Colour management and print quality control: densitometry, spectrophotometry and ICC profiles

In short

Colour management is the measured, repeatable control of colour from original file to printed sheet; print quality control is the set of instrument-based checks that prove a press is printing to aim. A densitometer measures how much light an ink film absorbs through broad filters and answers "how much ink"; a spectrophotometer measures reflectance every 10 nm and answers "what colour, as the eye sees it"; an ICC profile is the file that tells software how a particular scanner, monitor, press or proofer's numbers map to real colour. The subject is examined in Printing/Packaging Technology papers as process-control numericals (density, dot gain, print contrast, trapping, ΔE) and theory on ISO 12647 and ICC architecture, and is asked in every pressroom, prepress and quality interview — Manipal Technologies, Thomson Press, Pragati Offset, Parksons, TCPL, ITC PSPD, Huhtamaki, Avery Dennison, Repro India and brand-owner packaging QA teams. On the job it decides whether a pharma carton or FMCG laminate is accepted or rejected against the brand's colour tolerance.

Explained

Densitometry: measuring ink, not colour

• Optical density is the logarithm of the opacity of a sample; it compresses a huge reflectance range into roughly 0 to 2.5 and tracks ink film thickness almost linearly over the working range.
• Reflection density is used for prints, transmission density for films, plates and negatives. Definitions and geometry: ISO 5-3 (spectral conditions — it defines the status responses T, E, A, I) and ISO 5-4 (geometric conditions for reflection density, 0/45).
• Construction: stabilised lamp or LED, 0°/45° illumination-viewing geometry, defined aperture (commonly 1.5-3.5 mm), filter wheel or set (red, green, blue, visual) and a photodetector with log amplifier. Zeroed on the unprinted substrate (paper-relative) or on absolute white.
• Filter-ink pairing: cyan is read through the red filter, magenta through green, yellow through blue, black through visual. A filter reads the ink that absorbs its band.
• Status (response) types: Status T — wide-band, standard in North America and in most Indian pressrooms following US practice; Status E — European/DIN 16536 narrow-band blue, gives noticeably higher yellow density for the same ink film; Status A for photographic, Status I narrow-band. Densities from different statuses are not comparable — always state the status.
• Polarisation filter: a cross-polarised densitometer suppresses the surface-gloss difference between wet and dry ink, so a wet sheet at the press reads close to the dry sheet. It changes the absolute numbers, so wet and dry aims must come from the same instrument setting.
• What density cannot do: it says nothing about hue. Two inks of different pigment can give the same density and look obviously different, so density controls consistency, colorimetry controls appearance.

Densitometric process-control parameters

• Solid ink density (SID): the aim for the solid patch. Typical house aims on coated paper, Status T, polarised: C ≈ 1.40-1.50, M ≈ 1.40-1.50, Y ≈ 1.00-1.10, K ≈ 1.75-1.90; lower on uncoated and much lower on newsprint. ISO 12647-2 itself specifies CIELAB values, not densities — density aims are derived in-house to hit those Lab values with the ink and paper in use.
• Tone value / dot area: computed from a tint patch by the Murray-Davies equation; the Yule-Nielsen form adds an *n*-factor for optical (light-scattering) spread.
• Tone value increase (TVI, "dot gain"): measured tone value minus the nominal file value; it is mechanical (plate, blanket, impression, pressure, ink) plus optical (scatter in the substrate). Peaks near the mid-tone, where the dot perimeter is longest. Negative gain in highlights usually means plate or exposure error.
• Print contrast (K): compares the solid with the three-quarter tone; it rises with ink film until shadow dots fill in, then falls — so it finds the optimum inking point better than density alone.
• Ink trapping: how well a second ink lays on top of a first wet ink. Poor trapping gives weak reds, greens and blues even when all solids are on aim; controlled by ink tack sequence (descending tack), press speed and fountain solution.
• Hue error and greyness (Preucil): from the three filter densities of one ink — hue error shows how far the pigment is from an ideal subtractive primary, greyness shows unwanted absorption/contamination. Plotted on the GATF colour circle.
• Other checks: doubling and slur from a slur target, register from register marks, ink–water balance from scumming/tint, mottle, set-off, and a visual check against an OK sheet approved in ISO 3664 lighting.

Spectrophotometry and colorimetry

• A spectrophotometer disperses the reflected light (diffraction grating onto a photodiode array) and reports spectral reflectance typically 380-730 nm at 10 nm intervals — the full fingerprint of the colour, independent of any filter.
• From reflectance plus a standard illuminant (printing uses D50) and a standard observer (2° or 10°, CIE 1931/1964) you compute tristimulus XYZ, then CIELAB (L*, a*, b*), chroma C\* and hue angle h°. CIELAB is a near-uniform, device-independent space; CIE 1976 also defines CIELUV, mainly for emissive displays.
• Measurement conditions, ISO 13655: M0 (incandescent, legacy A-type source), M1 (D50-equivalent including UV — the condition specified for current graphic-arts work), M2 (UV-cut), M3 (polarised). Needed because optical brightening agents (OBAs) in paper fluoresce under UV: the same sheet can read several ΔE apart on M0, M1 and M2.
• Backing: measure with the specified black or white backing; the choice changes L* on thin stock. Aperture, number of averages and warm-up all matter.
• Colour difference: ΔE\*ab (1976) is simple but perceptually uneven, especially in saturated blues and in near-neutrals; ΔE\*94 adds lightness, chroma and hue weighting, and ΔE00 (CIEDE2000) adds revised weighting functions plus a blue hue-rotation term; these are what current standards and brand specifications use. Rule of thumb: ΔE00 ≈ 1 is a just-noticeable difference on a side-by-side comparison, 2-3 is a commercially acceptable production tolerance.
• Metamerism: two samples matching under D50 but differing under store lighting — detected only spectrally, never by a densitometer. Common when a spot colour is simulated in CMYK.
• Spectrodensitometer: one instrument that reports both spectral/Lab data and any density status; scanning versions (press-console strip readers, hand-held scanning spectros) read a whole control strip in seconds and feed closed-loop ink key control.
• Spectral data exchange: CxF/X-4 (ISO 17972-4) carries spectral values of spot colours and substrates in the job file — now routine in packaging specifications.

ICC colour management architecture

• Why: every device has its own gamut and its own relationship between numbers and colour. Managing n devices pairwise needs n(n−1) transforms; routing everything through one reference space needs only 2n profiles.
• PCS (Profile Connection Space): the hub — CIELAB or CIEXYZ at D50, 2° observer. Source profile converts device values → PCS; destination profile converts PCS → device values.
• The four parts of a managed workflow: PCS, profiles, a CMM (colour management module / engine — Adobe ACE, Apple ColorSync, LittleCMS, Heidelberg/GMG engines), and rendering intents.
• Profile classes: input (scanner, camera), display, output (press, proofer, digital printer), DeviceLink (one baked source-to-destination transform, used for ink-saving and press-to-press matching), abstract (creative edit in PCS), named-colour (spot libraries), colour-space conversion.
• Internals: matrix/TRC profiles (small, for RGB displays and spaces like sRGB, Adobe RGB, ProPhoto RGB) versus LUT-based profiles with A2B0/A2B1/A2B2 (device→PCS) and B2A0/B2A1/B2A2 (PCS→device) tables for CMYK. Specification: ICC.1:2010 = ISO 15076-1; v4 improved perceptual-intent definition (PRMG) over v2; iccMAX (ICC.2) supports spectral and non-D50 workflows.
• Rendering intents: Perceptual — compresses the whole gamut smoothly, best for photographs; Media-relative colorimetric — exact for in-gamut colours, clips out-of-gamut, white mapped to paper white, best for logos and spot colours; Saturation — keeps vividness, for business graphics; ICC-absolute colorimetric — simulates the source paper white, used for proofing. Black point compensation maps source black to destination black to stop blocked shadows.
• Separation controls in an output profile: total area coverage (TAC/TIL) — commonly about 300-330% on coated offset, 260-300% uncoated, 240-260% newsprint, lower still on some flexo/laminate jobs; UCR (removes CMY only in neutral shadows), GCR (replaces grey component with black throughout — steadier greys, less ink, better for packaging), black start, black width, maximum black.
• Building an output profile, step by step: 1) stabilise and clean the press/printer; 2) linearise and set ink limits per channel; 3) print a characterisation chart (IT8.7/4, ECI 2002, TC1617) with no colour management; 4) dry, then measure with a scanning spectrophotometer under M1; 5) average several sheets, build the profile with chosen TAC/GCR; 6) verify with an independent test chart and a control wedge; 7) iterate with a correction curve if TVI or greys are off.
• Monitor side: hardware calibration to D50 or D65 white, ~80-120 cd/m², gamma 2.2, in a dim neutral room; soft-proofing uses the press profile with absolute-colorimetric simulation plus paper-white simulation.
• Reference datasets such as Fogra 39/51/52 (and the ISO 15339 CRPC set) are characterisation data, from which standard profiles (e.g. ISO Coated v2, PSO Coated v3) are built — a profile is not the standard itself.

Standards, certification and Indian practice

• ISO 12647 family (process control for the production of half-tone colour separations, proof and production prints): Part 1 parameters and definitions; Part 2 offset lithography; Part 3 newsprint coldset; Part 4 publication gravure; Part 5 screen printing; Part 6 flexography; Part 7 off-press digital proofing (with its mandatory control wedge and ΔE pass/fail criteria); Part 8 validation print. Each part fixes substrate Lab, primary and overprint solid Lab, TVI curves and tolerances. Tolerances migrated from ΔE\*ab in older editions to ΔE00 in current revisions — always work to the edition named in the contract.
• ISO 13655 measurement and computation; ISO 3664:2009 viewing conditions — P1 critical appraisal at 2,000 lx ± D50 and P2 practical appraisal at 500 lx, with neutral grey surround and controlled UV content; ISO 15339 / CRPC characterised reference printing conditions; ISO 12640 digital data exchange; ISO 15790 certification/test targets.
• G7 (Idealliance) is a calibration *method*: match grey balance (NPDC — neutral print density curve) and substrate-relative grey rather than per-ink TVI, so different processes visually agree. G7 Master and Fogra PSO/PSD certifications are voluntary third-party schemes; several Indian printers (Pragati Offset and other large commercial and packaging houses) hold them and advertise them to export clients.
• Quality systems: ISO 9001:2026 is the current edition of the quality management standard; packaging plants additionally run BRCGS/FSSC-style audits for food and pharma cartons. BIS has adopted many graphic-technology and paper test standards through its sectional committees — cite an IS number only after checking its current edition, since several older printing and paper IS numbers have been revised or withdrawn.
• Where law touches print quality: the Legal Metrology (Packaged Commodities) Rules require declarations to be legible and in contrast with the background; drug and food labelling rules (CDSCO/FSSAI) fix minimum type sizes and specific colour marks — the green/brown vegetarian–non-vegetarian symbol being the clearest case where a measured colour is a legal matter, not just a brand preference. Pricing of colour-management instruments is in rupees and attracts GST at the applicable slab (goods are in the 5%/18%/40% structure in force since 22 September 2025).

Key points

• Optical density - log of the reciprocal of reflectance; tracks ink film thickness, not appearance.
• Status T / Status E - wide-band (US practice) and narrow-band DIN (European) densitometer responses; yellow readings differ most.
• Polarisation - cross-polarised filters make wet press sheets read like dry sheets.
• Murray-Davies - converts tint and solid densities into tone value; Yule-Nielsen adds an *n*-factor for optical dot gain.
• TVI - measured minus nominal tone value; largest near 50%, the single most-controlled press variable after SID.
• Print contrast - (Ds − D75)/Ds; finds the inking level that still holds shadow detail.
• Trapping - Preucil ratio of the second ink's density in the overprint to its density printed alone.
• Hue error / greyness - (M − L)/(H − L) and L/H from one ink's three filter densities; pigment quality indicators.
• Spectrophotometer - reflectance 380-730 nm at 10 nm; the only instrument that detects metamerism.
• CIELAB - device-independent L\*a\*b\* at D50, 2° observer; basis of all print colour tolerances.
• ΔE00 (CIEDE2000) - current perceptually weighted colour difference; ΔE\*ab 1976 overstates error in saturated blues.
• M0/M1/M2/M3 - ISO 13655 measurement conditions; M1 is the D50-with-UV condition needed for OBA-loaded papers.
• PCS - CIELAB/CIEXYZ at D50, the hub of ICC conversions; 2n profiles replace n(n−1) direct links.
• Rendering intents - perceptual, media-relative colorimetric, saturation, ICC-absolute (proofing).
• GCR/UCR and TAC - separation controls that trade ink cost, grey stability and drying against each other.
• ISO 12647-2 / -6 / -7 - aim values and tolerances for offset, flexo and digital proofing; ISO 3664 for viewing.
• G7 - grey-balance-based calibration method; a voluntary certification, not an ISO standard.
• Density controls, colorimetry qualifies - density for minute-to-minute running, Lab/ΔE for acceptance.

Formulas and examples

Example 1 — density from reflectance. A solid cyan patch reflects 3.5% of the incident red light. R = 0.035 → D = log₁₀(1/0.035) = log₁₀(28.57) = 1.46. If the film thickness is halved and the ideal Beer-Lambert relation held, D would fall to 0.73 and R would rise to 10^−0.73 = 18.6% — in practice the fall is less because of surface reflection.

Example 2 — tone value and TVI (Murray-Davies). Nominal 50% cyan tint: D_t = 0.45, solid D_s = 1.45 (both paper-relative, red filter, Status T).

• 10^−0.45 = 0.3548 → numerator = 1 − 0.3548 = 0.6452
• 10^−1.45 = 0.0355 → denominator = 1 − 0.0355 = 0.9645
• a = 0.6452 / 0.9645 = 0.669 → 66.9%
• TVI = 66.9 − 50 = +16.9 percentage points. That sits at the top of the usual mid-tone aim band for coated sheetfed offset (ISO 12647-2 mid-tone aims are in the region of 13-17% depending on paper type, screen ruling and the edition's curve) — acceptable, but the plate curve should be checked before it drifts further.
• With a Yule-Nielsen n = 1.8: numerator = 1 − 10^−0.25 = 0.4377; denominator = 1 − 10^−0.8056 = 0.8435; a = 51.9% — showing that most of the apparent 17% gain here is optical, not mechanical.

Example 3 — print contrast. D_s = 1.45, 75% patch D_75 = 1.12. K = (1.45 − 1.12)/1.45 = 0.33/1.45 = 0.228 → 22.8%. If inking is pushed to D_s = 1.60 and the 75% patch fills in to D_75 = 1.32, K = 0.28/1.60 = 17.5% — density went up but shadow separation got worse, so the first condition is the better inking point.

Example 4 — ink trapping (red overprint, yellow down first, magenta second). Read through the green filter (magenta's filter): overprint D_op = 1.25, yellow alone D_1 = 0.12, magenta alone D_2 = 1.35. T = (1.25 − 0.12)/1.35 × 100 = 1.13/1.35 × 100 = 83.7%. Below about 70% the red looks dull and dirty; fixes are lower tack on the second ink, reduced press speed, or resequencing.

Example 5 — hue error, greyness and ΔE. A magenta solid reads D_green = 1.40 (H), D_blue = 0.55 (M), D_red = 0.12 (L).

• Hue error = (0.55 − 0.12)/(1.40 − 0.12) × 100 = 0.43/1.28 × 100 = 33.6% (magenta inks typically run high — they lean towards red).
• Greyness = 0.12/1.40 × 100 = 8.6%.
• The same sheet's cyan is specified L\* 55.0, a\* −37.0, b\* −50.0 and measures L\* 53.2, a\* −35.1, b\* −47.4. ΔL\* = −1.8, Δa\* = +1.9, Δb\* = +2.6 → ΔE\*ab = √(3.24 + 3.61 + 6.76) = √13.61 = 3.69. Target chroma C\* = √(37.0² + 50.0²) = √3869 = 62.2, hue angle = 180° + atan(50/37) = 180° + 53.5° = 233.5°. Using ΔE\*94 with S_C = 1 + 0.045 × 62.2 = 3.80 and S_H = 1 + 0.015 × 62.2 = 1.93, the chroma and hue terms are divided down and ΔE\*94 comes out well below ΔE\*ab — which is why a highly chromatic cyan that fails on 1976 numbers often passes on 1994/2000 numbers.

Where it is used

• Commercial and publication offset: press-console scanning spectros driving closed-loop ink keys, CIP3/CIP4 pre-setting, OK-sheet approval under ISO 3664 P1 light — Manipal Technologies, Thomson Press, Repro India, Pragati Offset, Lovely Offset.
• Folding carton and corrugated packaging: brand colour tolerances written as ΔE00 in the buyer's specification, CxF spectral data for spot colours, GCR-heavy separations for grey stability across many print runs — Parksons, TCPL, ITC PSPD, Oji/Avery suppliers to FMCG and pharma.
• Flexible packaging and labels: flexo and rotogravure to ISO 12647-6 and -4; colour on films, metallised PET and white-ink underlays measured with the correct backing and M1 — Huhtamaki, Amcor, Uflex, Avery Dennison.
• Prepress and digital proofing bureaus: profile building, ink-limit setting, ISO 12647-7 proof verification with a Fogra/ISO media wedge before a job is released to press.
• Job roles: prepress/colour management executive, press quality controller, process engineer, QA analyst, application engineer with instrument and RIP vendors (X-Rite, Techkon, Konica Minolta, GMG, Esko, Kodak, Heidelberg) and in-house colour managers at brand owners.
• Textile, ceramic and paint sectors reuse the same colorimetry — spectrophotometric shade matching, ΔE tolerances and metamerism checks — so the skill transfers beyond print.

Interview questions

• Q: What is the difference between a densitometer and a spectrophotometer? A: A densitometer measures light absorbed through a few broad filters and reports density — how much ink is on the sheet; a spectrophotometer measures full spectral reflectance and reports CIELAB, so it tells you what colour it actually is and can compute density too.
• Q: Why is density logarithmic? A: Because ink absorption is multiplicative with film thickness and human response to lightness is roughly logarithmic, so density scales almost linearly with ink film thickness and spreads a wide reflectance range over a convenient 0-2.5 scale.
• Q: Which filter do you read each process ink through, and why? A: Cyan through red, magenta through green, yellow through blue, black through visual — each filter passes the band that the ink absorbs most, giving maximum sensitivity to film thickness.
• Q: Status T or Status E — does it matter? A: Yes. Status E uses a narrow-band blue response and gives a markedly higher yellow density than Status T for the same ink film, so aims, SOPs and supplier data must all state the status.
• Q: What causes dot gain and how do you separate mechanical from optical gain? A: Mechanical gain comes from plate, blanket, impression and ink spread; optical gain from light scattering inside the substrate. Murray-Davies lumps them together; a Yule-Nielsen calculation with the right n-factor estimates the mechanical part.
• Q: Why use print contrast instead of just maximising solid density? A: Density keeps rising past the point where shadow dots begin to fill in; print contrast peaks at the inking level that gives maximum shadow separation, which is the useful optimum.
• Q: What is ink trapping and how do you improve it? A: It is the ability of a wet first-down ink to accept the next ink; measured by the Preucil formula. Improve it with descending ink tack in the printing sequence, lower press speed, correct fountain-solution balance and resequenced colours.
• Q: Why does the same sheet measure differently on M1 and M2? A: Optical brightening agents in the paper fluoresce under UV. M1 includes a D50-equivalent UV component, M2 cuts UV, so the measured L\* and b\* of the substrate — and of light tints — differ, sometimes by several ΔE.
• Q: What is the PCS and why does ICC need it? A: A device-independent reference space (CIELAB or CIEXYZ at D50) that every profile converts to and from, so n devices need only 2n profiles instead of n(n−1) pairwise transforms.
• Q: When would you choose media-relative colorimetric over perceptual? A: For logos, brand spot colours and charts where in-gamut accuracy matters; perceptual is for photographic images where smooth gradation across a gamut mismatch matters more than exactness.
• Q: Which intent do you use to proof one printing condition on another device, and why? A: ICC-absolute colorimetric, because it reproduces the simulated condition's paper white and absolute colours rather than remapping to the proofer's own white.
• Q: Is ΔE\*ab = 4 acceptable? A: It depends on where in colour space it sits and which metric the contract names — 4 on a saturated cyan may be visually slight and pass under ΔE00, while 4 on a near-neutral skin tone or a brand grey is clearly objectionable. Always quote the metric, the measurement condition and the edition of the standard.
• Q: What does G7 calibrate that per-ink TVI control does not? A: Grey balance and the neutral print density curve across the tone scale, so visually similar results are obtained from offset, flexo and digital output even though their individual ink TVIs differ.

Common mistakes

• Comparing densities taken with different status responses, apertures, backing or polarisation — or comparing a wet press reading with a dry one.
• Forgetting that Murray-Davies needs paper-relative densities through the same filter; zeroing on absolute white inflates the computed tone value.
• Reporting TVI as a ratio or percentage of the nominal value instead of the difference in percentage points (66.9% at a nominal 50% is +16.9 points, not 33.8%).
• Reading trapping through the wrong filter — it must be the filter of the second-down ink, not of the overprint colour.
• Treating density as colour: chasing on-aim densities while the hue, greyness or paper shade has changed, and then being surprised the customer rejects the sheet.
• Judging or approving colour under shop lighting or a window instead of ISO 3664 D50 conditions, and ignoring the surround and the viewer's own colour vision.
• Assuming an ICC profile fixes an unstable press — profiling an uncalibrated, un-ink-limited machine just encodes today's drift; linearisation and ink limiting come first.
• Quoting a standard without its part and edition ("ISO 12647" alone), or quoting ΔE tolerances from an old edition when the contract is written in ΔE00.

Related topics

Colour theory · Printing substrates and inks: paper, board, films and ink chemistry · Prepress, screening and plate making · Offset, flexo and gravure printing processes · Digital printing and variable data · Statistical process control and ISO 9001 quality systems