ICC Profile Creation & Validation — Troubleshooting & Failure Guide

What the Press Sheet Is Telling You — Reading the Symptoms Correctly #

Three failure patterns show up repeatedly when an ICC profile is performing outside spec on a packaging press run.

The first is a global hue shift — typically a cyan-green cast in neutrals, or a warm red push in shadow areas — that appears consistent across the full sheet. Brand colours may still fall within tolerance individually, but skin tones or neutral backgrounds drift visibly. The second is a contrast collapse in midtones: shadow detail prints correctly, highlights are clean, but the 40–70% tonal range compresses. Gradients meant to transition smoothly look stepped or flat. Third is localised metamerism: a colour matches under D50 light on the press floor but fails badly under D65 or store fluorescent. The printed pack and the approved physical reference look like different colours in-store.

Each symptom points to different root causes. The diagnostic table below maps the most common patterns:

If you are seeing the first and third symptoms together, the diagnosis is almost always measurement condition error, not profile construction error. That distinction matters because the corrective actions are completely different.

The Root Cause Most Teams Misdiagnose — OBA Interaction Under M0 Measurement #

Optical brightening agents (OBAs) are added to coated and uncoated folding carton stocks during papermaking to increase apparent whiteness. Under a standard D50 illuminant with UV content (ISO 13655 M0 condition), OBAs fluoresce and raise the measured L* of the paper white. A profile built from M0 measurements assumes this fluorescence is real and stable. It is neither.

The problem is that OBA fluorescence varies with UV energy in the light source, and printing inks partially suppress OBA activity through optical masking. The result: your ink-free paper white measured at M0 might read L 97.2, but under actual press conditions with ink coverage, the effective white point in printed areas drops to approximately L 93–94. The profile has been told the paper is brighter than it actually behaves in the printed state. Every colour calculation downstream is anchored to a white point that does not exist on the finished pack.

This manifests as a systematic shift toward yellow-green in neutrals (because the profile is compensating for a blue-white that is no longer there) and as a general loss of apparent contrast, because the actual printed white point is lower than the modelled white point.

The ISO 13655:2017 M1 measurement condition was introduced specifically to address this. M1 uses a defined UV component in the measurement illuminant that approximates D50 UV content, giving stable, reproducible readings regardless of OBA intensity in the substrate. When we re-profile a job that is exhibiting neutral hue shift, our first action is always to re-measure the characterisation target under M1 and compare the Lab values to the original M0 dataset. If the paper white shifts more than 2.0 Delta E 2000 between M0 and M1 measurements, the substrate has high OBA content and the profile must be rebuilt under M1.

Confirmation measurement is straightforward: measure the unprinted substrate on your spectrophotometer in both M0 and M1 modes. We use a threshold of 2.0 ΔE 2000 between the two readings as our trigger criterion, logged under our CP-04 profile risk assessment procedure. Above that threshold, any profile built under M0 on that stock is suspect, regardless of how carefully the characterisation data was captured. Below 2.0 units, OBA influence is typically negligible and the press can continue under the existing profile with standard TVI monitoring.

This is where industry practice splits. Some prepress houses measure everything under M0 for legacy compatibility with older RIP profiles and customer-supplied data. Others have moved fully to M1 for all new profile builds. Our practice: M1 is the default for all new characterisation work on coated stocks above 115 gsm. For recycled or uncoated stocks with low OBA content, we continue to accept M0-built profiles from customers, but we flag the measurement condition on the job ticket and check the substrate delta at incoming.

Corrective Actions Ranked by Impact and Feasibility #

Once you have confirmed the root cause, the response needs to match the failure mode. These are ranked by how much they move the needle, not by how easy they are.

1. Rebuild the profile under M1 measurement conditions. This resolves OBA-driven neutral drift permanently. Requires re-running characterisation prints on the production substrate with a sufficiently dense patch set (we use ECI2002 R2 at a minimum — 1,485 patches — rather than the older IT8.7/4 at 928 patches). Turnaround on our profiling line is typically 4–6 hours from characterisation print to validated ICC profile. This fixes the majority of hue shift cases.

2. Re-linearise the press before re-profiling. If TVI has drifted from the profile reference (commonly seen after ink viscosity changes, blanket age, or substrate batch switch), re-profiling on top of a shifted press state embeds the error permanently. Check the 50% TVI against the characterisation target — if it has moved more than ±3%, re-linearise first. Skipping this step and just re-profiling is a common error; it delays resolution by a full press cycle.

3. Switch to M1-compatible spectrophotometer settings for all press-side verification. If your inline or handheld spectrophotometer is measuring under M0 while the profile was built under M1, every Delta E number on the press floor is wrong. This is a calibration and workflow configuration fix, not a hardware purchase — most i1Pro 2 and Konica Minolta FD-9 instruments support M1 natively. Configuration change takes 30 minutes. Impact on measurement accuracy on high-OBA stocks is significant.

4. Request substrate OBA content data from the paper mill. Most coated folding carton mills will supply a fluorescence index or OBA activation level on request. Qualifying new paper stocks against our CP-04 threshold before they enter the press schedule prevents the problem from reaching the press floor. This holds for SBS grades — for recycled boards and natural kraft stocks with no OBA additions, the M0/M1 delta is typically below 0.8 ΔE and the issue rarely applies.

5. Rebuild with a custom characterisation dataset, not a standard ICC reference profile. Using FOGRA51 or GRACoL 2013 as a profile target on a non-standard substrate introduces a second layer of mismatch on top of any measurement condition error. Both are valid frameworks under ISO 12647-2 and CGATS.21, but they were characterised on specific reference papers. If your production stock deviates from the reference substrate by more than 3 ΔE 2000 in white point, the standard reference profile will underperform a custom-built one. We see this most often on specialty coated stocks — soft-touch laminated cartons, aqueous flood-coated boards — where the optical characteristics differ substantially from coated offset reference papers.

Prevention — What to Specify Before a Profile Is Built #

In your supplier brief or profile specification document, include: measurement condition (M0 or M1), spectrophotometer model and geometry, target characterisation standard (ISO 12647-2 Condition 1 or equivalent), patch set size and format, and Delta E 2000 tolerance bands for primary colours and neutrals. Ask for the substrate OBA delta measurement as a deliverable at incoming inspection.

If you are commissioning a new profile build rather than using a factory standard, also request the raw characterisation data in CGATS format alongside the finished .icc file. This is the only way to audit the profile if press behaviour diverges later.

The document to request from your print supplier: a profile validation report including the characterisation print run date, substrate lot, instrument serial number, measurement condition, and Delta E summary across the full tone range, structured against ISO 15076-1 ICC profile architecture requirements.

Specification Notes for Brand Partners #

When you brief us on a new packaging print job involving colour-managed ICC workflows, the most useful thing you can send upfront is the approved substrate specification — mill name, grade, coating type, and if available, the fluorescence index from the mill data sheet. We receive briefs that specify Pantone references and Delta E tolerances in detail but omit the paper stock entirely, which means we cannot validate measurement condition compatibility until physical samples arrive and we run our CP-04 incoming check. That adds 2–3 working days before profiling can begin.

One brief gap that causes unnecessary sample iterations: customers approving soft proofs on a monitor calibrated to a D65 white point but expecting press output to match a D50 proof. The mismatch is not visible on screen but is clearly visible when the physical press sheet and monitor are compared side-by-side under D50 viewing conditions. We flag this during brief review, but if the monitor calibration report is included in the initial brief, we can catch it before samples are produced.

Our standard profiling-to-press-proof timeline for a new substrate is 5–7 working days from receipt of confirmed substrate stock. Jobs requiring metamerism evaluation under multiple illuminants (D50, D65, A) add 1–2 days.

What Delta E tolerance should we specify for brand colours?

For primary brand colours on packaging, we work to ΔE 2000 ≤ 2.0 for solid process builds and ΔE 2000 ≤ 3.0 for spot colour simulations in CMYK. Tighter than 1.5 ΔE is achievable on stable, well-characterised stocks but requires controlled measurement conditions on both sides — press-side and approval-side — using the same illuminant geometry. If your approval workflow uses a different spectrophotometer model than your factory’s inline system, the instrument-to-instrument variation alone can account for 0.4–0.8 ΔE, which eats into your tolerance budget before ink is laid down.

Can we use a standard FOGRA51 or GRACoL profile instead of a custom-built one?

The premise here needs examining. Standard reference profiles work well when your production substrate closely matches the reference characterisation paper — coated offset above 130 gsm with moderate OBA content. On specialty stocks, heavily coated boards, or uncoated recycled grades, a standard profile may introduce systematic errors that no press calibration can correct, because the error is in the profile’s assumption of paper white, not in the press behaviour. The decision depends on how closely your actual stock matches the reference paper’s Lab white point and TVI curve. We measure this at incoming and make the call based on data, not defaults.

If the press sheet looks good to the eye but fails the spectrophotometer check, which takes precedence?

The spectrophotometer, always — but the failure mode matters. A visual pass with an instrument fail usually points to one of two things: a metamerism condition (looks fine under the press room’s D50 viewing booth, fails under a different illuminant that the spec references) or a geometry mismatch between how the approval proof was measured and how the press sheet is being measured. Before pulling the job, confirm both instruments are measuring under the same M condition and the same geometry. We have seen jobs held up by a 45°/0° vs. d/8° geometry discrepancy that accounted for nearly the entire apparent Delta E — per ASTM E308 colorimetry measurement principles, geometry differences on glossy substrates can produce substantial Lab variation that is an artefact of measurement setup, not colour error.

---

Planning a packaging project? Contact our team to request a complimentary specification review and sample quote.