Your print file is lying to you

A design gets approved on a laptop, at whatever size the window happened to be, on a screen that emits light. Then it gets printed with ink, on cotton, at a size somebody had to decide.

Every one of those steps is a chance for the thing on the shirt to not be the thing that was signed off, and the person who signs off is almost never the person who finds out. We got burnt by this enough times that we built a system for it: 28 garment types with a measured print area, 1,439 designs carrying 2,284 sized placements, 20 blanks calibrated against what actually came off the press. That is not because we enjoy spreadsheets.

The three that get us most often

Scale. Artwork approved at screen size lands on the garment at a real measurement in centimetres, and a design that looked confident in a browser can arrive looking like a pocket print. Approve in centimetres or do not approve.

Placement. Print position is measured down from the collar seam, not from the top of the garment, and the right distance is different on a small and a 3XL. Fix one number for the whole run and the print creeps up the chest as the sizes go down. We now store a collar-seam start for every garment that needs one, and scale from there.

Colour. Screens are backlit and cotton is not. A blue that glows on a monitor sits flat on a dark garment. Anything with a gradient into the shirt colour is a gamble.

Nobody has ever worn a PDF.

The habits that fix it

• Approve print size in centimetres against a real garment measurement, never in pixels.
• Scale placement per size rather than using one number for the whole run.
• Print one before you print four hundred, and look at it in daylight rather than under the workshop lights.
• Keep the approved sample. When a reprint comes back different, memory is not evidence.

It is boring, careful work. It is also the difference between finding a problem on one shirt and finding it on a pallet, and we have done it both ways.