Skip to content

Filament colour consistency: lot variation, ΔE and matching

Why can two spools with the same colour name look different side by side? How filament colour is made and measured, and how to keep the same look for months.

AMENT Research TeamPublished 10 min read

Short answer

Filament colour can shift slightly from lot to lot because of the pigment or masterbatch batch, the base polymer's own tint, dosing and processing temperature. The difference is measured as ΔE in the CIELAB colour space: as a rough rule, below ΔE 1 is imperceptible, 1 to 2 is barely noticeable on close inspection, and 3 or more is clearly visible. To keep the same look, record lot numbers, keep a reference print and run important jobs from a single lot.

Key takeaways

  • Most colour drift comes not from the pigment itself but from small changes in the recipe and process around it.
  • ΔE is the distance between two colours; around ΔE 1 is usually invisible, ΔE 3 is a difference most people will see.
  • Ask which ΔE formula is used: the same two colours give different numbers under different formulas.
  • Matte, silk and translucent surfaces look different even with the same pigment; keep light and layer height constant.
  • On a print farm, colour consistency is a stock-planning job: lot records, a reference part and an agreed tolerance.

Why filament colour varies from lot to lot

Filament gets its colour from pigments mixed into the molten polymer. In production the pigment is usually added not as loose powder but as masterbatch: pellets that carry a high concentration of pigment in a carrier polymer. The line blends this masterbatch into the base polymer at a set ratio to reach the target colour.

Every link in that chain is a small source of variation. A colour difference rarely comes from one big mistake; it is usually several small shifts stacking up:

  • The masterbatch batch: two batches made to the same recipe can still differ slightly in pigment strength.
  • The base polymer's tint: natural PLA and PETG are not perfectly clear or white; they can carry a faint yellowish or greyish undertone that varies by batch.
  • Dosing: a small drift in the masterbatch ratio shows up most in light and pastel colours.
  • Thermal history: high temperatures and long dwell times can slightly yellow some polymers and pigments.
  • Changeovers: if the line isn't fully purged between colours, the first metres can carry a trace of the previous one.

A good maker can't remove these variables, but it measures and limits them. Prusament, for example, says each of its colours is locked to a documented spectroscopic recipe and that every batch is tested against that standard before it ships. The useful question isn't “does the colour never change”, but “how much does it change, and who measures it, and how”.

How colour is measured: CIELAB and ΔE

The human eye is good at comparing colours and poor at putting numbers on them. Industry therefore measures colour with a spectrophotometer and usually expresses it in the CIELAB (L*a*b*) colour space. Every colour there has three coordinates: L* for lightness (black to white), a* from green to red, and b* from blue to yellow.

+a*−a*+b*−b*a*: green ↔ red · b*: blue ↔ yellow · L*: lightnessClose upΔE 3ΔE 1referencesampleL*
In CIELAB a colour is defined on three axes. ΔE is the distance between reference and sample; the rings in the close-up mark distances of 1 and 3 units.

The difference between two colours is given as ΔE (delta E): the distance between the reference colour and the sample. The smaller the ΔE, the closer the two colours. The measurement is usually taken not on the thin, round strand but on a flat sample, a printed or moulded colour chip, so that light reflects the same way every time.

Which ΔE: 1976, 1994 and 2000

There is no single ΔE formula. The original 1976 formula (ΔE*ab) is the straight-line distance between two points. But CIELAB isn't perfectly uniform to the eye: we notice small differences more easily in some regions than in others. Later formulas (CIE94 and CIEDE2000) add weighting factors to correct for this; according to Datacolor's overview, CIEDE2000 matches human perception of small colour differences better.

What a ΔE of 1 versus 3 actually means

There are widely used rough thresholds for turning a ΔE number into what the eye sees. Konica Minolta's quality-control guide sums them up like this:

Rough perception thresholds for ΔE (Konica Minolta)
ΔEHow it looksWhat it means in practice
Below 1ImperceptibleParts side by side look the same
1 to 2Barely noticeable on close inspectionAcceptable for most uses
3 or moreClearly visibleUnacceptable in many industries
What ΔE means to the eyeRough thresholds. Each pair is drawn on screen at a computed ΔE from the reference.ΔE 0.5ΔE 1ΔE 2ΔE 3ΔE 5012345ImperceptibleBarely noticeableClearly visibleSource: Konica Minolta · Screen colours vary by device
In each pair the left half is the reference and the right half a sample at a computed ΔE. Screens render colour differently, so treat this as a rough impression.

These thresholds are a starting point, not a rule. The same ΔE is perceived differently in different parts of colour space; Datacolor notes that the eye is more sensitive to changes in chroma than in lightness. And small differences are far easier to see when parts sit side by side than when they are in different rooms.

In practice, the real question is whether the parts touch. For a single-colour figurine, ΔE 2 may bother no one. If two halves of one product are printed from different spools and joined, or the same item is lined up on a shop shelf, the same difference can jump out.

Matte, silk and translucent: why one colour looks different

Colour isn't only pigment; the eye sees whatever light the surface sends back. The same pigment can leave a very different impression on a different surface.

Same pigment, three different looksMatteScatters lightSilkA bright bandTranslucentDepends on thickness
A rendering of one pigment on three surfaces: matte scatters light, silk reflects a narrow bright band, and in translucent material the light behind and the thickness change the colour.
  • Matte: microscopic roughness scatters light in all directions. The colour looks softer and a little lighter, and layer lines draw less attention.
  • Silk: the smooth, glossy surface reflects light strongly at one angle. The same part can look bright or dark depending on where you stand, which makes silk colours hard to compare against a fixed reference.
  • Translucent: light enters the part. Wall thickness, infill pattern and the light behind it change the colour directly: thin walls look light, thick walls look dark.

That is also why the colour on the spool and the colour in the print differ. On the spool the filament is a thick, stacked mass and usually looks darker and more saturated. When choosing a colour or comparing two lots, look at a printed sample, not at the spool.

Layer height, light and print settings

Layer height affects colour too. Thick layers cast more pronounced shadows at the edge of each layer, and the surface can look slightly darker. Thin layers reflect light more evenly. Print temperature and speed also change surface gloss and so shift how the colour reads; when comparing two parts, use the same print settings.

The light source is the most forgotten variable. Two colours can match in daylight and split apart under a warm LED; this is called metamerism. Always compare colours under the same light, ideally a steady source close to daylight.

Matching colour on reorders

The first time you print a product, its colour is whatever comes off the spool. When you print it again three months later, you need a reference to compare against. A simple routine prevents most surprises:

  1. 1

    Record the lot number

    Note each spool's lot number along with the job you used it for. When you see a colour difference, knowing where it came from is half the answer.

  2. 2

    Keep a reference part

    For each colour and lot, print a small flat test chip and store it in a closed box, away from light. Parts left in the sun can fade; the reference itself must not change. The same goes for spools: closed, dry and dark. Details are in the storage and drying guide.

  3. 3

    Compare the new spool

    Print the same chip from the new spool with the same settings. Compare the two under the same light, side by side and from a few angles.

  4. 4

    Run important jobs from one lot

    Print an order for one customer, or parts that will be joined, from a single lot. If needed, set aside enough spools from the same lot before you start.

  5. 5

    Mix lots deliberately

    If you must use two lots, put one on hidden internal parts and the other on visible surfaces. Avoid switching lots on the same visible face.

What sits behind a lot number is explained in our lot data guide. What to check when a spool arrives is in our filament quality checklist.

The AMENT palette: places in Türkiye

AMENT's colour names come from places in Türkiye, Samsun first: Karadeniz, Kunduz and Bafra. The rest of the palette is Pamukkale, Nemrut, Kapadokya, Sümela, Uzungöl, Efes, Ağrı, Van and Salda. Today these colours are a concept; they are being prepared for launch and the products are not on sale yet.

Colour consistency is part of the quality system we are building: a colour check against a reference sample for every lot, and a record of colour pigment batches, are among our goals. The measurement method and results will be published with production. You can see the planned products and colours on the materials page.

Frequently asked questions

Why do two spools of the same filament colour look different?

If the spools come from different lots, the pigment or masterbatch batch, the base polymer's tint and the processing temperature can all create small differences. If they come from the same lot, the difference usually comes from print settings, layer height or lighting.

What is delta E (ΔE)?

ΔE is the distance between two colours in the CIELAB colour space. As a rough rule, below ΔE 1 is imperceptible, 1 to 2 is barely noticeable on close inspection, and 3 or more is clearly visible. What the number means depends on the formula used, such as CIEDE2000, and on the measurement conditions.

How do print farms avoid colour differences between filament batches?

Record each spool's lot number, keep a reference test chip for every colour, and set aside enough stock from a single lot for products you will print over a long period. Before a new lot goes into use, compare it with the reference under the same light.

Why does matte and silk filament look different in the same colour?

A matte surface scatters light in every direction and makes the colour look soft. A silk surface reflects light strongly at one angle, so it looks bright or dark depending on the viewing angle. Even with the same pigment, the surface treats light differently.

Are AMENT filament colours on sale?

No. AMENT's colours are a concept palette named after places in Türkiye; they are being prepared for launch and the products are not on sale yet. Colour consistency is one of the goals of the quality system we are building.

Sources

  1. 1What Is Delta E (ΔE)? Understanding Colour Difference in Quality Control (Konica Minolta Sensing Europe)
  2. 2CIE ΔE* (Color Difference) Equations (Techkon / Datacolor)
  3. 3Prusament PLA product page: colour recipes and batch testing (Prusa Research)

Hear about AMENT filaments first

Our filaments are pre-launch. Be the first to know when measured technical data is published.