Woven and Knitted Fabrics: How Each Is Made, and Which Your Product Needs

The two ways cloth is made

A woven fabric is built on a loom. One set of yarns, the warp, is held under tension running the length of the fabric; a second set, the weft, is passed across it, over and under the warp yarns at right angles. The result is a grid. Because the threads are held in tension against each other rather than curled around themselves, the structure barely moves: pull a woven fabric across its width and it resists, which is exactly why a formal shirt keeps its collar line all day.

A knitted fabric is built from loops. A single continuous yarn is drawn through previous loops by a bank of needles, forming interlocking rows. Nothing is under tension against anything else — the loops themselves can flatten and extend — so the fabric stretches in every direction and recovers. That is why a t-shirt pulls over your head without a single fastening, and why a woven shirt of the same dimensions needs a full placket of buttons.

Everything else in this guide follows from those two paragraphs. Stretch, drape, dimensional stability, how a piece is cut, what it costs to sew, how it frays, how it is printed, even which factory can make it — all of it descends from whether the cloth is a grid or a field of loops. Understanding the mechanism means you stop memorising which fabric does what and start predicting it.

The vocabulary is worth fixing early, because it is used loosely in the clothing industry and precisely in textile engineering. Woven and knitted describe construction, not fibre. Cotton can be either. Wool can be either. Polyester can be either. A cotton shirt and a cotton t-shirt share a fibre and share almost nothing else, and confusing fibre with construction is the single most common mistake in a first fabric brief.

There is a commercial layer to the same distinction. The two constructions are made on different machines, in different factories, by people with different skills, and they are bought on different minimums and different lead times. A brand adding a woven shirt to a knit range is not adding a style — it is adding a supply chain. Knowing that before the season is designed is worth more than any amount of fabric vocabulary, and it is why this guide spends as long on manufacturing as it does on structure.

Inside a woven fabric: warp, weft and weave structure

Warp yarns run the length of the roll and are loaded onto the loom first, under constant tension. They take the most stress during weaving, so they are usually the stronger, more tightly twisted yarn. Weft yarns — sometimes called filling — run across the width, interlacing with the warp as the loom cycles. When a garment is described as having grain, the grain is the warp direction, and cutting a pattern off-grain is what makes a finished shirt twist on the body after a wash.

The pattern in which warp and weft cross defines the weave, and there are three foundational ones from which nearly everything else is derived.

A plain weave alternates one over, one under, across every row. It is the simplest structure, the most stable, and the most matte in appearance. Poplin, chambray, canvas, muslin and chiffon are all plain weaves at different yarn weights and densities — which is a good illustration of how far a single structure stretches, since chiffon is a barely-there evening fabric and canvas is stiff enough to build a jacket from.

A twill weave passes the weft over two or more warp yarns then under one, stepping the offset on each row. That step produces the diagonal line visible on the face of chinos, gabardine and drill. Twills drape more softly than plain weaves at the same weight, hide soil and creasing better, and are more durable in abrasion because the longer floats let the yarns shift rather than tear. Denim is the twill everyone knows, though it is not a fabric we produce.

A satin weave floats the weft over four or more warp yarns before tucking under one. Very few interlacing points means very little to interrupt the surface, so light reflects evenly and the fabric reads as lustrous. That same long float is the weakness — satin snags easily, because there is more exposed thread to catch.

Beyond the three, dobby and jacquard looms vary the interlacing to weave patterns directly into the cloth rather than printing them on. It is more expensive, and it is why a woven-in stripe survives the life of the garment while a printed one fades.

Two numbers describe a woven fabric beyond its weave, and buyers routinely confuse them. Yarn count measures how fine the thread is — counter-intuitively a higher count means a finer yarn, so a 100s poplin is lighter and smoother than a 40s. Thread count is how many warp and weft yarns sit in a square inch, which is a density measure rather than a quality one: a high thread count in a coarse yarn tells you very little on its own. Weight in grams per square metre is the figure that actually predicts how a shirt feels in the hand, which is why shirting specifications are written in GSM and why a spec quoting only a thread count is incomplete.

Three structural terms are worth knowing because they turn up in production problems rather than in marketing. The selvedge is the self-finished edge running down both sides of the roll, formed where the weft turns back on itself; it does not fray, and it is the reference edge a cutting room squares everything against. The grain is the warp direction, and a pattern piece cut off-grain produces a garment that twists on the body — a shirt side seam migrating towards the front after two washes is almost always this. The bias is the 45-degree diagonal, the one direction in which a woven fabric genuinely stretches and drapes, which is why bias-cut dresses hang the way they do and why bias binding curves around a neckline that straight-grain tape would pucker.

Inside a knitted fabric: courses, wales and the two knitting families

A knitted fabric is described by courses, which are the horizontal rows of loops, and wales, the vertical columns. Loop size and the gauge of the machine — how many needles per inch — set the density, the weight and the hand. Knits divide into two families that behave quite differently.

Weft knitting builds the fabric one row at a time from a single yarn travelling across the needles. It is what circular knitting machines do, and it produces nearly all the jersey in the world. It stretches generously in both directions, and because the whole fabric is one continuous yarn, a broken loop can run — the ladder that travels down a stocking.

Warp knitting feeds a separate yarn to every needle and interlocks them vertically as the fabric advances. It is faster, more stable, will not ladder, and produces tricot, mesh and most performance and lingerie fabrics. It stretches less than weft knit, which is often the point.

Within weft knitting, the structures worth naming by name are the ones that appear on a spec sheet. Single jersey is the basic one-face knit: smooth on the front, looped on the back, light and fluid, and the standard construction for a t shirt. Interlock is two single jerseys knitted together so both faces are smooth — heavier, more stable, less prone to curling at a cut edge, and the usual choice for polos and premium tees. Rib alternates knit and purl columns so it contracts across its width and recovers, which is why every cuff and hem band on a sweatshirt is rib. French terry leaves uncut loops on the inside face, and brushing those loops open turns the same base structure into fleece.

Sweater knitwear sits slightly apart. A cable knit or a fine-gauge cardigan is made on a flat-bed machine that knits each panel to shape rather than cutting it from a roll, which is a different process, a different factory and a much higher minimum — 500 pieces per style against 50 for cut-and-sew knitwear.

Gauge is the number that decides how a knit reads. It counts needles per inch on the machine, so a high gauge produces a fine, dense, smooth fabric and a low gauge a chunky one with visible loops. It is not the same as weight, though the two move together: a coarse-gauge single jersey and a fine-gauge one at the same GSM feel quite different, one slightly rustic and the other almost papery. A usable specification needs gauge, weight and composition together, because any two of the three still leave the fabric ambiguous.

Two behaviours are unique to knits and both show up on a production floor. Curling is the tendency of single jersey to roll at a raw edge, because the loops on the two faces sit under different tension — it is why a jersey hem is turned and twin-needled rather than left raw. Laddering is the run that travels down a weft-knitted fabric when one loop breaks, since the whole structure is a single continuous yarn. Warp knits do not ladder, which is exactly why tricot and mesh are chosen where a run would be unacceptable.

How a knit garment is assembled splits the same way sweaters do. Cut-and-sew knitwear takes fabric off a roll, spreads it, cuts pattern pieces and sews them — every t shirt and hoodie in the world — and it runs on low minimums because the fabric is a commodity input. Fully fashioned knitwear knits each panel to its finished shape with the shaping built into the knitting, so there is almost no waste and no cut edge, but the machine time per garment is far higher. The visible tell is a fine chain of stitching where panels meet on a good sweater, against the plain overlocked seam of a cut sweatshirt.

How the two behave differently, and why

Every practical difference between woven garments and knitwear traces back to the grid and the loop. Setting them side by side is the fastest way to build the intuition.

The named fabrics on each side

Common examples are worth learning as a group, because a fabric name usually encodes both a structure and a weight, and recognising the name saves a paragraph of specification.

Manufacturing consequences: two different factories

A knit factory and a woven factory are not the same building, and in most of the industry they are not the same company. Knitwear runs from a knitting hall to cutting to sewing lines optimised for few operations and high throughput. Woven production is a longer, more skilled chain — fabric inspection, careful spreading on grain, cutting to tighter tolerance, then sewing lines built around collars, plackets, cuffs and pockets, followed by pressing that is itself a craft.

That split has three consequences a brand feels directly. Minimums differ, because a woven run carries a separate fabric and cutting setup that does not amortise over a small order: our knitwear starts at 50 pieces per style and woven garments at 300. Lead times differ, because Bangladesh has a deep domestic knit mill base while woven fabric is more often imported, which can add weeks at the front of the schedule. And sampling matters more on wovens, since a non-stretch fabric shows every fit error a jersey would quietly absorb.

Three stages in a woven factory have no real equivalent on a knit floor, and they are where the extra cost and the extra time live. Fabric inspection comes first: woven rolls are graded before cutting, usually on the four-point system, where faults are scored by size and a roll above a threshold per hundred square yards is rejected. Knit fabric is inspected too, but a flaw in jersey is often absorbed by the cut, while a slub or a missing warp end runs the entire length of a woven roll and can write off a whole lay.

Then the cutting room. Woven fabric has to be spread flat, square to the selvedge and completely relaxed, because tension left in the lay becomes a twisted garment. The marker — the arrangement of pattern pieces on the fabric — has to respect grain on every piece, and on a stripe or a check it also has to match across seams, which can push fabric consumption up by a fifth. Knit fabric is spread relaxed too, but with no grain rule as unforgiving and rarely a pattern to match, so the marker is tighter and the wastage lower.

Finally, pressing. On a woven garment, pressing is construction rather than tidying: collars, cuffs, plackets and creases are set with heat and steam at several points during assembly, not once at the end, and a skilled presser is as much a part of shirt quality as the sewing operator. Knitwear is steamed and relaxed rather than pressed into shape, because the structure will simply spring back. This is the single most underestimated line in a woven quotation.

None of that is visible in a quotation, which is why the useful conversation in apparel manufacturing is with a production manager rather than a sales desk. Ask how the style will be laid, how many operations the sewing line will run, and what the fabric consumption per garment works out at. Those three answers predict the real price of bulk production far better than a unit rate does, and they tell you whether the factory has planned your garment or simply priced something that resembles it.

The practical answer for a brand running both is one point of contact across the two. We knit and sew knitwear on our own floor and place woven garments with audited partner factories, managed by our team against a single quality standard, so a range spanning both constructions ships together rather than arriving in two shipments from two suppliers with two different navies.

We produce both constructions — knitwear on our own floor and wovens through audited partner factories, which is set out on our page for woven garment manufacturing.

Decoration, trims and finishing

Printing behaves differently on each. A flat, stable woven surface takes a screen print cleanly and holds fine detail, but it will not tolerate the stretch a knit does, so a print that spans a seam or a stress point needs care. Knits move under the squeegee and move again in wear, so inks have to flex — a rigid ink on jersey cracks along the fold line within a season. Sublimation is a special case that ignores the argument entirely: it works on polyester regardless of structure, because the dye becomes part of the fibre.

Embroidery is the reverse. It sits happily on a woven shirt because the base is stable enough to support the stitch density; on a light single jersey the same design puckers unless it is backed properly. Any factory quoting embroidery on a lightweight knit without mentioning backing has not thought about it.

Trims and accessories follow the structure too. Wovens need interfacing at collars, cuffs and plackets to give those parts their body, and buttons, zips and pocket bags all belong in the bill of materials. Knits need far fewer components, which is a real part of why they cost less. Labels are the exception that behaves identically: a woven label is a woven label whether it is stitched into a shirt or a hoodie.

Both constructions can also be treated after they are made, and the finish often matters as much as the structure. Mercerising swells cotton fibers so they take dye more brilliantly and gain lustre. Peaching and sueding raise a fine nap for softness. Enzyme and stone washes break the surface down for a worn hand. Calendering presses a woven flat and glossy, while anti-pill finishes stop a knit surface bobbling. A fabric named on a spec sheet without its finish is only half specified — two rolls of the same apparel-grade twill can feel entirely unalike depending on what was done to them afterwards.

Care behaves differently enough to be worth writing onto the label deliberately rather than copying from a competitor. Wovens hold their dimensions but crease, so they are pressed or hung; a cotton poplin shirt that is never ironed simply looks wrong, and that is a property of the grid rather than a fault. Knits resist creasing but relax and can drop out of shape under their own weight, so a heavy sweater is folded rather than hung and a jersey tee is dried flat if it matters. Shrinkage runs higher on knits for the same reason they stretch — the loops close up in heat — which is why knit measurements should be specified after wash and woven measurements can usually be specified as finished.

Choosing the right construction for your product

Work from the garment backwards, not from the material forwards. Three questions settle it almost every time.

Does it need to move with the body? If the answer is yes and there is no fastening — a tee, a hoodie, activewear, anything pulled over the head — the answer is a knit, and the only remaining decision is weight and structure. Does it need to hold an imposed shape? A collar that stands, a trouser crease, a jacket that keeps its line on a hanger: that is a woven, and no amount of heavy knit substitutes for it. Does it need to survive abrasion rather than flexing? Workwear, outerwear and anything worn against a surface points to a dense woven.

Two rules of thumb are commonly reliable. Anything worn next to the skin and expected to move is suitable for a knit; anything expected to hold a line is suitable for a woven. Where a garment has to do both — a shirt that must not crease but must also move — the essential decision is not the construction at all but whether to add elastane to it, which buys stretch in a woven at a small cost in recovery and a real one in fabric price.

Cost and volume then filter what is left. A knit range is cheaper per unit, reaches a small minimum, and can be sampled quickly, which is why almost every new brand starts there and adds wovens once a core range is selling. Budget, order volume and how fast you need to respond to customers matter as much as the material properties — the right product is the one you can actually fund, produce and reorder.

A worked example makes the filter concrete. A brand launching a small streetwear capsule wants a heavyweight tee, a hoodie, a cap and an overshirt. The first three are knit or headwear and all sit on a 50-piece minimum, so the whole capsule can be tested on a few hundred units. The overshirt is woven, needs 300 pieces of one style on its own, and adds imported fabric to the front of the schedule. The right product decision is usually to launch the three, prove the customers, and bring the overshirt in on the second order once the range has earned the commitment — not because the overshirt is a bad idea, but because it is a different kind of bet.

One more consideration for anyone selling into the European Union or other regulated markets: fibre composition labelling is a legal requirement, and it is stated by fibre, not by construction. Your care and content label has to name the fibres and their percentages whether the fabric is woven or knitted, and it is far easier to collect that from your supplier at sampling than to chase it a week before shipment.

A short history, and why both survive

Weaving is ancient and knitting is not. Woven cloth predates written history by thousands of years, and the loom was mechanised early enough to become the engine of the first industrial revolution. Hand knitting appears comparatively late, and machine knitting later still — the stocking frame in the sixteenth century, then circular knitting machines in the nineteenth, which is the point at which knitted cloth became something a whole industry could be built on.

The mechanisation matters more than the dates. A loom is a machine for holding many threads in tension and lifting a subset of them in sequence, which is a problem that yields to gears and punched cards — the jacquard loom was programmable a century and a half before computing, and it is the reason woven pattern is cheap at scale. A knitting machine has to form and transfer loops without dropping them, a mechanically fussier job, which is why knitted cloth stayed a cottage product for so much longer and why a modern circular knitting machine is a more delicate thing than a loom.

That sequence explains a lot about the textile industry today. Formal clothing is woven because formal clothing was codified when woven was the only option; the t shirt is a twentieth-century garment because affordable jersey is a twentieth-century material. Neither structure is being replaced. They solve different problems, and a wardrobe or a collection with only one of them is missing half the answer.