Illumination side: sets the excitation
A filter on the illumination side determines how much UV reaches the sample. It changes the emission condition of the sample and therefore directly affects the result for fluorescent materials.
Color measurement · Spectrophotometric
Matched to the colour path: illumination spectrum and detection band
For portable, benchtop and in-line colour measurement systems. Working from the illumination source, the dispersion structure, the detector response and the sample behaviour, we design UV-cut, out-of-band rejection, channel selection and intensity-control filters for material colour measurement and optical module integration.

A colorimeter filter is a spectral-selection or intensity-control element in a colour measurement path. By changing transmission, reflection or absorption in specific bands it adjusts the spectrum reaching the sample, or limits the spectral range entering the detector.
A spectrophotometric colorimeter first acquires the spectral information of the sample as a function of wavelength, then computes tristimulus values and colour parameters under the selected standard illuminant and standard observer. It therefore provides not only a colour difference number but also the spectral data behind it.
In a typical grating colorimeter the grating and array detector carry the main dispersion and detection function. The filter supports UV control, non-target band suppression or intensity adjustment; such an instrument should not be described simply as "measuring colour with red, green and blue filters".
Two colour measurement approaches have to be distinguished:
| Approach | Basic characteristics | Typical filter function |
|---|---|---|
| Spectrophotometric colour | Measures spectral information at many wavelengths, then computes colour parameters | Illumination spectrum control, out-of-band rejection, order sorting, auxiliary selection |
| Tristimulus colour | Approximates tristimulus values through several weighted response channels | Forms the target channel response together with source and detector |
This page covers the spectrophotometric approach. Instruments that use filter-based tristimulus channels are described on the colour difference meter filter page; the two should not share one set of specifications.
A filter does not change colour itself; it changes the spectrum illuminating the sample and the spectrum reaching the detector.
The spectrum the sample receives is set by the source and the illumination path together. A filter on the illumination side changes the UV, visible or other content reaching the sample.
The instrument collects reflected light according to its illumination and viewing design. The measurement geometry determines how diffuse and specular components are handled and is essential for comparability.
The collected light enters the detector through the dispersion system. With reference calibration the system obtains the spectral measurement result.
For non-fluorescent samples, XYZ can be computed from the spectral reflectance, the specified illuminant spectrum and the standard observer colour matching functions, then converted to L*a*b* and similar parameters. Comparing sample and standard gives the colour difference using the agreed formula.
Filter elements on the illumination side and the detection side do different jobs: one sets the light the sample receives, the other sets the light the detector sees. Mixing their specifications is the most common source of error in colorimeter filter selection.
With fluorescent samples, colour measurement is no longer pure reflection — it becomes reflection plus emission.
Paper, textiles and plastics may contain fluorescent whitening agents or other fluorescent components. These materials absorb shorter wavelengths and can emit fluorescence at longer wavelengths, so the light collected by the instrument may contain both reflected light and fluorescence.
When the UV content of the illumination changes, the degree of fluorescent excitation can change as well, affecting whiteness, chromaticity and colour difference. For such samples, matching only the visible detection band is not enough to guarantee the same measurement conditions.
A UV-cut filter can reduce a specified UV range in the illumination; systems with UV adjustment also need source control, reference materials and calibration methods to define the actual condition. Instruments differ: some use a mechanical filter change, others use multiple sources or numerical processing.
A filter on the illumination side determines how much UV reaches the sample. It changes the emission condition of the sample and therefore directly affects the result for fluorescent materials.
A filter on the detection side changes which wavelengths reach the detector. It does not change how much the sample was already excited — only how much light gets through.
Visible fluorescence already generated in the sample cannot be removed by a UV-cut filter on the detection side. This is where the two positions are most often confused.
The structures below are grouped by where they sit and what they solve; the combination depends on the instrument architecture and the sample behaviour.
Reduces transmission over a specified UV range while keeping the visible light required, for UV exclusion or UV condition adjustment. The design has to consider UV suppression, the cut-off edge and short-wave visible transmission; an edge that reaches too far into the visible region can also change blue-violet illumination conditions.
Keeps the visible measurement range required and reduces the share of non-target NIR entering the detection channel. Whether it is needed depends on source output, detector response and the existing dispersion system; an instrument that "measures visible colour" cannot be assumed to be unaffected by non-visible bands.
For dispersion structures affected by higher-order overlap, suppressing non-target light that may overlap the target band. Its form and position depend on whether the architecture is scanning or array based; in an array system that detects a wide band at once, a single long-pass element cannot simply block short wavelengths when those wavelengths are also measurement targets.
For colour modules that use a filter array, a filter wheel or another multi-channel selection method, sampling the spectrum with several passbands. Channel count, bandwidth, overlap and the following algorithm all affect spectral reconstruction and the colour result; a few broad channels do not equal full spectrophotometric capability.
Attenuates intensity over a specified range for detector dynamic-range matching, reference channel adjustment or a test path. Colour work also needs the spectral flatness of the attenuation: a large wavelength-dependent difference can change the received spectral balance and has to be handled in design and calibration.
The five options are not mutually exclusive: one instrument may use illumination-side UV cutting and detection-side rejection at the same time, so each position should be confirmed separately.
The table lists the parameters to confirm item by item; fabrication ranges and performance figures are confirmed against the project requirement and validated fabrication and inspection results.
| Parameter | Customisation | Explanation |
|---|---|---|
| Position | Illumination side, detection side, reference channel or inside the dispersion assembly | The function differs with the position |
| Working band | Set by the actual spectral coverage of the instrument | A single "visible range" is not applied universally |
| Spectral structure | UV cut-off, NIR cut-off, bandpass or neutral density | Selected according to the system goal |
| Cut-off edge | Position, criterion and permitted deviation | State the transmission level used to define the edge |
| Passband transmission | Minimum, average or peak | Defined separately; the three cannot be interchanged |
| Passband flatness | Transmission variation within a specified interval | Affects spectral shape and channel response |
| Out-of-band rejection | Transmission limit or OD over a specified range | Not equivalent to the colour difference accuracy of the instrument |
| Channel bandwidth | FWHM and guaranteed passband | Applies to bandpass parts and multi-channel sets |
| Incidence conditions | Chief ray angle, angular spread and polarisation | Spectral shift assessed against the actual path |
| Substrate and thickness | Spectral fit, mechanical strength, process compatibility | Substrate absorption and possible self-fluorescence also matter |
| Outline and clear aperture | Per holder, filter wheel or array drawing | Distinguish the outline from the guaranteed optical area |
| Surface and edge | Surface defects, chamfer and edge requirements | Agreed by scatter sensitivity and assembly need |
| Environment and lifetime | Temperature, humidity, light exposure and cleaning | Assesses long-term spectral change |
| Inspection and delivery | Spectral curves, dimensions and batch documents | State test conditions and acceptance method |
Colorimeter filters require spectral performance and batch consistency together. A high transmission on one sample does not mean that different batches will give the same colour result once installed. The released specification should define the critical spectral intervals, tolerances and inspection conditions, with an acceptable range confirmed by instrument validation.
These four come up most often in colorimeter filter discussions, and are most often treated as the same thing.
The sampling interval is how often the instrument reports a data point along wavelength; the spectral bandwidth is the wavelength range a single measurement channel actually receives. They are different figures, and a smaller interval does not automatically mean higher real spectral resolution.
SCI includes the specular component and SCE excludes it. They concern the measurement geometry and how specular reflection is handled; they are not two filter colours and cannot be achieved simply by changing a bandpass element.
Selecting D65 for the calculation does not mean the instrument contains a physical source with exactly that spectrum. For non-fluorescent samples the colour can be computed under the specified illuminant from the measured reflectance; for fluorescent samples the actual excitation conditions also have to be considered.
Filter transmission, OD or centre wavelength tolerance cannot be converted directly into an instrument dE. The result also depends on the source, measurement geometry, wavelength accuracy, detector, calibration and sample condition.
Requirements differ widely with the material measured; fluorescence is the first dividing line.
| Application | Typical samples | Filter design focus |
|---|---|---|
| Textile colour measurement | Dyed and optically brightened fabrics | UV conditions, fluorescence effects, batch consistency |
| Paper and print colour measurement | Paper, board, printed patches | Fluorescent whitening agents, illumination conditions, spectral response |
| Plastics and coatings | Plastic parts, masterbatch, coatings | Visible spectral matching, geometry, stability |
| Benchtop colorimeters | Laboratory material samples | Spectral purity, reference channel, calibration fit |
| Portable colorimeters | On-site colour difference checks | Compact size, assembly consistency, environmental robustness |
| In-line colour inspection | Continuously produced material | Long-term light stability, thermal drift, maintenance conditions |
For metallic, pearlescent and other effect materials, the filter has to be assessed together with a multi-angle measurement structure; a single filter arrangement rarely covers colour that changes with viewing angle.
Requirement review → spectral design → mechanical adaptation → sample inspection → instrument validation → specification release
Reflection or transmission, sample type, and whether fluorescent components are present.
Source, dispersion method, detector and filter mounting position.
Passband, blocking regions, edges, tolerances and working angle.
Thickness, clear aperture, fixing method and switching mechanism.
Check spectra and test typical, dark and applicable fluorescent reference samples.
Agree batch consistency, inspection conditions and delivery documents.