Appendix G: Color Management Reference

The chapters taught you to see and steer color: hue, saturation, and value as the three dimensions of any color (Chapter 7); white balance, HSL, grading, and the saturation-versus-vibrance discipline (Chapter 27); and RAW as the file that keeps color decisions adjustable (Chapter 2, developed in Chapter 26). This appendix is the plumbing underneath all of that — the part most photographers never learn and then blame the printer, the phone, or "the colors looking wrong online."

Color management is the boring, invisible system that decides whether the red you graded at your desk is the same red on someone else's screen, in a print, or on a phone in daylight. Get it wrong and a portfolio you loved goes muddy on Instagram or comes back from the lab too dark and too magenta. Get it right and it disappears — which is the whole goal. This is reference material: skim it once to understand the ideas, then come back to the decision table in §G.7 when you actually have to choose an export setting.

None of this changes the central promise of the book. The eye is still the instrument. But when you already know what color you want, this is how you make it survive the trip from your file to someone else's eyes.

G.1 The one idea that makes the rest make sense

A digital color is just a set of numbers (red, green, blue values per pixel — Chapter 2). Numbers are meaningless until something tells you which red those numbers mean. "R: 220" is not a color until you know the rulebook being used to turn it into light.

A color space (or color profile) is that rulebook: a defined relationship between the numbers in a file and the actual colors they represent. The same RGB numbers point to different real colors in different color spaces. Tag the file with the wrong space — or strip the tag and let a device guess — and the colors shift.

FIGURE G.1 — Why the same number is two different colors

   THE FILE SAYS:  R 220   G 30   B 40   (just three numbers)

        │ interpreted in sRGB              │ interpreted in Adobe RGB
        ▼                                  ▼
   ┌──────────────┐                   ┌──────────────┐
   │ a normal,    │                   │ a deeper,     │
   │ web-safe red │                   │ more saturated│
   └──────────────┘                   │ red           │
                                      └──────────────┘
   Same numbers. Different rulebook. Different red.
   This is why an "untagged" image looks wrong on the web: the browser
   assumes sRGB, but the file's numbers meant something else.

Everything below is just two questions asked over and over: which rulebook is this file using, and does the next device in the chain know it? Color management is the discipline of keeping that answer consistent from capture to screen to print.

🚪 Threshold Concept: A color in a file is not a color. It is a number plus a rulebook (the profile). Once you internalize that an RGB value is meaningless without its color space, every "the colors changed" mystery becomes solvable: somewhere, a rulebook was lost, swapped, or ignored. Manage the rulebook and you manage the color.

G.2 The color spaces, and when each one matters

A gamut is the range of colors a given space (or device) can actually represent — its color vocabulary. A wider gamut can describe more saturated, more extreme colors; a narrower one can't say those words at all. The common working spaces differ mainly in how wide their gamut is.

FIGURE G.2 — Working-space gamuts, smallest to largest (schematic)

   all colors the eye can see  ████████████████████████████████  (huge)
   ProPhoto RGB                ██████████████████████████░░░░░░  (very wide)
   Adobe RGB (1998)            ████████████████░░░░░░░░░░░░░░░░  (wide)
   sRGB                        ███████████░░░░░░░░░░░░░░░░░░░░░░  (standard)
   a typical phone/laptop      ███████████░░░░░░░░░░░░░░░░░░░░░░  (≈ sRGB)

   Wider gamut = can DESCRIBE more saturated colors. It does NOT make a
   photo more colorful by itself — it only matters if your colors, your
   editing, your screen, AND your output can all use that extra range.

The plain-language version. Edit wide, deliver narrow. Many photographers do their RAW developing in a wide space (Adobe RGB, or ProPhoto for heavy work) so no color is thrown away while they push sliders, then convert a copy to sRGB on export for anything going to a screen, and keep Adobe RGB only for prints whose lab explicitly wants it. The wide space is the workshop; sRGB is the shipping box for the web.

📝 A note on ACES. ACES (the Academy Color Encoding System) is a wide, standardized color framework developed in the motion-picture industry (by the Academy of Motion Picture Arts and Sciences) to keep color consistent across many cameras, software tools, and displays in a film pipeline. You will hear it in video and VFX work. For still photography it is almost always more machinery than you need — the sRGB / Adobe RGB / ProPhoto trio above covers stills. Know the name so it doesn't intimidate you; reach for it only if a video or color-pipeline workflow specifically requires it.

⚠️ Common Mistake: Exporting Adobe RGB or ProPhoto for the web. You grade a gorgeous file in a wide space, export it still tagged Adobe RGB, post it — and on every phone that ignores the tag it looks washed out and grey, because the browser/app reads your wide-gamut numbers as if they were sRGB. The fix is always the same: convert to sRGB before anything goes online. Wider is not "better" for delivery; it's only better if every link in the chain understands it, and the web mostly doesn't.

🎒 Gear Note: "Wide gamut" is sold as a camera and monitor feature, and it is genuinely useful for print work — but it is not a shortcut to better photographs. A wider gamut only helps if your subject actually contains those extreme colors, your screen can show them, and your output can reproduce them. A phone shooting and posting to social media lives almost entirely in sRGB, and that is completely fine — most images the world loves were delivered in sRGB. Match the space to the destination, not to the spec sheet.

G.3 Bit depth: 8-bit vs 16-bit, and where banding comes from

Gamut is how wide the range of colors is. Bit depth is how many steps sit between the darkest and lightest version of each color — the smoothness of the gradient.

Bit depth is the number of distinct tonal steps recorded per color channel. 8-bit gives 256 steps per channel; 16-bit gives many thousands. More steps mean smoother transitions and far more room to push tones in editing before the gradient breaks.

FIGURE G.3 — Why a smooth sky bands (8-bit vs 16-bit)

   A GENTLE SKY GRADIENT, edited hard (big exposure / contrast moves):

   8-bit  (256 steps)   ░░░▒▒▒▓▓▓███   ← only 256 stops to spread across the
                        │  │  │  │      gradient; after heavy edits the steps
                        └──┴──┴──┘      become VISIBLE as bands/stripes = BANDING

   16-bit (thousands)   ░░░░▒▒▒▒▓▓▓▓███ ← thousands of steps; the same heavy edit
                        (no visible seams) still reads as a smooth, seamless gradient

Banding (also posterization) is the visible stepping of what should be a smooth gradient — most obvious in clear skies, soft studio backdrops, and out-of-focus washes — caused by too few tonal steps surviving the edit. 8-bit files banding under heavy adjustment is the usual culprit.

The practical rules:

• Edit RAW files in 16-bit. RAW captures more tonal information than 8-bit can hold (Chapter 26); developing in 16-bit preserves it so big exposure, contrast, and grading moves don't tear the gradients.
• ProPhoto RGB must be 16-bit. A gamut that wide spread over only 256 steps will band; the two go together (wide space + high bit depth) or not at all.
• 8-bit is fine for delivery. JPEG and PNG-for-web are 8-bit, and a finished image rarely bands once you've stopped editing it. Do the heavy lifting in 16-bit, then export 8-bit sRGB JPEG for the web. The banding risk is in editing 8-bit, not in viewing it.
• If you see banding in a sky, it usually means an 8-bit file was pushed too hard. Re-develop from the RAW in 16-bit; a tiny amount of added grain/noise can also mask mild banding.

🔬 The Physics: (Optional — skip without penalty.) Why 256 steps for 8-bit? Each channel stores 8 binary digits, and $2^8 = 256$ possible values (0–255) per channel; 16-bit stores $2^{16} = 65{,}536$ possible values. Your eye can distinguish only a couple hundred brightness steps in a single smooth gradient at viewing time, which is why a finished 8-bit image looks smooth. The trouble is editing: when you stretch tones (raise exposure, add contrast), you spread those few steps further apart, and gaps open between them that the eye now resolves as bands. 16-bit keeps thousands of intermediate steps in reserve, so even after a hard stretch the survivors are still close enough together to read as continuous. You edit in 16-bit for headroom, then deliver 8-bit because the result needs far fewer steps than the process did.

G.4 White balance and color temperature in practice

You met white balance as a creative warm/cool lever (Chapter 7, §7.5) and as the first correction step (Chapter 27, §27.1). Here is the quick-reference version of the numbers behind it.

Color temperature is measured in kelvin (K) and runs counterintuitively: warmer-looking light has the lower number. (Consistent with the glossary and Chapter 5.)

FIGURE G.4 — The kelvin scale (warmer light = LOWER number)

   WARM / ORANGE  ◄──────────────────────────────────────►  COOL / BLUE
   low K                                                     high K

   ~1800 K     ~2700–3000 K     ~5200–5500 K     ~6500 K        ~7000–9000 K
   candle      tungsten bulb    midday sun       overcast       open shade
   firelight   warm interior    "daylight"       cloudy sky     blue-hour shade

   Your camera's WB setting does the OPPOSITE of the light to cancel it:
   under warm (low-K) light it adds blue; under cool (high-K) light it adds warmth.

⚙️ Settings Box — White balance, fast.

Situation	Start at	Phone-equivalent path
Color-critical work (product, skin, repro)	Shoot RAW, set Custom off a grey card	Pro/Manual mode → WB → "Custom"/tap a grey card; or shoot RAW (ProRAW/Expert RAW) and set WB after
Keep a golden-hour glow	Cloudy or Shade (don't let Auto cool it)	Pro mode → WB → Cloudy/Shade (or drag the Temp slider warm)
Keep a blue-hour / night mood	Tungsten, or dial a low Kelvin	Pro mode → WB → Incandescent (or drag Temp slider cool)
Unsure / changing light	Auto, but shoot RAW as the safety net	Default camera app is Auto; switch on RAW so WB stays editable

Phones: most stock camera apps hide WB inside a Pro / Manual / Expert mode as a "Temperature" or "WB" slider; many flagships also shoot a RAW (often labeled ProRAW or Expert RAW) that lets you set white balance losslessly afterward, exactly like a camera RAW.

The grey-card move (the reliable one). Photograph a grey card held in your subject's light at the start of a shoot (Chapter 27, §27.1). In the develop, click the white-balance eyedropper on the card, then copy that white balance to every frame shot in the same light. Now the whole set is neutral, objectively, with no guessing. A folded white coffee filter or the back of a white card works in a pinch.

⚠️ Common Mistake: Trusting Auto white balance in mixed light. Warm lamps plus a cool window plus a flash give the camera three different color temperatures at once, and Auto can apply only one global correction — so it splits the difference and gets all of them slightly wrong (faces orange, windows cyan). Decide which light is on your subject's face, set WB for that, and let the rest fall where it may — or shoot RAW and resolve it later (see the Mixed lighting entry in the glossary, Chapter 5).

G.5 Gamut and "out of gamut": what it means and what to do

A color is out of gamut when it is more saturated or extreme than the target space (or the printer, or the screen) can reproduce. The number exists in your wide working space, but the destination has no way to say that color. Something has to give.

FIGURE G.5 — Out-of-gamut: a color the destination can't say

   YOUR EDIT (wide space):   ● a blazing, ultra-saturated orange flower

   SENT TO sRGB / a printer:   that exact orange is OUTSIDE the box
                               ┌───────────────┐
                               │   sRGB / ink   │   ● ← actual color sits
                               │     gamut      │       just OUTSIDE the edge
                               └───────────────┘
   The system must MAP it back inside. How it maps = "rendering intent":

   • PERCEPTUAL    → squeezes ALL colors inward a little to keep RELATIONSHIPS
                     smooth (best for photos with many saturated colors)
   • RELATIVE      → keeps in-gamut colors EXACT, clips only the out-of-gamut
     COLORIMETRIC    ones to the nearest edge (best when most colors already fit)

What you actually do about it:

• Soft-proof before you print. Most editors have a soft-proofing view that simulates how your image will look in a chosen output profile (a specific printer/paper, or sRGB) and can flag out-of-gamut colors. This is the single most useful color-management habit for anyone who prints.
• Pick a rendering intent. For photographs, Perceptual usually looks most natural because it preserves the relationships between colors as it pulls them in; Relative Colorimetric is good when only a few colors are out of gamut and you want the rest left untouched. If unsure, compare both in soft-proof and trust your eye.
• Pull the saturation back yourself. If a saturated red or blue is clipping (losing its inner detail and going to a flat block — the saturation-limit tells from Chapter 27, §27.5), gently lower that color's saturation in HSL until it sits inside the destination's range. Controlled by you beats clipped by the machine.
• Don't panic over screen warnings. A color being out of your monitor's gamut doesn't mean it's wrong in the file — it means your screen can't show it. That is a display limit, not a file error. It matters for judging, which is why calibration (next) exists.

G.6 Monitor calibration: trusting what you see

Every screen has its own bias — too blue, too bright, too contrasty — straight out of the box. If you grade to make an image look right on a screen that runs warm, it will look cool everywhere else. Monitor calibration is how you remove your screen's personal opinion so the colors you edit are the colors actually in the file.

Monitor calibration is measuring and correcting a display so it shows color accurately — neutral greys, correct brightness, and predictable hues — rather than the manufacturer's default bias. Done properly it uses a hardware colorimeter (a small sensor you hang on the screen) and software that builds a corrective display profile.

FIGURE G.6 — Calibration removes the screen's opinion

   BEFORE                          AFTER
   ┌──────────────┐                ┌──────────────┐
   │ screen runs  │                │ neutral grey, │
   │ a bit BLUE & │   colorimeter  │ correct       │
   │ too BRIGHT   │ ───measures──► │ brightness,   │
   │ (its bias)   │   & corrects   │ honest color  │
   └──────────────┘                └──────────────┘
   You stop editing AGAINST your screen's lie. What you see ≈ what's in the file
   ≈ what a calibrated lab will print. The three finally agree.

Low-cost habits that help even without hardware:

• Edit in consistent, neutral room light. Bright, colored, or changing ambient light fools your eye far more than people expect. A dim, neutral room beats a sunlit one for judging color.
• Don't edit at full screen brightness. Most laptops ship far too bright, which makes you under-expose and over-darken your edits; they then look muddy on other screens. Bring brightness to a moderate level.
• Check on a second device. Your phone and your laptop disagreeing tells you something is off even if you can't afford to fix the cause. The second screen is a free reality check.
• Use a known-neutral reference. When your adapting eyes have "fixed" a cast for you (Chapter 27, §27.1), glancing at a truly neutral grey resets them.

🎒 Gear Note: A hardware colorimeter is a real, worthwhile purchase once you print or deliver color professionally — it's the difference between guessing and knowing. Below that threshold, your money is better spent on a good neutral viewing environment and the habit of checking on a second device. As with every gear note in this book: buy the tool when the work demands it, not because the spec sheet flatters you. Phone and tablet screens generally cannot be hardware-calibrated, which is exactly why they make good "second-device" sanity checks but poor primary grading displays for print work.

G.7 Choosing an export color profile: web vs print vs social

This is the table to come back to. When you finish an image and hit export, two settings carry all the color risk: the color space (profile) and the bit depth / format. Match them to the destination.

FIGURE G.7 — The export decision (where is this image going?)

                    WHERE IS THIS IMAGE GOING?
                              │
        ┌─────────────────────┼─────────────────────┐
        ▼                     ▼                     ▼
   WEB / SOCIAL          PROFESSIONAL PRINT      KEEP / RE-EDIT
   ───────────          ──────────────────      ──────────────
   sRGB                 ask the LAB:             Keep the RAW +
   8-bit JPEG           usually Adobe RGB        a 16-bit master
   (embed the profile)  16-bit TIFF if they ask  (ProPhoto/Adobe RGB)
   sRGB if they say so  embed the profile        Export delivery copies
                                                 FROM the master as needed

The four rules that prevent most color disasters:

1. For the web and social, export sRGB. Every time. This single rule prevents the most common "why does it look washed out online" failure.
2. For professional print, ask the lab what they want and give them exactly that — usually Adobe RGB, often 16-bit TIFF, profile embedded.
3. Embed the profile on export when the option exists. An embedded profile lets a smart app interpret your numbers correctly; a stripped file forces every app to guess (and most guess sRGB).
4. Keep a 16-bit master and the RAW. Export copies to whatever space each destination needs. Convert wide → narrow on the way out; never edit your only file down to 8-bit sRGB and lose the headroom.

💡 Why It Works: Why "edit wide, deliver narrow"? Because editing is where you push color — and pushing benefits from every extra step (bit depth) and every extra color (gamut) you can keep in reserve, so gradients stay smooth and saturated tones don't clip prematurely. Delivery is the opposite problem: the destination has a fixed, usually narrow vocabulary, and your job is to translate cleanly into the words it actually knows. A wide master keeps your options open; a correctly-targeted export makes sure the viewer sees what you meant. One file optimized for the work, another optimized for the trip.

♿ Accessibility & Inclusion: Color management is also an accessibility concern. First, the same export that "looks fine" to you may shift on the cheap, uncalibrated screens many viewers actually use — which is another reason to deliver in safe, universally-understood sRGB rather than betting on wide-gamut support. Second, do not let the meaning of an image ride on hue alone: roughly 1 in 12 men has some red-green color-vision deficiency (consistent with Chapters 7 and 27), so back any color-coded meaning with differences in value or shape too, and name the colors explicitly in your alt text and image descriptions so the message survives for someone who can't perceive the distinction or is using a screen reader. Good color management and inclusive color are the same instinct: don't assume everyone's screen — or everyone's eyes — see exactly what yours do.

G.8 Quick glossary for this appendix

A compact recap of the terms above (fuller definitions live in the book's glossary):

• Color space / color profile — the rulebook mapping a file's numbers to actual colors (sRGB, Adobe RGB, ProPhoto RGB, ACES).
• Gamut — the range of colors a space or device can represent; a wider gamut can describe more extreme, saturated colors.
• Out of gamut — a color too saturated/extreme for the target space, screen, or printer to reproduce; must be mapped inward (via a rendering intent).
• Rendering intent — how out-of-gamut colors are pulled into range: Perceptual (keeps relationships smooth) or Relative Colorimetric (keeps in-gamut colors exact, clips the rest).
• Bit depth — tonal steps per channel; 8-bit = 256 steps (fine for delivery), 16-bit = thousands (use while editing).
• Banding / posterization — visible stepping in what should be a smooth gradient; from too few tonal steps surviving heavy edits.
• Color temperature (kelvin, K) — the warm-to-cool color of light; counterintuitively, warmer light has a lower K number.
• White balance — the global warm/cool correction that neutralizes (or deliberately keeps) the color of the light.
• Monitor calibration — measuring and correcting a display (ideally with a hardware colorimeter) so it shows color accurately.
• Soft-proofing — previewing how an image will look in a specific output profile (printer/paper or sRGB) before exporting or printing.
• Embed (a profile) — saving the color-space tag inside the exported file so other apps interpret its numbers correctly.

G.9 Sources and where this came from

The mechanisms here are reasoned from the chapters, especially Chapter 2 (RAW vs. JPEG and how a sensor records color), Chapter 7 (hue/saturation/value, white balance as a creative choice), Chapter 26 (developing RAW), and Chapter 27 (color correction, HSL, grading, vibrance vs. saturation, the grey card). The color spaces named — sRGB, Adobe RGB (1998), ProPhoto RGB — and ACES are established, publicly documented color standards; the kelvin values for common light sources match the book's glossary and Chapter 5. For deeper, hands-on workflow guidance, the further-reading sections of Chapters 26 and 27, and the foundational craft references cited throughout the book, are the right next step. Where a precise lab requirement, a specific monitor model, or an exact device gamut matters for your work, confirm it with the lab or manufacturer directly rather than relying on a remembered number — color management is a place where the specific destination always wins over a general rule.