Photo2PatternFree grid drawing tool
Grid drawing tool that prints both halves of the method
Drop a reference and this prints two pages: the photograph ruled into numbered squares, and the same grid left empty at the size you are drawing. Between them it works out the pitch of one square on each sheet, the enlargement as a linear and an area ratio, and the distance from the corner to every line in millimeters and in sixteenths of an inch — the list you need when the support is bigger than any sheet a printer can feed. No fee, no login, and the reference is opened where you dropped it.
- 100% free
- No signup
- 3 to 24 squares
- Prints at true scale
- Marks in 1/16 in
Press Ctrl+V anywhere on the page to drop in a reference you copied out of a browser tab — no need to click the field first.
Rows follow from the shape of the picture. Eight is the traditional start for a head and shoulders: coarse enough to rule in two minutes, fine enough to fix an eye within a tenth of the face. Detailed work goes to 16 or 24 — past that the ruling takes longer than the drawing.
A 10 mm margin all round and a 6 mm gutter for the labels leaves 184 × 271 mm for the picture, and it is fitted inside that whichever way round it is.
The width of the paper, board or canvas you are drawing on, from 2 to 98 in. Its height is not yours to choose — it comes from the photograph, and forcing it is how a face gets narrow.
How to set up a drawing grid that keeps its proportions
The width of the surface you are drawing on is the entry people skip, and every measurement below it is wrong without it.
Choose the squares before you choose anything else
The count decides how much of the drawing the grid does for you and how much you do. Eight across is the traditional start for a head and shoulders: it fixes an eye to within a tenth of the face and takes two minutes to rule. Sixteen or twenty-four suits an architectural subject or a full figure, where a misplaced window is unforgivable and a misplaced eyelash is not. The row count is not yours to set — it comes from the proportions of the photograph, and the panel prints it as soon as a file is open.
Enter the width of the thing you are drawing on
Not the width of the paper coming out of the printer — the width of the board, canvas or pad in front of you, in inches or centimeters. Everything downstream is that one number: the pitch of a square on your support, the enlargement as a linear ratio and as an area ratio, and whether the empty sheet can print at full size or has to print as a reduced map of itself. A 16 in support from a photograph printed 184 mm wide across A4 is an enlargement of 2.21, which is 4.9 times the area to cover.
Print at 100%, then mark out from one corner
Set the print dialog to 100% or Actual Size rather than Fit to Page: both sheets carry a true millimeter width, and fit-to-page quietly rescales them by three to six percent, which is enough to make the printed square disagree with the table. If your support is bigger than the paper, take the mark-out list instead and measure every line from the same corner. Stepping a rule one square at a time adds the sixteenth-inch rounding of each square together, and by the far edge that drift is most of a square.
Technical specifications
| Squares across | 3 to 24, with the row count taken from the photograph rather than typed. The tool refuses a grid where a square would hold fewer than 24 source pixels, and offers the coarsest count that clears it |
|---|---|
| Reference sheet | A4 (210 × 297 mm) or US Letter (215.9 × 279.4 mm). A 10 mm margin all round and a 6 mm gutter for the labels leaves 184 × 271 mm on A4 and 190 × 253 mm on Letter, and the picture is fitted inside whichever way round it is |
| Support width accepted | 2 to 98 in, or 5 to 250 cm. The height is not an input — it follows from the photograph, and the panel prints both so a mismatch with your canvas is visible before you rule anything |
| Pitch on your support | Given to two decimal places in millimeters and rounded to the nearest sixteenth of an inch, with the worst rounding error over the whole sheet stated beside it — a single square rounded to a sixteenth is out by up to 0.79 mm |
| Enlargement reported | As a linear ratio and as an area ratio. Doubling the width of a drawing quadruples the surface to cover, and the area figure is the one that predicts the afternoon |
| When a square stops being square | Only if you ask. By default the squares stay square and the bottom row is left short by the remainder, printed as a percentage of a full square; the alternative gives a whole row count with oblong cells, and the panel states the cell aspect and the resulting distortion in percent |
| What it hands over | Two print pages carrying a true millimeter width, the ruled reference as a PNG at 305 dpi, and the mark-out list as CSV — every line's distance from the corner, along the top edge and down the left |
| Where the reference is read | In this tab. The two sheets are drawn on a canvas in the page rather than fetched as a PDF from anywhere, which is why they appear the moment a control moves |
Frequently asked questions
How many squares should I use?
Use the fewest that still pin down the thing you are worst at placing. A portrait at eight squares across puts each eye inside a known square and leaves the modeling to you, which is the point of the method — it is a positioning aid, not a tracing. Sixteen is right when the subject has hard structure that reads as wrong when it moves: window frames, a boat's sheer line, lettering. Above twenty-four the grid stops helping, because ruling twenty-five lines accurately on a large support takes longer than the drawing and every one of them has to be erased afterwards. The tool refuses to go finer than a square holding 24 pixels of your file, because below that there is nothing inside the square to copy.
Why is the bottom row of squares shorter than the others?
Because a photograph's proportions almost never divide into a whole number of squares, and this tool leaves the remainder rather than hiding it. A 4:3 frame at eight squares across wants exactly six rows, but a 3:2 frame wants 5.33 and a phone photograph in 19.5:9 wants 3.69. The honest options are a short bottom row or squares that are not square, and the default is the short row because the alternative distorts every measurement you take from the grid. Turn the checkbox off and you get a whole row count with slightly oblong cells, and the table prints the cell's aspect and exactly what percentage the drawing stretches by if you rule it as a square anyway.
My photo is 3:2 and my canvas is 4:5. What do I do?
Decide what to lose before you start, because nothing further along can decide it for you. The grid here follows the photograph, so entering a width gives you the height the photograph implies, not the height of your canvas — a 16 in wide support from a 3:2 photograph is 10.7 in tall, and a 4:5 canvas 16 in wide is 20 in. Either crop the photograph to your canvas shape first and drop the cropped file in, or accept a band of unused canvas at top and bottom. What you must not do is squeeze the grid to fit, which is the same as squeezing the subject.
Can I print the blank grid at full size for a 16-inch board?
No, and it deliberately does not try. Reaching 16 in at true size on A4 would take four sheets across and three down, and twelve taped seams move under a straightedge in a way nine pencil lines do not — the tape alone is thicker than the tolerance you are working to. When the target is bigger than the sheet, the second page prints as a reduced map of the grid and the real work is done from the mark-out list: eight distances along the top edge and however many down the side, each given in millimeters and as a fraction a steel rule is actually divided into.
Should I draw the grid on the drawing itself?
Rule it lightly and in the right medium, or keep it off the support entirely. A 2H pencil at the lightest pressure that still leaves a visible line erases cleanly from smooth paper; a soft pencil ruled hard leaves a groove that shows through a wash for the life of the piece, and graphite under watercolor or acrylic lifts into the paint as a gray haze. The other approach is to rule the grid on tracing paper or acetate, tape it over the support, and lift it off once the main placements are in — slower to set up and the only method that leaves nothing behind at all.
Is the grid method cheating?
It is one of the oldest recorded drawing aids in Western painting, which is a different question from whether it suits you. Leon Battista Alberti described a veil of threads stretched on a frame in De pictura in 1435, and Albrecht Durer's woodcut of a draftsman working through a gridded screen was printed in his book on measurement in 1525 — both were teaching devices, published so that the technique would spread. What the grid does is remove one specific failure, the slow drift where a head grows and a shoulder wanders, so that attention goes to value and edge instead. It does not make a mark for you, and it will not save a drawing whose observation is wrong inside the square.
Does printing in gray help, or is it just a preference?
It removes a real distraction, and this page converts through linear light rather than by averaging the three channels. Hue is the part of a photograph most likely to mislead when you are working in graphite or charcoal, because a saturated red and a mid green can be the same value while looking nothing alike, and the eye insists they are different. Flattening to value shows what is actually going on the paper. The naive conversion — adding red, green and blue and dividing by three — darkens midtones by several percent because it averages gamma-encoded numbers rather than light, so the reference comes out heavier than the subject.
About the grid method, and the two places it goes wrong
The method is nearly six hundred years old and has barely changed. Alberti described a veil of threads on a frame in 1435 as a way of holding a subject still in the eye; Dürer’s woodcut of 1525 shows the same screen with a draftsman sighting through it onto a gridded sheet. What both were solving is a specific failure of freehand copying: not that the marks are wrong, but that they drift. A head enlarges by a few percent as you work down it, a shoulder wanders, and because every step looked right relative to the last one there is no moment where the mistake is visible. A grid converts one hard global judgment — where does this go on the whole sheet — into a series of easy local ones inside a square you can see all of at once.
The first place it goes wrong is the cell shape. Most grid tools take the frame and cut it into a fixed number of parts in each direction — four by four, three by three — which gives cells that are only square when the picture happens to be. A 3:2 photograph cut into 8 by 6 gives cells 1.125 times wider than they are tall, and copying those into the square cells somebody ruled on a canvas stretches the whole drawing horizontally by twelve and a half percent. Nothing about the result looks obviously broken; the face is just slightly wrong in a way nobody can name. This page keeps the squares square and leaves the bottom row short instead, and if you want the whole row count it tells you precisely what the distortion is before you commit to it. If squares are not what your subject wants at all, a hexagonal grid divides the same area with no aligned rows to drift along, and if you want the ruled sheet with no photograph on it, that is printable graph paper at an exact pitch.
The second is measuring. Once the support is larger than a sheet of paper the grid has to be ruled by hand, and the standard mistake is to set a rule to one square and step it along. Every step carries its own rounding — a pitch of 41.28 mm read on a rule divided into sixteenths becomes 41.275, and eight of those add up — so the last column lands most of a square out and the error is all in one place. Marking every line from the same corner keeps the error under half a millimeter anywhere on the sheet, which is why the list here gives cumulative distances rather than a pitch. For a subject where the problem is not placement but value, flattening the reference to tone helps more than any grid does, and reducing the photograph to its edges or to numbered flat regions takes that further. If the squares are going to be filled with something countable rather than drawn in — beads, stitches, plates — then the grid is a chart and the tool you want is the bead chart, which counts every square for you. And if all you need is the ruling burned into the picture file itself, that is the grid overlay, one page across.
Where the reference stays
Your picture is opened by JavaScript running in the tab you are reading this in. It is never uploaded, never sent to a server and never stored anywhere — there is no upload step to undo, because there is no upload. Close the tab and nothing of it remains.
The two print sheets are drawn on a canvas inside the page, so there is no PDF being generated on a server and nothing to fetch when you change the square count — which is also why the sheets redraw as fast as you can move the slider.