Stitch pitch around curves and corners
Whether stitch spacing holds where a seam turns, and whether the line keeps its distance from the edge.
The reading sequence
4 steps, starting at surface and finish and running to read the same curve on both sides. Evidence on this point is read at E1, and 1 of the 4 steps carries a condition that ends the assessment outright.
Count two spans of one seam
Get a straight count and a curved count from the same seam, so any change in pitch can be measured instead of sensed.
Evidence: E1 — one macro per span with a ruler in frame, both taken from the same distance and the same angle
- Both spans arrive with the ruler in frame and the same stitch type visible.
- Over 20 mm the curved span carries one stitch fewer at most than the straight span.
- The two spans are shot from different distances, so the counts cannot be compared directly.
- One span sits on the flap and the other on the gusset, where two machines may have been used.
- The curve carries about half the stitches of the straight run across the same 20 mm.
- No ruler is in frame and the seller will not re-shoot with one.
Follow the distance from the edge
Decide whether the stitch line holds a constant margin from the panel edge as the seam turns.
Evidence: E1 — a frame that follows the whole curve, taken square to the surface rather than along it
- The margin between edge and stitch line stays within about half a millimetre along the curve.
- The curve is smooth, with no flat spot where the sewer paused and restarted.
- The margin widens by about a millimetre at the sharpest point of the turn.
- The frame is taken at an angle that makes the margin look narrower near the top.
- The stitch line crosses the edge or leaves the panel at any point on the curve.
- The margin doubles at the corner, so the seam allowance is visibly uneven.
Search the curve stitch by stitch
Find stitch faults that survive a colour match and will not be corrected by conditioning the leather.
Evidence: E1 — a macro of the full curve at the highest resolution the seller can send, read one stitch at a time
- Every stitch is present, and none sits on top of another.
- The thread lies into the leather rather than standing above the surrounding surface.
- One stitch is doubled where the machine restarted, at a point hidden while the bag is carried.
- Two stitches sit deeper than their neighbours at a place where the material thickens.
- A gap longer than two stitch lengths interrupts the run.
- Thread stands proud along a section, which means the tension was wrong for that whole pass.
Read the same curve on both sides
Use the bag's own symmetry as a second reference, which catches uneven work that a single frame cannot show.
Evidence: E1 — two frames of the left and right curves taken at the same distance and the same angle to the surface
- Counts from the two sides agree within one stitch over the same 20 mm.
- The two curves mirror each other, with the same radius to the eye and the same margin from the edge.
- The two sides are shot at different angles, so the comparison stays loose.
- One curve sits on a pocket flap and the other on the body, where a difference is built in.
- The two sides differ by three stitches or more across the same span.
- One curve is visibly flatter than the other, so the panel was cut to a different shape.
Why the thresholds sit where they do
Stitch pitch is the cheapest workmanship question on a bag to answer with a number, and the easiest to argue about without one. A 20 mm span is long enough to show a machine's rhythm and short enough to count from any macro frame that includes a ruler. The one-stitch allowance on a curve reflects how sewing machines behave: pitch shortens as the work turns, so demanding identical counts on a straight run and a corner would fail almost every correctly made bag. The half-millimetre margin band was chosen after measuring the same seam on several photographs, where changing the camera angle moved the apparent margin by roughly that amount. Symmetry is the fallback when one side is photographed badly, because a maker who sews both sides on one machine produces two curves that agree.
Common misreads
2 pairs recorded for this point. Each runs in both directions: 1 finding that looks worse than it is, and 1 that looks better.
- looks like Return actually Pass
Stitches that crowd together at a corner. Every machine shortens its pitch as it turns, so a tighter corner is normal. The fault is a change that happens in the middle of a straight run.
- looks like Pass actually Return
One curve, clean and well lit, with the other side never photographed. The two sides can differ by three stitches and one panel can be cut flatter, and a single frame shows neither.
Check this point against what you were sent
Insufficient evidence
Nothing decided on curve stitching evenness yet: 0 of 4 steps have been read.
0 of 4 steps judged
| Step | Reading | Evidence | Stop met |
|---|---|---|---|
| 1. Count two spans of one seamneeds E1 | — | ||
| 2. Follow the distance from the edgeneeds E1 | — | ||
| 3. Search the curve stitch by stitchneeds E1 | A gap of more than two stitch lengths sits on a seam that carries load. Each remaining stitch then takes more force, and the seam opens outward from that gap. | ||
| 4. Read the same curve on both sidesneeds E1 | — |
Sources and checks
- Public community thread — Photo sets where stitching was counted and the two sides disagreedn=49
- Editor analysis — 20 mm span, one-stitch curve allowance and half-millimetre margin band set togetherno sample
Last checked · next review due · status current.
Return to Bags: pre-purchase judging points, or read how verdicts are defined on the method page.