How tight is a proof-to-print match under ISO 12647-7
The digital proof must first pass ISO 12647-7 on its own, then you ask whether the print can follow it. In 12647-7:2016 a contract proof’s mean ΔE00 across all patches is about 2.5 or less, 95% of patches about 5 or less, with a separate cap on paper white. Do not accept a proof with the solid ΔEab 5 from 12647-2 Table 7. Use the edition the contract cites.
The digital proof must first pass ISO 12647-7 on its own, then you ask whether the print can follow it. In 12647-7:2016 a contract proof’s mean ΔE00 across all patches is about 2.5 or less, 95% of patches about 5 or less, with a separate cap on paper white. Do not accept a proof with the solid ΔEab 5 from 12647-2 Table 7. Use the edition the contract cites.
Do not reverse the direction
12647-2 is how wide production can sit. 12647-7 is whether the digital proof used for sign-off is good enough to be a target. If the proof is looser than production, the target itself is skewed; the tighter the print chases it, the farther it sits from the real reference. Measure the proof against the reference data first. When the proof passes, the print goes against the proof or against the target the contract names.
When you compare, M condition, backing, and geometry must match how the proof was made. Proof in M1, print in M2, and fluorescent paper will fail the whole sheet.
Who against whom
| Comparison | Whose tolerance | Do not use |
|---|---|---|
| Digital proof vs reference data | 12647-7 | Solid ΔEab 5 from 12647-2 |
| Print vs a proof that already passed | Production tolerance in the contract | Adding the proof tolerance a second time |
| Proof never measured | Do not print yet | Printing as if it were correct |
When you record, write
- The proof’s measurement conditions and mean color difference.
- Whether the print is against the proof or against reference data.
- If the contract names a year edition, write that year. Do not let “about” stand in for that table.
