The same CIEDE2000 result is reproduced when the identical two opaque #RRGGBB sRGB inputs are compared using the documented D50 conversion path and unchanged default weighting parameters kL = kC = kH = 1.
That makes repeatability a matter of controlling the inputs, calculation method, and saved record. It does not make the on-screen swatches colorimetric evidence or guarantee that two printed, painted, or photographed objects will appear identical. The Color Difference Calculator provides the numerical comparison; your documentation supplies the conditions needed to reproduce the decision later.

What must stay fixed for a repeatable result
A CIEDE2000 result is calculated from two specific color coordinates. Changing either coordinate changes the input to the formula, while changing the conversion or formula method can produce a different value even when the displayed HEX colors look unchanged. For a repeatable workflow, keep the following controls constant:
| Control | Value to keep fixed | Why it affects repeatability |
|---|---|---|
| Input colors | The same two exact #RRGGBB values | Each hexadecimal coordinate represents a specific sRGB color and therefore affects the calculation. |
| Opacity and color space | Opaque sRGB | The calculator deliberately rejects formats that would require hidden parsing or compositing assumptions. |
| Lab reference | D50-adapted CIELAB | The sRGB values pass through D65 XYZ and a Bradford D65-to-D50 adaptation before Lab coordinates are calculated. |
| Difference formula | CIEDE2000 | CIEDE2000 is not a straight Euclidean distance between Lab coordinates. |
| Parametric weights | kL = kC = kH = 1 | These are the calculator’s common default weights and should not be confused with a different metric or weighting setup. |
| Calculation precision | Unrounded CIELAB values | Visible Lab coordinates are rounded for display, but the unrounded values continue into the difference formula. |
Identical colors are a useful starting point because they return zero. Also, with the same default conditions, comparing A with B produces the same result as comparing B with A, subject only to floating-point display precision. These properties make the calculator deterministic, but they do not remove the need to save the original inputs.
Run the comparison with valid inputs
Use the following procedure whenever you need to reproduce a result or check a previous calculation:
- Identify the two original colors. Take them from the same source used for the earlier comparison. Preserve the leading #, all six hexadecimal digits, and the intended order of the pair.
- Confirm that each value represents an opaque sRGB color. The accepted format is exactly #RRGGBB. If a source contains three-digit shorthand, alpha, a named color, Display P3 notation, or malformed text, resolve it into the intended opaque six-digit sRGB value before entering it. The calculator rejects unsupported formats instead of silently reinterpreting them.
- Enter one color in each field. Do not paste surrounding spaces, notes, or extra characters into a HEX field. An empty value, missing hash, shorthand, alpha value, non-hexadecimal character, or extra digit produces an explicit error and no calculated result.
- Select Compare colors. The calculator converts both sRGB inputs through D65 XYZ, Bradford-adapts them to D50, calculates their CIELAB coordinates, and then applies CIEDE2000 with kL = kC = kH = 1. Everything runs locally in the browser, and the entered color values are not uploaded.
- Read the numeric Delta E 2000 value and both D50 CIELAB coordinate sets. Treat the displayed Lab values as inspection figures because they are rounded. Use the reported Delta E value as calculated from the unrounded coordinates.
- Save the original pair and result. Record both entered HEX values, the Delta E 2000 number exactly as reported, and the tolerance and viewing or production context that give the comparison meaning.
If transparency matters, composite both colors over the same known background before comparing them. A translucent foreground is not one standalone opaque sRGB color, so changing the background or compositing method between runs can change the color ultimately submitted to the calculator.
| Input form | Example | Calculator behavior |
|---|---|---|
| Accepted opaque sRGB HEX | #A1B2C3 | Accepted for calculation. |
| Three-digit shorthand | #ABC | Rejected; write out the intended six-digit value before comparison. |
| Alpha-bearing HEX | #11223344 | Rejected because the tool compares opaque colors. |
| Named or other color-space notation | rebeccapurple or color(display-p3 ...) | Rejected instead of being interpreted under an assumed color space. |
| Malformed text | Missing #, empty field, invalid digits, or extra digits | Validation occurs before calculation, so an error appears without a result. |
Keep the conversion and formula unchanged
The calculator follows the W3C CSS Color 4 conversion and DeltaE2000 sample code path. Each sRGB channel is normalized to a value from zero through one, linearized with the sRGB transfer function, and converted into CIE XYZ relative to D65. A Bradford matrix then adapts the XYZ values from D65 to D50 before the calculator converts them to CIELAB.
CIEDE2000 then accounts for differences that a basic straight-line Lab distance would miss. It adjusts the Lab a axis near neutral colors, handles hue angles across the zero-degree boundary, weights the lightness, chroma, and hue contributions, and includes a rotation term in the blue region. Keeping the metric identified as CIEDE2000 and its default parameters unchanged prevents a later reader from applying a different interpretation to the same HEX pair.
The implementation is deterministic and is checked against eight sign-sensitive reference pairs from the Sharma, Wu, and Dalal supplemental CIEDE2000 implementation data. Those checks support correct handling of hue wrapping and weighting terms. For your own records, distinguish the rounded Lab values shown for inspection from the unrounded values used internally; manually recalculating from the visible coordinates may produce a slightly different result.
Save a result another person can reproduce
A useful repeatability record identifies not only the number but also everything needed to interpret it. Create a QA note, ticket, design-system annotation, or other durable entry containing these items:
- The two exact entered HEX strings, including each leading # and the order in which they were compared.
- The Delta E 2000 result exactly as displayed, without replacing it with a rounded or reformatted substitute.
- The D50 CIELAB coordinates shown for each color, labeled as rounded display values if they are included.
- The method: opaque sRGB, the documented D50 conversion path, CIEDE2000, and default kL = kC = kH = 1 weights.
- The workflow tolerance and where that tolerance came from, if the comparison supports a pass, fail, revision, or approval decision.
- The viewing or production context, including relevant display calibration, ambient lighting, surface, surrounding colors, process, or physical sample conditions.
- For composited colors, the background and compositing method used before the resulting opaque HEX values were entered.
Record the order even though the default calculation is symmetric. It removes ambiguity when several colors or revisions are being tracked. Also avoid inventing a universal acceptance limit: the calculator provides a formula-defined difference, while each production process must define its own tolerance and conditions.
| Use case | What to record | Important limitation |
|---|---|---|
| Web or design-token revision | Original HEX values, Delta E result, tolerance, and target rendering context | The calculation compares digital sRGB coordinates; it cannot prove how every display will render them. |
| Sampled image comparison | The sampled HEX values, sample location or method, Delta E result, and comparison context | The calculator does not pick colors from an image or measure an object. |
| Print, paint, or textile review | The digital comparison as a reference plus the physical process and acceptance conditions | Physical matching requires calibrated measurement equipment and process-specific illuminant, observer, profile, substrate, and tolerance conditions. |
Separate numerical repeatability from visual agreement
A larger Delta E 2000 number represents a greater formula-defined difference, but the calculator intentionally does not label a fixed value as pass, fail, identical to every observer, or acceptable for every application. Perceived agreement also depends on viewing conditions, display calibration, ambient light, surface texture, surrounding colors, and observer vision.
The two swatches provide a quick visual reference, yet browser rendering, monitor gamut, color profiles, brightness, and calibration can alter their appearance. The numerical calculation is based only on the entered sRGB coordinates and the documented conversion path. It cannot measure a real object, infer a printer profile, or establish that two physical samples match.
This distinction also separates CIEDE2000 from accessibility testing. Closeness between two colors does not establish readable foreground and background contrast. For text and interface readability, use the Color Contrast Checker and its WCAG-based contrast rules instead.
Check what changed when a result differs
If a later run does not match the saved result, work through these checks in order:
- Compare the exact input strings. A changed digit or a different color source changes the numerical input even if both swatches look nearly identical on the current display.
- Confirm both values remain valid opaque #RRGGBB inputs. Unsupported input should produce an error rather than a hidden conversion.
- Check any background compositing. For transparent colors, the same known background and compositing method must be used in both runs.
- Confirm the method. The expected result is CIEDE2000 with D50 Lab coordinates and default weights, not a different color-difference formula.
- Do not rebuild the result from the visible Lab coordinates. They are rounded for display, while the calculator uses unrounded values in the formula.
- Preserve the reported precision. Copy the result as shown instead of introducing an extra rounding step.
- Separate calculation changes from viewing changes. If the number is unchanged but the appearance differs, investigate display, lighting, profile, surface, or surrounding-color conditions rather than changing the saved HEX pair.
Two sanity checks can help isolate a transcription error: an identical pair should return zero, and reversing the order should preserve the same default-condition result, subject only to floating-point display precision. The complete repeatable unit is therefore the two exact HEX inputs, the fixed conversion and formula, the full reported result, and the context in which the number is used.
Related reading: Simplify Color Difference With a CIEDE2000 Calculator.
Related reading: How to Repeat Color Palette Generator Results.