TL;DR: Most prepress failures that cause reprints aren’t caught at preflight — they’re embedded in how the file was built, and they only show on press at full production speed.
TL;DR: In our experience, misregistered spot colors and wrong bleed setups account for roughly 60% of the first-sample rejections we log under our PR-09 prepress failure tracker.
What the Finished Sheet Tells You About the File That Built It #
Three symptoms show up repeatedly when a file has underlying prepress problems — and each one points to a different root cause.
Symptom 1: Color banding or unexpected tonal shifts in flat tinted areas.
You’ll see this as a stripe of lighter or darker tone running parallel to the print direction, most visible in 20–40% tints of cyan or black. This is almost never an ink problem. It maps to three causes: raster interpolation artifacts from low-resolution placed images (below 300 PPI at final print size), incorrect color mode conversion from RGB to CMYK done inside a layout application rather than via a calibrated ICC profile, or gradient meshes built in Illustrator that weren’t flattened before output.
Symptom 2: Hairline white gaps between adjacent color panels or along die-cut edges.
These appear after lamination or cutting and were invisible on the soft proof. Three causes: bleed was set at 1mm or less on a job that went through a guillotine trim with ±1.5mm mechanical tolerance, adjacent spot colors were set to knockout rather than overprint in panels smaller than 6pt, or the artboard in the source file was sized to the finished trim, leaving zero bleed margin on one or more edges.
Symptom 3: Barcode grade failure on the finished carton.
If your GS1-compliant barcodes verify at grade C or below per ISO/IEC 15416, the file is usually the cause before the press is. Root causes: barcode generated at screen resolution and placed as rasterized JPEG, bar width reduction not applied for flexo (minimum 0.1mm reduction required for 133 lpi flexo plates), or barcode colors built from process CMYK mixes rather than a solid black (100K) or dedicated spot.
| Symptom | Primary Root Cause | Secondary Root Cause |
|---|---|---|
| Banding in flat tints | RGB-to-CMYK conversion in layout app | Low-PPI image rasterization |
| White gaps at trim / between panels | Insufficient bleed (under 2mm) | Knockout set on adjacent spot colors |
| Barcode grade C or below | Rasterized barcode placement | Missing bar width reduction for flexo |
| Misregister on small reverses | No trapping on fine type | Plate tolerance not factored in design |
| Unexpected color on proof vs. press | No output intent embedded in PDF | Wrong ICC profile on export |
The Cause Most Prepress Teams Attribute to Press When It Lives in the File #
Dot gain miscalculation on light tints is the failure mode we see misdiagnosed most often. The press operator gets blamed for “heavy ink” when the problem was written into the file weeks earlier.
Here is the mechanism. Every print process has a tonal value increase (TVI), sometimes called mechanical dot gain, where a 20% dot in the file prints as something closer to 28–34% on the substrate after ink spread and impression pressure. For sheet-fed offset on coated SBS board, we calibrate our curves to a TVI of 12–14% at the 50% tonal value, per ISO 12647-2:2013. For UV flexo on uncoated corrugated liner, TVI at 50% runs 18–22% on our lines.
The failure happens when a designer specifies a delicate 15% tint for a background panel — chosen because it looks right on an uncalibrated studio monitor — and no one applies a linearization curve or device link profile before platemaking. That 15% tint prints at 22–25%, and the background that was meant to be barely-there reads as a solid-looking wash.
The confirmation measurement is straightforward: pull a densitometer reading on a test wedge printed in the first 500 sheets of the run. If your 25% patch is measuring above 35% dot area, the curve was either absent or wrong. Our press standard is a tolerance of ±2% TVI across the tonal range — anything outside that triggers a curve recalibration before we proceed to full production quantity.
Where this gets misread: flexo on foil substrates shows even higher TVI because there’s no ink absorption into the surface, so ink spread dominates. We’ve measured TVI as high as 28% at the 50% patch on metalized BOPP. Files destined for foil or metallic substrates need a separate dot gain compensation curve, not the standard offset or corrugated profile. Using the wrong profile is the source of the misdiagnosis — the press operator adjusts anilox pressure trying to control gain that the file already guaranteed.
Corrective Actions, Ranked by How Much Ground They Recover #
-
Re-export the PDF with the correct output intent and ICC profile embedded. This fixes color-mode and rendering issues in roughly 70% of the cases we receive where the brief says “color looks wrong on proof.” Takes under 30 minutes if the source file is intact. Use PDF/X-4 with an embedded destination profile (ISO Coated v2 for European offset, FOGRA39, or GRACoL 2013 for North American sheetfed). Zero press downtime required.
-
Rebuild bleed and safe-zone margins to spec before platemaking. Minimum bleed: 3mm on carton jobs, 5mm on flexible pouch side seams. Safety margin (live area inset from trim): minimum 4mm on carton. If the file arrives with 1mm bleed, it needs to be rebuilt — not stretched in a PDF editor, which distorts live content proportionally. This step takes 2–4 hours of studio time but eliminates rework after print.
-
Apply process-specific dot gain compensation curves at the RIP stage. If the file cannot be recalled and rebuilt, your prepress operator can apply a TVI correction curve at the RIP. This is a partial solution — it controls the midtone spread but cannot undo incorrect ICC conversions already baked into placed images. We use this as a stabilizing step, not a replacement for correct file preparation.
-
Regenerate barcodes natively in a vector application at final print size. Never scale a barcode after generation. A GS1-128 barcode generated at 80% magnification and then scaled to 100% in a layout file accumulates rounding errors in bar width. Regenerate at the correct magnification, export as vector EPS or native PDF, and confirm X-dimension per GS1 General Specifications before platemaking. The minimum X-dimension for standard retail is 0.495mm.
-
Run a device-level proof against a certified contract proof before plate output. A calibrated Epson or Canon large-format proofer running a certified proofing profile (ISO 12647-7) will surface 90% of remaining color issues before a plate is imaged. The cost of one contract proof set (typically 4–8 sheets) is a fraction of a plate remake, which on a 6-color carton job runs to real money.
What to Lock Down Before the Job Reaches Prepress #
Specify output intent and color mode in your purchase order, not your email. The PO or spec sheet should state: color mode (CMYK only, no RGB or LAB objects), PDF standard (PDF/X-4 preferred), minimum image resolution (300 PPI at 100% final size for continuous tone, 1200 PPI for line art and barcodes), bleed dimension, and which press-standard ICC profile the job is being prepared for.
Ask your prepress contact for a signed-off preflight report and a calibrated soft proof before approving for plate output. The preflight should be run against a packaging-specific profile, not a generic Acrobat preflight. Request our PR-09 Prepress Submission Checklist — it covers 34 parameters specific to folding carton, flexible, and label work and is what our team uses internally for every incoming file.
Specification Notes for Brand Partners #
When you brief us on a new packaging job, the two things that most often cause unnecessary sample iterations are missing output intent information and undefined varnish knockout areas.
For output intent: tell us whether your studio works in Adobe RGB, sRGB, or is already CMYK-profiled, and which press standard your agency has been designing to. If the file arrives as CMYK but built to a monitor-default profile, we will catch it — but the correction takes time and sometimes requires going back to the originating studio.
For varnish and surface finishing: spot UV and soft-touch matte lacquer both require a separate spot channel in the file, and the boundary between coated and uncoated areas affects the perceived color of the underlying print. Files without a varnish separation layer cause at least one extra sample cycle in roughly 40% of the luxury carton jobs we handle.
Our standard sampling timeline for folding carton is 10–15 working days from approved file to first physical sample. That timeline extends by 5–7 working days if the file requires structural rebuild or color profile correction. Send us the file early — we run a complimentary prepress check before we quote.
Does low image resolution always cause visible problems on press?
Not always — it depends on viewing distance and substrate. A 200 PPI image placed on a corrugated shipper viewed from 1 metre is usually acceptable. The same 200 PPI image on a cosmetics carton viewed at arm’s length will show visible pixelation in curves and gradients. Our standard threshold is 300 PPI at final print size for consumer-facing carton, which aligns with ISO 12647-2 output expectations for commercial offset.
Our designer set bleed to 3mm — is that enough for all packaging types?
For sheet-fed offset carton, 3mm is our minimum and it’s sufficient when cutting tolerance is within ±0.5mm on a die-cut job. For flexible packaging with side-seal trim, we require 5mm bleed on the seal edge because the mechanical tolerance on a pouch form-fill-seal line runs ±2–3mm. The 3mm rule doesn’t transfer from carton to flexibles without adjustment.
Can we submit an AI or InDesign native file instead of PDF?
We accept native files in specific circumstances — primarily when the job requires structural artwork integration with our dieline. But native files from studios running different font libraries or linked asset structures frequently arrive with substituted fonts or missing links. PDF/X-4 is our preferred format because it packages everything and embeds the output intent. For rush jobs, a native file with all fonts outlined and all links embedded is workable.
We approved the digital proof but the printed sample looks different — whose fault is that?
This question carries an assumption worth examining. If the digital proof was produced on a calibrated proofer running ISO 12647-7, and the press was running to ISO 12647-2 (or the relevant substrate standard), and the proof was approved as the color target, then a significant departure on press is a press calibration issue. But if the digital proof was a screen PDF viewed on an uncalibrated monitor, it cannot be used as a binding color reference. We specify contract proofs as the approval standard for any job where brand color accuracy is contractual — screen approval is only accepted for structural/layout sign-off.
Planning a packaging project? Contact our team to request a complimentary specification review and sample quote.