Needlepoint pattern maker — for canvas, not aida

No, and four things differ before the first stitch. The ground is canvas measured in mesh, meaning holes per inch, rather than aida measured in woven blocks. The canvas must be covered completely, so bare mesh showing between stitches is a fault rather than part of the look. The stitch is a tent stitch, a single diagonal over one intersection always slanting the same way, not a cross made of two. And because coverage is total, the thread has to be thick enough to fill the hole — several plies of Persian wool or six to twelve strands of cotton, where cross stitch on 14-count is usually two. A chart drawn for aida will physically fit a canvas of the same count, but it will be costed wrong, worked in the wrong stitch, and threaded too thin to hide the mesh.

Start at four plies of three-ply Persian wool for 10 mesh, three for 12 and 13, two or three for 14, two for 18 and one for 24; in six-stranded cotton that is roughly twelve strands down to two. Then test it, because there is no published coverage standard and shops differ by a ply — the same wool covers differently depending on how hard you pull and which dye lot you have. Work a dozen stitches out in the bare margin, hold the canvas up to a window, and add a ply if daylight comes through. Discovering it halfway through the sky means unpicking the sky.

No — the chart is identical for all three, and only the yardage and the blocking change. Half cross, continental and basketweave put the same slanting stitch in the same intersection with the same shade; they differ in how the needle travels on the back, which is where most of the thread is spent and where the distortion comes from. Continental pulls every back stitch the same way, so a piece worked entirely in it comes off the bars as a rhomboid and has to be blocked square. Basketweave alternates the pull and stays flat, which is why it costs 3.41 stitch-widths a stitch instead of 2.41 and why finishers ask for it on anything that will be sat on.

Yes, in continental or basketweave, but not in half cross. Plain mono is one warp thread crossing one weft with nothing holding the crossing, which is why it takes paint so smoothly and why it is what most stitch-painted canvases are printed on. A half cross stitch leaves almost nothing on the back to anchor it, so the canvas threads shift and the stitches creep off their intersections and let mesh show between them. Interlock mono twists two fine threads around each weft, locks every crossing and does not unravel when cut, which is the ground that forgives a half cross. Penelope has a further trick: prick its thread pairs apart with the needle and the same canvas takes petit point at double the count, so one piece can carry a coarse background and a fine face.

Because a finer mesh gives the match more intersections to spend, and each one averages a smaller patch of the original. At 13 mesh a cheek is thirty intersections and four of DMC's tans cover it; at 18 the same cheek is fifty-eight, the averages separate further, and seven tans get reached. That is more modeling in the finished face and more cuts of wool to buy and change between. The fold-in control is the way back down that scale once you have seen the cost — it is deliberately the one number on this page that removes shades rather than adding them.

Because a needlepoint shop cuts wool off the wall by the yard and will not sell you two thirds of one. Every figure here is that shade's stitch count times the inches a stitch takes at your mesh, plus a waste allowance, rounded up to the next whole yard. Read them as working thread — what ends up in the needle — because that is not the same as what you buy: Persian is sold as a whole three-ply strand, so at 18 mesh you pull two plies and one purchased yard gives you one and a half, while at 10 mesh you pull four and need a third more strand than thread. The panel converts both ways, and to six-strand skeins at 8.75 yards for anyone working the chart in cotton. All of that is estimate; the stitch counts beside them are not — those are counted intersections, and they are the numbers to trust.

Full coverage removed the ground, and the ground was doing more work than you think. On aida, unstitched fabric shows around and between the crosses and lightens the whole field, which separates adjacent midtones for free; on a covered canvas two neighboring shades sit edge to edge with nothing between them, so any separation has to come from the shades themselves. Wool compounds it — a matte, slightly hairy surface scatters light and swallows the specular highlight a photograph relies on for shape. The fixes are to raise the contrast of the source before charting, to work at a finer mesh so more of the value range gets reached, and to keep the very darkest and lightest shades in the list rather than folding them out for being rare.