the capture, and on the RAW global develop from Chapter 26 for the base. We pick up where those left off and do the color — this chapter's job. We're not re-teaching how to expose a night scene; we're learning how to color one.

Gear and capture settings

You need almost nothing. A camera or phone that shoots RAW (Chapter 25/26), something to brace it on for the slower shutter speeds night demands, and an editor with white-balance, HSL, and shadow/highlight color controls. The capture is the easy part; the develop is the lesson.

⚙️ Settings Box — Capturing the night block (a starting point, not a recipe)

Setting	Start at	Why
Mode	Aperture-priority or Manual	You want control over depth and a steady exposure
File format	RAW	Non-negotiable for color work — JPEG bakes in a white balance you can't undo (Ch.26)
White balance (capture)	Daylight / ~5500K, fixed	Don't let auto WB lie to you frame-to-frame; fix it and correct later. Daylight preserves the real color of each light source
Aperture	f/2.8–f/5.6	Wide enough for hand-holdable shutter; enough depth to keep the subject sharp
Shutter	1/60 s if a person is moving; longer (braced) if the scene is still	Freeze the person; let the lights be lights
ISO	As low as the shutter allows; accept some noise	Noise is fixable (Ch.26); blur from too-slow a shutter on a person is not
Focus	Single-point AF on the subject's near eye	The eye is what the photo is about

The one capture rule that makes the whole develop easier: set white balance to a fixed Daylight value, not Auto. Auto white balance will try to "fix" the warm streetlamps and cool storefronts on every frame, guessing differently each time and fighting you in the edit. A fixed Daylight capture records each light source's true color, leaving every decision for the develop where you can make it deliberately.

The setup

Here's the scene we framed, drawn as a top-down map of the light:

FIGURE CS27.2 — The night-block setup (top-down view)
   ☀ = available light source (here, artificial). ( S ) = subject. [CAM] = camera.

        ▣ NEON SIGN (magenta-pink)            ☀ LED STOREFRONT (cool white/blue)
        (window, behind & camera-right)        (across the street, camera-left)
                    \                              /
                     \                            /
                      ( S ) ──────────────► facing camera, half-turned to the neon
                     /        \
        ☀ SODIUM    /          \   reflected glow
        STREETLAMP (warm orange, from above-left, spills onto the wet pavement)
                    |
                 [ CAM ]  (low, ~chest height, subject filling left third)

   The problem in one sentence: three light sources, three different colors, one face
   caught in all of them. There is NO single "correct" white balance — only choices.

The subject — let's say a friend who agreed to stand on the corner for ten minutes (consent matters; we cover it properly in Chapter 32) — is lit from three directions by three colors: warm sodium from the streetlamp above-left, cool LED from the storefront across the street, and the magenta-pink of a neon sign in the window behind-right rim-lighting their shoulder. The wet pavement (it had rained, which is a gift — always shoot a city block at night after rain) reflects all of it in long colored streaks.

We shot maybe twenty-five frames, working the scene — different framings, the subject looking into the neon and away, waiting for the street to clear. We're going to develop one.

The develop, in phases

This is the heart of the case study. We'll go through the exact chapter workflow on this one file, and at each phase show a Described Photograph of where the image is and say what we changed and why.

Phase 0 — The global develop (Chapter 26, briefly)

Before any color decisions, the Chapter 26 base: we set exposure so the face is properly lit and the brightest signs aren't fully blown (a glance at the histogram confirms the highlights aren't jammed against the right wall), recovered a little shadow detail in the dark pavement, and applied lens corrections and a touch of noise reduction. No color grading yet — this is just getting an honest, well-exposed RAW base. The white balance is still the fixed Daylight value from capture, so the streetlamp reads orange, the storefront reads blue, the neon reads magenta — every light its true color.

FIGURE CS27.3 — "Night block, base develop"        [constructed teaching example]
  THE FRAME    A person on a wet corner, filling the left third, half-turned toward an off-frame neon glow.
               Storefronts recede down the right; colored reflections streak the pavement toward us.
  THE LIGHT    Three sources, true-colored: warm orange streetlamp from above-left on the face's near side;
               cool blue LED from across the street on the far side; magenta neon rim on the shoulder.
  THE COLOR    Honest but chaotic — the face is split warm/cool down the middle, which looks slightly
               unwell; the blues and oranges fight; nothing leads.
  THE EFFECT   A correct but unresolved record. Everything is *there*; nothing is *decided*. This is our
               starting line.

Phase 1 — Correction: tame the chaos without killing the night

Now §27.1. The temptation is to "fix" the white balance to neutral — but neutral is wrong here, because the point of a neon night is its colored light. If we neutralized to grey, we'd throw away the whole reason we came. So correction here is surgical, not global:

• We do not drag the global white balance to neutral. We leave the scene's warm/cool color intact — it's intentional, not a lying cast (the §27.1 distinction).
• We do fix the face, which is the one place the mixed light is doing damage. The cool LED on the far cheek had pushed it slightly green-blue, reading unhealthy. Using HSL (jumping ahead to §27.2 because correction and HSL blur together in practice), we lift the orange/red band's luminance and saturation a touch to bring the skin back toward believable, recovering the far cheek without warming the whole frame.
• We sample a genuinely neutral spot — a grey concrete pillar catching mostly the cool LED — just to check how cool that light really is, then back off, keeping most of the blue because we want it.

FIGURE CS27.4 — "After correction"        [constructed teaching example]
  THE COLOR    The colored lights are all still there (orange lamp, blue storefront, magenta neon) — but the
               FACE now reads as a believable person, the green-blue pallor on the far cheek lifted back to
               skin. The night's color is preserved; only the skin was rescued.
  THE EFFECT   Still chaotic, but now the one thing that *had* to be honest — the human face — is honest. We
               kept the night and fixed the person. This is the "accurate" base we'll grade from.

💡 Why It Works: On a mixed-light night scene, "correction" doesn't mean "make it neutral" — it means "rescue the things that must read true (skin, a known-white sign) while preserving the colored light that's the whole point." This is the §27.1 rule at its sharpest: neutralize casts that lie (the green on the cheek), keep casts that tell the truth (the colored streetlights).

Phase 2 — HSL: make the colors lead instead of fight

Now §27.2 in earnest. The corrected base is honest but the colors compete. We use HSL to build a hierarchy — to decide which colors lead and which recede:

• The magenta neon rim is the most distinctive color and the most "this is a neon night" — we boost the magenta/purple band's saturation slightly and lift its luminance so the rim on the shoulder glows.
• A garish, too-bright green from an old fluorescent tube deep in one storefront was pulling the eye to a dead corner. We drop the green band's saturation and luminance hard — silencing it (the §27.2 "silence the distraction" move) so it recedes into the background where it belongs.
• The blue LED is good but a touch too bright and cyan; we nudge the blue band's hue slightly deeper and drop its luminance a little so the storefronts feel like night, not daylight.

FIGURE CS27.5 — "After HSL"        [constructed teaching example]
  THE COLOR    The magenta neon now leads — the brightest, most saturated color, glowing on the subject's
               shoulder and reflected on the wet street. The distracting green is gone, dropped into the
               dark. The blues are deeper and more nocturnal.
  THE EFFECT   For the first time the image has a *hierarchy*: eye to the lit face, then to the glowing
               magenta rim, then down the colored reflections. The chaos is becoming a composition. Still no
               "grade" — this is all isolation and emphasis.

Phase 3 — Grading: build the mood with shadows and highlights

Now the hinge, §27.3. We name the feeling in one word: "electric loneliness" — the specific mood of one person alone on a glowing, rain-slick city corner at night. That word drives every move.

• Shadows toward deep blue-teal. We drag the shadow grading wheel toward a cool blue-teal. The dark pavement, the recessed doorways, the night sky all pick up a cold, nocturnal blue — "loneliness," "night," "cold."
• Highlights toward warm magenta-amber. We drag the highlight wheel toward a warm magenta-amber, so the brightest things — the neon, the streetlamp glow, the highlights on the wet street — get hotter and more "electric." This deepens the complementary split (cool shadows, warm highlights) that builds depth and drama.
• Midtones: left almost alone. This is the discipline from §27.3 and §27.5 — the face lives in the midtones, so we barely touch them. A whisper of warmth, no more.

FIGURE CS27.6 — "After the grade"        [constructed teaching example]
  THE COLOR    The shadows are now cold blue-teal (the night, the wet dark street); the highlights glow warm
               magenta-amber (neon, lamplight, wet reflections). The complementary split makes the frame feel
               deep and designed. The subject, in the midtones, is the one calm, human-colored thing in a
               glowing, cold, electric night.
  THE EFFECT   "Electric loneliness," exactly. The grade did what the capture couldn't: it turned a correct
               record of a corner into a *feeling* about being alone in a city at night. The color is now
               carrying the emotion.

Phase 4 — Protect the skin (the part everyone gets wrong)

We check the face, hard, because §27.5 says this is where grades go to die. And sure enough: the cool teal shadow grade has crept up into the shaded side of the face, giving the cheek a faint greenish cast again. So we do the skin-protection move:

• Back on the orange/red HSL bands, we lift the skin's luminance and restore a little of its warmth and saturation — just on those bands — recovering the face while leaving the cold teal grade on everything that isn't skin.
• We compare before/after on the face specifically, toggling, until the person reads as a believable human in an unbelievable light.

♿ Accessibility & Inclusion: This step is not optional, and it's not only about realism — it's about respect. A grade that leaves a person's skin looking grey, green, or ashen is a failure no matter how beautiful the rest of the frame is, and the risk is greater for darker complexions, whose richness a heavy cool grade can mute fast. Whoever your subject is, check their face specifically under the grade and protect its natural luminance and warmth. The person is never the place to be careless with color.

FIGURE CS27.7 — "After skin protection"        [constructed teaching example]
  THE COLOR    Identical to CS27.6 everywhere EXCEPT the face, which now reads as warm, living skin against
               the cold electric night — the greenish cast on the shaded cheek rescued, the rest of the
               grade untouched.
  THE EFFECT   Now it works completely: a real, human person, honestly colored, standing inside a graded,
               moody, electric night. The contrast between the believable face and the unbelievable light is
               the whole picture.

Phase 5 — Pull it back, then make the accurate twin

Two final steps, both from §27.6:

1. Pull the grade back ~25%. Our first instinct was, as always, too strong. We reduce the whole grading effect by about a quarter, so it reads as atmosphere rather than filter. The mood survives; the heavy-handedness goes.
2. Make the accurate twin. We go back to the Phase 1 corrected base (face fixed, night preserved, no mood grade) and finish that as a clean, honest version — a faithful record of a colorful night corner. Now we have both: the accurate version and the evocative one.

The two-grade decision

Here they are, side by side:

FIGURE CS27.8 — "Night block," two grades        [constructed teaching example]

  VERSION A — ACCURATE
    The corner as it honestly looked: colored streetlights preserved, face believable, no mood grade. A
    truthful, pleasant, slightly busy record of a rainy neon night. Good for: documenting the place, a news
    or location-scouting context, anyone who needs to know what was actually there.

  VERSION B — EVOCATIVE ("electric loneliness")
    Cold blue-teal shadows, warm magenta-amber highlights, the neon leading, the green silenced, the skin
    protected, the whole grade pulled back to atmosphere. A point of view about being alone in a glowing city
    at night. Good for: a portfolio, an editorial piece, a personal project — anywhere feeling is the value.

We slept on it (the §27.6 rule — and it's real). In the morning, Version B was unmistakably the stronger photograph — it had a point of view the accurate version only documented, and the skin protection meant the mood didn't come at the cost of the person. So B is the keeper for a portfolio. But we kept A, labeled, because the day a client or a publication needs the honest version of that corner, we'll be glad we have it — and because the pair is proof we can do both, on purpose, which is the entire skill this chapter teaches.

⚠️ Common Mistake: Grading a night scene to neutral. The single most common error on mixed-light night photos is "correcting" the white balance until the colored light is gone — turning a magical neon corner into a flat grey street. The colored light is the content. Correct the things that must be true (skin, a known-white sign), and grade with the colored light, never against it.

Troubleshooting the night grade

Mixed-light scenes fail in predictable ways. Here is the field-repair guide — match the symptom to the fix.

The throughline of every fix is the same lesson: on a night street, correct narrowly and grade deliberately. The colored light is your raw material, the face is the one thing that must stay true, and the grade is the mood you build on top — pulled back far enough to feel like the night rather than like an app.

🔄 A quick self-check before you export any night grade: 1. Does every face read as a believable, living person — checked specifically, not assumed? 2. Did the colored light survive correction (it should), and does one color clearly lead? 3. Is any channel clipped on the histogram (back it off if so)? 4. If you pulled the grade back 30%, would the image be better? (Usually yes — do it.)

Discussion questions

1. We deliberately captured with a fixed Daylight white balance instead of Auto. Explain exactly why that made the develop easier, in terms of what Auto would have done to a series of mixed-light frames.
2. In Phase 1 we refused to neutralize the global white balance but did fix the face. Articulate the rule that distinguishes the casts you removed from the casts you kept.
3. The midtones were "left almost alone" through the entire grade. Trace why that decision protected the image, and connect it to where skin lives in the tonal range.
4. We chose Version B (evocative) as the portfolio keeper but kept Version A. Describe a specific situation where the exact same file should be delivered as Version A instead, and why.
5. The wet pavement was called "a gift." Why does rain make a neon night easier and more rewarding to grade?

Your turn (extension)

Run this exact shoot-along on your own city block at night.

1. Capture: find a corner with at least two differently-colored light sources (warm lamp + cool storefront, or any neon). Shoot RAW, fixed Daylight white balance, braced, with a person if you can (get their consent). Shoot twenty-plus frames; work the scene.
2. Develop, in the chapter's order: global base (Ch.26) → correct the face, keep the night → HSL to build a color hierarchy (lead with one color, silence one distraction) → name a feeling-word → shadow and highlight grade toward a complementary split → protect the skin with HSL → pull the grade back.
3. Make the accurate twin from the corrected base.
4. Sleep on the pair, then choose the keeper and write one sentence justifying it by the photo's purpose.

Submit both versions and your feeling-word. If your skin reads true under a moody grade and your colored lights survived correction, you have learned the whole chapter — on the hardest possible material.

Key takeaways

• A mixed-light night scene is the ultimate color-correction-and-grading exercise: many colored sources, no single "correct" white balance, and a face caught in all of it.
• Capture with a fixed Daylight white balance (RAW), so each light source records its true color and every decision is left for the develop — Auto WB only fights you across a series.
• Correction ≠ neutralization here. Rescue what must be true (skin, a known white), but preserve the colored light that is the whole content of a neon night.
• Build a color hierarchy with HSL (lead with one color, silence the distractions) before grading, then grade shadows/highlights toward a complementary split for mood, keeping midtones — and skin — nearly untouched.
• Protect the face specifically under any grade (especially a cool one), recovering skin on the orange/red bands while leaving the grade on everything else — a respect issue as much as a realism one.
• Deliver the two grades, sleep on them, and choose by purpose: the evocative version for a portfolio or editorial; the accurate version kept for whenever truth is the job.