Granite Color Difference Acceptance: No Universal Figure, So Fix the Sample First
There is no universal range for granite color difference. What belongs in the contract is not a percentage but three things: an approved sample, the conditions of comparison, and a remedy when a piece falls outside. Compare under one light source, in the same dry or wet state, from one metre back. A 100 m² minimum order is 278 pieces at 600×600, and a 20ft container of 346 m² is 961 pieces, so piece-by-piece checking is out of reach and sampling goes by the pack.
What range of color difference is acceptable in granite? First separate the three kinds
For granite color difference, no single figure holds for every project. What works in practice is to split color difference into three kinds: patches of light and dark inside one slab, variation between slabs in the same batch, and variation between one batch and the next. The first is the grain of the stone itself. Acceptance does not call it color difference, and it cannot be selected out. The other two are what acceptance controls, and they come from three different places: another block from the same quarry, the crust-to-core drift within one block, and a polished face against a flamed one. So the useful question is not how much is allowed, but which slab is the reference, under what conditions we compare, and what happens when a piece falls outside.
Why can the contract not simply say color difference stays within a stated percentage?
Because color difference is a judgement made by eye, and no instrument returns a reading for it. Neither side can measure the percentage, so the argument ends in words. A clause that can be executed has three layers: a sample signed by both sides, the conditions of comparison, and a remedy for pieces that fall outside. Without the first layer the other two have nothing to stand on. The clause also has to name who looks and when, at the factory before shipment or at the port after opening, because the light and the state of the stone are not the same in those two places. Leave those three points open and one sentence gets two readings on site.
| What the order usually says | Why it cannot be executed | What to write instead |
|---|---|---|
| "Color difference within a stated percentage" | No instrument measures that percentage on a slab | "Measured against the approved sample, no visible difference from one metre" |
| "Same as the sample" | Samples collect dust and shift in color, and no sample is named | "Three signed and numbered samples held by buyer, factory and supervisor" |
| "Reject if the color is far off" | "Far off" has no line, and nobody owns the freight | "Pieces outside the line are replaced by the pack, freight paid by the responsible side" |
Under what conditions does a color comparison actually count?
Four conditions together. First, light: diffused daylight or one uniform white source, no direct sun and no warm yellow lamp. Warm light pushes red stone toward yellow, and cool white light pulls grey stone toward blue. Second, moisture: the same slab is two different colors dry and wet, so compare dry with dry or wet with wet. Third, distance: step back one metre and read the overall tone. Up close you only see mineral grains. Fourth, position: lay the slabs flat and view them at the same angle. A slab standing on edge reflects light and shifts the color. Set the sample and the test piece on the same flat surface, with no slab of another color between them, or the eye follows the neighbour. One more point: the same difference reads stronger on a polished face and weaker on a flamed one. A polished face concentrates the reflection, so a light piece beside a dark one separates at once. Flamed and bush-hammered faces scatter light, and the grain covers part of the difference in base color. On a project with polished main elevations, set the standard on the polished face. Drop any one of the four and the other side can overturn your reading.
How should the approved sample be made so it holds up at acceptance?
The sample is taken after the order is placed and before the block is cut, not after the goods land. Three pieces, signed by both sides: one to the buyer, one kept at the factory, one to the supervisor or a third party. Write the stone code, the finish, the batch and the date on the back of each. Name the finish explicitly, because polished and flamed do not read the same and comparing across finishes leads nowhere. The pieces need not be large, as long as the base color and the grain show, but all three must come off the same batch. Store them away from light and damp, and do not let them be used as a pad under something else. On G564 Pink Hemp, the three pieces must come from one batch, and a new batch means a new sample. A 100 m² minimum order works out at about 278 pieces of 600×600, and a 20ft container of 346 m² at about 961 pieces, so checking every piece is not realistic. Check by pack: open each one and look at the first and the last piece.
A container holds close to a thousand pieces, so how do numbering and layout hide the variation?
Number in three levels: container, pack, piece, and match those numbers on the packing list. Before laying the full area, set out a few pieces as a trial patch and look at it from a distance. If the light and dark are uneven, change the order there rather than after the whole elevation is fixed. On site, keep one elevation on consecutive numbers but skip as you lay, taking one piece and leaving the next. The change reads softer than laying in strict order. Put the darker pieces on a shaded face, low down, or at an edge that is not seen, and keep the middle of the color range for the main elevation. This needs no agreement from the other side, and it takes effect the day you lay.
If a piece falls outside on site, how do you settle it without losing money?
Three steps. First, stop laying. Once pieces are fixed, it is hard to say whose problem it is, so leave the pieces already fixed in place as evidence. Second, re-check under the same conditions: all three parties present, the same light and moisture, the disputed piece beside the approved sample, and the light and distance written into the photo record. Third, settle in tiers: replace by the pack where stock allows, move the darker pieces to a secondary face, or take a reduction on area and write it into the acceptance sheet rather than agreeing by word. Keep the photo record. It settles more than an argument over how far off the color is.