A flat spec never tells the truth about cloth. The same wool can look stiff on one form and liquid on another. What designers really need is the shape the fabric takes when it hangs, not the number on the bolt.

3D capture of a draped garment gives that shape. It turns “looks about right” into a file the pattern team can measure. The catch is that fabric moves, so the rig must hold it still.

Define the drape question before you scan

• Name the behaviour you are checking. “Nice drape” is not a brief. “Skirt holds an A-line to the knee without clinging” is. Write that down, and the scan has a verdict, not just a picture.
• Separate the form from the fabric. A naked mannequin and a dressed one answer different questions. Scan the form for the base, then scan the cloth on it for the drape. Mix the two and the team measures the wrong thing.
• Pick the drivers that matter. Hem flare, front fall, and side swing decide most fit calls. Choose three per style. After that, the review stops arguing taste and starts reading data.
• Keep the toile as the truth. The mesh supports the room. It never replaces the sewn sample. Say so, or the factory cuts a shape nobody tried on.

Where fabric capture quietly fails

• The cloth shifts during the pass. A light skirt drifts as the air moves. As a result, the scan freezes a wave that never settled, and the hem reads crooked. Weights at the hem and a still room solve most of it.
• Glossy synthetics throw glare. Satin and coated nylon bounce light back, so the scanner reads shine as fold. The surface then looks ridged where it is smooth. A matte mist and a soft tent remove the fake creases.
• Hand rotation never repeats. An operator turning a dressed form by hand shifts the angle each time. Consequently two scans will not register, and the before-and-after fit study fails. Lock the motion, and the shape stays honest.
• Seams get lost in the mesh. A princess seam and a print line fall below the sensor’s depth. Then the pattern team cannot see the true panel edges. Move closer, slow the pass, and raise overlap so the seams survive.
• Colour and form drift apart. A model with the right shape but wrong tone misleads the buyer. Lock white balance per fabric, and bake colour into the same file as geometry, so the room sees the cloth as it is.
• No record of the base form. A dress on form A and a dress on form B look alike in a viewer. Without a tag, the team compares two sizes and calls it one. Mark the form in the filename, and the asset stays comparable.

Build a rig the studio can repeat

• Fix the form and move the view. Handheld sweeps around a mannequin look lively in a reel. Yet they change path every time, so overlap quality swings with the person. A steady stage keeps the path constant, and the drape stays in the data.
• Automate the spin for even coverage. For a hanging garment, even step angles beat raw resolution. Studios that use a programmable turntable, such as the ComXim motorised turntable, get the same angular view on every sample. Therefore the merge needs less repair, and a junior capturer produces a senior-grade file.
• Light it like a studio, not a shop floor. Harsh overhead tubes streak satin and flatten the fall. Use diffuse panels and lock exposure before the first pass. After that, surface sheen stops leaking into the shape.
• Mark the stage, not the garment. A pin on a finished dress scars the very thing you show. A mark on the turntable travels with the spin and still gives the software a stable frame.
• One naming rule for the whole rack. Style, size, fabric, form, date, capturer. Six fields, one order. Six months later that rule separates a library from a rail of unknowns.

Turn scans into pattern decisions

• Compare the drape, do not admire it. A rendered skirt impresses no pattern cutter. A deviation map against the block tells the team where to add sweep. Build the comparison view so it appears automatically.
• Let the digital drape be the source. Physical toiles get pinned, washed, and “nudged” by whoever is nearest. A versioned file stops that drift. Then grading and prototyping pull from one shape.
• Feed the block from the same capture. The silhouette you scanned already holds the fall line. Reuse it rather than re-draping a fresh toile. Fewer handles means fewer chances to change the hang.
• Close the loop with returns. Find which sizes came back, then re-scan those fits. After two seasons the pattern is clear, and your size table reflects real bodies instead of a 1990s chart.
• Keep a golden form for calibration. Three known shapes, scanned each month, expose drift before a buyer notices a crooked hem.

A pre-scan checklist

• Write the drape brief on the job card.
• Weigh the hem and still the room before the pass.
• Matte any glossy cloth with a light mist.
• Lock light and white balance for the whole fabric.
• Run one known form to prove the rig behaves.
• Name every file with the six-field rule.

Drape is a behaviour problem, not a scanner problem. Start with one style, one form, and one clear brief. Once that loop works, the rest of the range follows.