The source of truth

How I calculate every scale conversion

Every percentage on STLACCESS comes from one rule: (target height ÷ source height) × 100. I use a 175cm human baseline with eyes at 160cm. Gaming labels describe eye level and ratio labels are defined from full height, but I compare eye-level equivalents for both, then verify the resulting numbers on my own prints.

Last reviewed August 15, 2026 · Written and tested by me, not generated from a generic scale table.

Skip to content

The baseline I chose

I needed one fixed reference so that 28mm, 32mm, 1:35 and every other value on the site mean the same thing from page to page. I use a 175cm adult human with the eyes at 160cm. That leaves 15cm for the head above the eye line, which matches the 5–10% gap you see between eye-level and top-of-head miniature measurements.

Scale familyDefined fromExampleWhy
Gaming / displayEye level — 160cm28mm, 32mm, 35mm, 40mm, 75mmHow tabletop ranges are conventionally labelled
RatioTop of head — 175cm1:35, 1:48, 1:56, 1:72, 1:100A strict division of the full-height object

Those definitions are not the numbers I divide in the chart. A scale label tells you where the millimetre value comes from; a conversion tells you how tall two figures will look side by side. For that comparison I use the eye-level value in every row: a 1:35 figure is 50.0mm to the crown but 45.7mm to the eyes, so it is 45.7mm that I set against 28mm or 32mm. A 1:48 figure works the same way: 36.5mm to the crown, 33.3mm to the eyes.

All 16 scale baselines are published as JSON and CSV so you can check my inputs rather than take my word for it.

The formula, worked through

(target ÷ source) × 100 = slicer %

I take the eye-level millimetre value of the scale I want, divide it by the eye-level millimetre value of the file I have, and multiply by 100. The result is the number I type into the slicer's uniform scale field with X, Y and Z locked together.

28mm → 32mm

32 ÷ 28 × 100

114.3%

32mm → 28mm

28 ÷ 32 × 100

87.5%

1:35 → 32mm

32 ÷ 45.7 × 100

70.0%

The same rule gives 1:48 → 28mm = 84.1% (28 ÷ 33.3) and 28mm → 1:35 = 163.2% (45.7 ÷ 28). The full set of pairs is on the miniature size chart and in the conversion dataset.

How I verify a number before publishing it

  1. 1

    Confirm the sculptor's convention

    I note the source scale and whether the label is a gaming height or a ratio. If the sculptor's measurement point is missing, I measure a known figure from the same range and compare eye level to eye level.

  2. 2

    Calculate from the baseline

    I work out the percentage from the two standard heights. I do not round until the final step, then publish one decimal place.

  3. 3

    Print and measure

    I print a test piece in PLA or PETG on my Bambu Lab A1 Mini, let it cool, and measure it with digital calipers rather than trusting the slicer's bounding box alone.

  4. 4

    Compare against a trusted model

    I stand the test print next to a factory figure from the target range. If the head, shoulders and base all agree, the percentage stays. If not, I revisit the measurement convention.

What the method does not claim

This is a measurement standard, not a law of miniature sculpting. Heroic proportions, oversized helmets, sculpted bases and manufacturer drift all move a real model away from the tidy number. A conversion within a few percent is usually closer than the labelling on the package.

Thermal shrinkage is handled separately. I do not fold an estimated 0.5% into every result, because that would disguise the actual geometry and make the chart harder to audit. If your material needs compensation, measure a coupon and apply that correction on top of the scale percentage.

I do not sell models, filament or printers, and I do not upload your files. Everything on this site is a tool, a chart or a written explanation you can check against the published data.

Methodology FAQ

Why are gaming scales measured to the eyes instead of the top of the head?

Because that is how the hobby's conventional numbers are used. A 28mm or 32mm miniature is described by the height of a 175cm human's eyes, which sit at 160cm. If you measure to the crown or a helmet you are using a different convention, and mixing the two produces models that are visibly too tall or too short.

If ratio scales are defined from full height, why do the conversions use eye level?

A ratio label is a strict fraction of the real object: 1:35 means a 175cm human becomes 50.0mm to the top of the head. But that same figure's eyes sit at 45.7mm. Since gaming scales are labelled to the eyes, I compare eye-level equivalents on both sides. Comparing 50.0mm to 28mm would make the 1:35 figure visibly too small next to a true 28mm model.

What percentage do I type for 28mm to 32mm?

114.3%. The calculation is 32 ÷ 28 × 100. If you are measuring both models to the top of the head instead, the variant is 113.1% (34.5 ÷ 30.5 × 100). Use the convention the sculptor used, not the one that happens to be easiest to measure.

Does shrinkage change the conversion percentage?

Usually not enough to matter. PLA and PETG move by a few tenths of a percent as they cool, which is smaller than normal variation between slicers, nozzles and filaments. I only add compensation after measuring a real test piece, and I document that separately rather than hiding it inside the headline percentage.

Are these percentages exact for every sculpt?

No. They are exact for the stated baseline and measurement convention. Sculptors exaggerate heads, helmets and bases, and manufacturers drift between ranges. Treat the chart as a starting point, then compare a printed test model against a figure you already trust.

Where to go next