Lessons · Lesson 2 of 6
What the simulation needs from the factory
The four inputs a virtual sample runs on, what it costs a factory to produce the one that is usually missing, and how many times that work gets reused.
Lesson 2 of 6 · 19 min
Four inputs, and only one of them is normally missing
A virtual sample of KA-7318 runs on four things, and a factory has to produce all four.
- The pattern, as clean digital pieces with grain, notches and internal lines. The CAD course in this track owns how that is made. Here it is simply an input, and a factory already doing digital pattern work has it.
- The assembly: which edge is sewn to which edge, in what order, with what seam and what ease. People underestimate this part. It is not a picture of a construction. It is a specification of one, and it is where a merchandiser's knowledge of the garment goes into the model.
- A body. The avatar has to match the buyer's size specification, not a generic mannequin. Sizing and virtual fit belong to their own course in this track and are not re-taught here.
- The cloth, as numbers.
Nine factories out of ten have the first three within a month of buying the system, and never finish the fourth. It is the input that decides whether the middle band of lesson 1, the 23.5% about how the cloth behaves, can be answered at all.
Measuring a cloth is a bench job, not a laboratory job
The properties a simulation wants are not the ones on a mill's specification sheet. Composition, construction and weight tell you what the cloth is. A simulation needs to know what it does under very small forces, because that is how a hanging garment lives. How far it extends under a light pull along the warp, along the weft and on the bias. How easily it bends. How it distorts when pushed out of square. Its thickness. Its weight. And how much its surface grips itself.
Kerem measures those on a bench in the sample room, with a rig the factory bought outright, in about ninety minutes per cloth.
| Test strips, warp, weft and bias, 1.4 m at USD 5.20 | USD 7.28 |
| Technician, ninety-five minutes at USD 6.40 an hour | USD 10.13 |
| The measuring rig, USD 9,600 over five years across the fabrics it reads in a year | USD 40.00 |
| Entering and sanity-checking the record, forty minutes at USD 11.50 an hour | USD 7.67 |
| Validation panel, 0.7 m of cloth | USD 3.64 |
| Sewing and hanging that panel, twenty-five minutes at USD 4.90 an hour | USD 2.04 |
| One measured fabric | USD 70.76 |
Two lines in that table are worth defending at your own factory.
The rig is the largest line, and it is the one that falls fastest. It is a fixed cost divided by the number of fabrics you read. Kalthoum assumed forty-eight cloths a year, which is modest. A factory reading double that pays half as much per fabric, and the whole table falls by more than a quarter.
The validation panel is not optional, and it is the cheapest insurance in this course. It is a sewn rectangle of the real cloth, hung beside the simulated one and photographed. It takes twenty-five minutes, and it is the only thing standing between you and lesson 6.
The work that gets reused, and the work that gets used up
Here is the structural difference between a physical sample and a digital one. It is the reason the arithmetic in lesson 6 works at all.
A physical sample is used up by the question it answers. Sew a proto to check a pocket position and you have a garment that answers a pocket position. The next style starts from nothing.
A digital input is not used up. The Rustan viscose-linen record served four styles in Aldemere's autumn range, so its cost per style was USD 17.69. The shirt block underneath KA-7318 served three. The avatar served all eleven.
That is the whole economic case in one sentence: the factory's digital inputs are shared across many styles and its physical samples are not. Everything else in this course follows from it.
The per-style bill
Once the four inputs exist, putting one style through costs hours, not materials.
| Hours | At USD 11.50 | |
|---|---|---|
| Pattern prepared for simulation: seam lines, internal geometry, clean-up | 3.1 | USD 35.65 |
| Assembly, ease and avatar set-up | 1.6 | USD 18.40 |
| Trims, colour, print placement, rendering | 1.2 | USD 13.80 |
| Simulation runs and the iteration in between | 2.4 | USD 27.60 |
| First virtual sample | 8.3 | USD 95.45 |
| A later virtual round on the same style | 1.9 | USD 21.85 |
| A style taken through two virtual rounds | USD 117.30 |
The one-off bill, and what is deliberately not in it
| Hours | At USD 11.50 | |
|---|---|---|
| Avatar built to Aldemere's size specification and checked against a fitted garment | 22 | USD 253.00 |
| Digital blocks for the four base silhouettes in the range | 46 | USD 529.00 |
| System configuration and the factory's own library structure | 14 | USD 161.00 |
| Training, two people | 96 | USD 1,104.00 |
| One-off | 178 | USD 2,047.00 |
Read what is missing from that table, because it matters. There is no software licence in it and no hardware. That is deliberate. Those prices move, they differ enormously by how a vendor packages them, and a figure published in a course gets quoted back three seasons later when it is wrong. Every price above belongs to Kalthoum's own people and its own bench, and you can rebuild the table from your own payroll in an afternoon.
So the break-even style count in lesson 6 is a floor, not an estimate. Add your own licence and your own workstations to the one-off column and the true figure rises. It rises by an amount you can work out, and this course cannot.
Check yourselfYour mill sends a specification sheet with composition, construction, weight and shrinkage. Can your 3D team simulate that cloth?Show the answer
No. Not one of those is a mechanical property in the sense a simulation needs. Composition and construction describe what the cloth is made of and how. Weight and shrinkage describe outcomes. None of them tells the software how far the cloth extends under a light load, how easily it bends, or how it distorts. Two cloths with identical specification sheets can hang completely differently, which is the same finding the knitwear course makes about retractive force. The mill's sheet is necessary for buying the cloth. It is not enough for simulating it. Somebody has to put a piece of it on a bench.
What to take away
- Four inputs: pattern, assembly, body, cloth. The first three arrive with the system. The cloth is the one that gets skipped, and it decides whether a screen can answer how a garment hangs.
- A mill specification sheet is not a fabric measurement. The simulation needs low-load mechanical behaviour, not composition and weight.
- Measuring one cloth at Kalthoum cost USD 70.76: a bench, a technician, ninety minutes and a sewn validation panel.
- That cost is divided by the styles that use the cloth. Across four styles it was USD 17.69 each, and the reuse rate is the single most important number in your own business case.
- Physical samples are used up by the question they answer. Digital inputs are not. That difference, and nothing else, is what makes the programme pay.
- The steady-state figure of 8.3 hours a style is where you finish, not where you start. The one-off table deliberately leaves out licence and hardware, so treat every break-even in this course as a floor.