Free Grid Overlay Tool

A uniform grid divides the frame into equal cells for copying a drawing cell by cell. Nothing about it is compositional: the lines fall wherever equal division puts them, and a four-by-four grid on a portrait says nothing about where the eyes should be.

Cell count controls accuracy, because each cell holds less of the picture as the count rises and errors stay local instead of propagating. Four by four is a loose sketch grid; eight by eight is a working default; sixteen by sixteen is for detailed copying and takes considerably longer.

The practical limit is legibility. Past about twenty divisions on a normal screen the lines start to obscure more of the photograph than they organise, which is why the range here stops there.

Rows and columns are set independently on purpose. A panoramic frame divided equally on both axes produces cells that are wide and short, and comparing a wide cell to a square one on your paper is exactly the confusion the grid is supposed to remove.

The rule of thirds tool places lines at one third and two thirds — 33.3% and 66.7% of the frame — where the rectangular transfer grid divides the frame into as many cells as a copy needs. The golden ratio places them at 38.2% and 61.8%, the divisions produced by the ratio 1:1.618.

The gap is about five per cent of the frame, which on a 4,000-pixel image is roughly 190 pixels — small enough that the two overlays look similar at a glance, large enough that a subject placed on one is visibly off the other.

Both are offered here without an argument about which is correct, because the honest position is that neither is a law. The thirds grid is a teaching device of the nineteenth century that has become the default in cameras; phi divisions have a longer history in design and a weaker claim to being perceptually special than is usually asserted. What they share is that they move the subject off dead centre, which is the part that reliably helps.

Corner-to-corner diagonals do one thing precisely: they cross at the true centre of the rectangle. That is useful because the perceived centre of a frame sits slightly above the geometric one, and a subject placed at the true centre often looks low.

They also reveal corner relationships. A line in the photograph running parallel to a diagonal, or a subject sitting where a diagonal meets an edge, produces a relationship the eye registers even when the photographer did not intend it. It is a diagnostic overlay rather than a placement one.

The obvious implementation draws the image at whatever size it appears on screen, draws the grid on top, and exports that. It is simpler, and it silently caps every export at the size of the browser window.

This tool sets the canvas to the image's natural dimensions and draws into that. A 4,000 by 3,000 photograph exports at 4,000 by 3,000 regardless of how small the preview is, which matters the moment the gridded file is going to be printed rather than looked at.

Line weight follows from the same decision. Specifying thickness in screen pixels would produce a line that is prominent on a small image and invisible on a large one. Expressing it as a proportion of the shorter side keeps the visual weight constant across sources, which is why the same setting looks right on a phone photo and a scan.

Most gridded references get printed rather than worked from on screen, and the print is where a low-resolution export shows. Lines that look crisp in a browser at 900 pixels wide become soft grey bands once the same file is enlarged onto A4, which is the main practical argument for compositing at source resolution.

The other printing consideration is that the grid should be lighter on paper than it looks on screen. A monitor is emissive and a print is reflective, so the same eighty per cent opacity reads heavier on paper. Dropping the opacity a little before exporting a file destined for print usually gives a reference that is easier to draw from.

White lines read on most photographs because most photographs have more mid and dark tone than highlight. On snow, sand, pale skin or a white studio background they disappear, and black is the answer.

Full opacity is rarely right. At one hundred per cent the grid competes with the picture and hides detail exactly where the lines cross, which tends to be where the subject is. Around eighty per cent keeps the lines readable while letting the photograph show through them.

This page does one grid on one picture and gives you the file. That is deliberately narrow, because the query it answers is narrow, and answering it with an eighty-two-overlay application would be answering a different question.

Stacking several overlays, transforming each independently, saving workspaces, sharing a state by link, exporting to PDF or SVG at a chosen DPI — all of that lives in the full tool. If you find yourself wanting the second overlay, that is the point to move across. The two families are catalogued apart for the same reason: the basic drawing grid overlays measure, and the composition category judges.