Optical filters are defined by the way they control light across a specified wavelength range. Some transmit a narrow region of the spectrum and reject wavelengths on either side. Others separate shorter wavelengths from longer ones, reduce optical intensity or divide different spectral regions between separate optical paths.
These differences are not simply variations in construction. They reflect the different tasks optical systems need to perform. A fluorescence instrument may need to isolate a narrow emission band from a much stronger excitation source, while an imaging sensor may require broad visible transmission combined with strong infrared rejection. In another system, the objective may be to reduce optical power without substantially changing the spectral distribution of the light.
For this reason, optical filters are usually classified according to their spectral response: the relationship between wavelength and the amount of light transmitted, reflected or rejected by the filter.
Understanding these response characteristics provides a more useful way of comparing filter types than considering their physical appearance alone.
Understanding Spectral Transmission and Blocking
The behaviour of an optical filter is commonly represented by a spectral transmission curve. Wavelength is plotted along the horizontal axis and transmission along the vertical axis, allowing the regions transmitted and rejected by the filter to be seen directly.
A filter designed to transmit a specific wavelength region will show a passband within that part of the spectrum. Outside the passband, transmission falls as the filter enters its blocking regions.
The transition between high transmission and blocking is also important. Some applications tolerate a gradual spectral transition, while others require a steep edge so that closely spaced wavelengths can be separated effectively.
Filter classifications such as bandpass, longpass and shortpass describe the overall shape of this spectral response. They do not, however, define the complete performance of the component. Two filters belonging to the same category may have substantially different bandwidths, transmission levels, blocking performance and transition characteristics.
The filter type therefore establishes the required spectral function; the detailed optical specification determines how precisely that function is performed.
Bandpass Filters
A bandpass filter transmits a defined region of the spectrum while rejecting wavelengths both above and below that region.
The transmitted region may be relatively broad or extremely narrow depending on the application. Narrow bandpass filters are particularly useful where the optical system must distinguish a specific spectral feature from neighbouring wavelengths or background radiation.
In spectroscopy, for example, the passband can be positioned around a wavelength associated with a particular measurement. Fluorescence systems use carefully selected excitation and emission filters to separate relatively weak fluorescent signals from the illumination used to generate them. Imaging and sensing systems may use bandpass filters to isolate spectral information that improves contrast or enables a particular material or feature to be detected.
The position of the transmission band is generally described using a centre wavelength, while its width is commonly expressed as full width at half maximum (FWHM). The amount of unwanted light transmitted outside the passband is determined by the filter’s blocking performance.
These parameters need to be considered together. A narrow passband is of limited value if unwanted wavelengths outside it are insufficiently suppressed, just as deep blocking provides little benefit if transmission within the required band is too low.
Bandpass filters therefore represent a balance between spectral selectivity, transmission and rejection appropriate to the optical system.
Longpass Filters
A longpass filter separates the spectrum at a transition region, transmitting wavelengths longer than the specified cut-on while rejecting shorter wavelengths.
Unlike a bandpass filter, there is no second transition immediately beyond the transmitting region. Once the filter enters its passband, transmission continues towards longer wavelengths within the operating range for which the filter has been designed.
This behaviour is useful where an optical system needs to remove shorter-wavelength radiation while retaining a broad range of longer wavelengths. Longpass filters are used in fluorescence, spectroscopy, imaging and sensing systems, among other applications.
The transition between blocking and transmission is particularly important. A steep spectral edge allows wavelengths relatively close to the cut-on region to be separated more effectively, while a more gradual transition may be entirely adequate where the unwanted and required spectral regions are widely separated.
As with other interference filters, the stated cut-on wavelength is meaningful only in relation to the conditions under which the filter is specified, particularly the angle of incidence.
Shortpass Filters
A shortpass filter performs the complementary function to a longpass filter. Shorter wavelengths are transmitted while wavelengths beyond the transition region are rejected.
This makes shortpass filters useful where longer-wavelength radiation needs to be removed without sacrificing transmission at shorter wavelengths.
One familiar example is the control of unwanted infrared radiation in visible imaging systems. Silicon-based imaging sensors can remain sensitive beyond the visible spectrum, allowing near-infrared radiation to influence the recorded image if it is not adequately controlled. An appropriately designed shortpass or IR-cut filter can suppress this response while maintaining transmission across the required visible region.
Shortpass filters are also used within scientific instruments, detector systems and optical assemblies where longer wavelengths would otherwise interfere with the measurement being made.
The distinction between longpass and shortpass is therefore straightforward, but their performance depends heavily on the position and shape of the spectral transition rather than the category name alone.
Notch and Band-Rejection Filters
A notch filter reverses the basic behaviour of a bandpass filter. Instead of transmitting a defined spectral region and blocking wavelengths around it, the filter rejects a selected wavelength band while allowing wavelengths on either side to pass.
This is useful when a strong or unwanted spectral feature must be removed without discarding the surrounding optical information.
Laser-line rejection provides a typical example. An optical system may need to suppress radiation from a particular laser wavelength while continuing to observe wavelengths immediately above and below it. A suitably designed notch filter creates a narrow region of high attenuation centred on the unwanted wavelength.
The width and depth of the rejection region depend on the application. Removing a single narrow spectral feature requires different performance from rejecting a broad section of the spectrum, and the filter must be specified accordingly.
Notch filters demonstrate why optical filter selection is better understood in terms of spectral function than simple categories. Their purpose is not to make the entire system darker, but to remove a particular part of the optical signal while preserving information elsewhere.
Neutral Density Filters
Neutral density filters differ from wavelength-selective filters because their principal purpose is to control optical intensity rather than isolate a particular spectral region.
An ideal neutral density filter attenuates light by a defined amount while maintaining a broadly consistent spectral response across its specified wavelength range. The relative distribution of wavelengths is therefore preserved while the overall optical power is reduced.
This is useful where a detector would otherwise saturate, where illumination needs to be controlled or where measurements must be made at several known attenuation levels.
Neutral density filters can achieve attenuation through absorption, reflection or a combination of optical mechanisms depending on their construction. The choice becomes particularly important in systems handling substantial optical power, as an absorptive filter retains rejected energy as heat while a reflective design redirects a greater proportion of it away from the optical path.
Attenuation is often expressed in terms of optical density, providing a logarithmic description of the relationship between incident and transmitted light.
Unlike the blocking specification of a wavelength-selective filter, however, the optical density of a neutral density filter describes the intentional attenuation of the working beam itself.
Dichroic Filters
Dichroic filters use wavelength-selective thin-film coatings to transmit one spectral region while reflecting another.
Rather than simply removing unwanted light from the system, they can therefore separate an incoming beam into two useful spectral paths. This makes them particularly valuable in optical systems where different wavelength regions need to be directed towards separate detectors or processing channels.
Fluorescence microscopy provides a well-known example. A dichroic element can reflect excitation light towards a specimen while transmitting the longer-wavelength fluorescence emission towards the detector. Similar principles are used in imaging, projection, analytical instrumentation and multispectral systems.
The transition between reflected and transmitted wavelengths is determined by the coating design and must be considered alongside angle of incidence and polarisation. Dichroic filters are frequently used at non-normal incidence specifically so that the reflected and transmitted beams follow different physical paths.
Their function is therefore both spectral and geometric: wavelength determines which light is transmitted or reflected, while the orientation of the filter determines where those beams travel within the system.
Infrared Optical Filters
Infrared optical filters operate beyond the visible spectrum and are used to control radiation across regions including the near infrared and mid infrared.
The underlying filtering principles remain the same, but moving into the infrared changes many of the practical considerations involved in filter design. Substrate materials that perform well in visible systems may no longer provide the required transmission, while coating materials and deposition processes must be appropriate for the intended wavelength range.
Infrared filters are used in imaging, sensing, spectroscopy, environmental monitoring, aerospace instrumentation and other systems in which spectral information beyond the visible range needs to be isolated or controlled.
Some transmit a defined infrared band while rejecting radiation outside it. Others provide broader spectral separation or form part of multispectral optical assemblies.
The combination of wavelength range, detector response and operating environment means infrared filter design needs to be considered as part of the complete optical system rather than simply as a visible filter shifted to a longer wavelength.
IR-Cut Filters
IR-cut filters are designed to prevent unwanted infrared radiation from reaching an imaging detector while maintaining the required transmission through the visible spectrum.
This becomes particularly important with silicon-based CMOS and CCD sensors. Their spectral sensitivity extends beyond the wavelengths perceived by the human eye, so near-infrared radiation can contribute to the recorded signal unless it is suppressed.
Without appropriate infrared control, colour reproduction and image response can be affected by radiation that is invisible to the observer but detectable by the sensor.
An IR-cut filter introduces the required spectral boundary between visible transmission and infrared rejection. Its performance needs to complement the spectral response of the detector and the other components within the imaging system.
The design is therefore more sophisticated than simply “blocking infrared”. Visible transmission, transition position and out-of-band rejection all contribute to the final imaging performance.
Multiband and Multispectral Filters
Not every optical system can be divided into a single transmitting region and a single blocking region.
Multiband filters are designed with two or more distinct transmission bands separated by controlled blocking regions. This allows several selected wavelength ranges to pass through a single optical component.
The approach is valuable in instruments that collect information from multiple spectral channels. Rather than changing filters mechanically or using several independent components, a carefully designed multiband filter can provide the required spectral structure within a single coating system.
Multispectral optical systems take this principle further by using information from separate wavelength regions to distinguish features that may appear similar when observed within only one part of the spectrum. Applications range from scientific imaging and remote sensing to specialist aerospace and surveillance systems.
Producing several accurately positioned spectral regions within one filter places significant demands on thin-film design and deposition control. Each transmission and blocking region forms part of the overall coating response and must operate correctly alongside the others.
Filter Type and Optical System Requirements
Terms such as bandpass, longpass, shortpass and neutral density are useful because they describe the basic function of an optical filter. They should not, however, be mistaken for complete specifications.
A requirement for a “bandpass filter”, for example, leaves many important questions unanswered. The centre wavelength and bandwidth still need to be defined, together with transmission through the passband, blocking outside it, angle of incidence and the spectral range over which that blocking must be maintained.
The detector, illumination source and optical geometry also influence the required response. In infrared systems, substrate transmission and environmental performance may become equally important. Where a filter operates at an angle, spectral shift and polarisation effects may need to be included from the beginning of the design process.
For this reason, engineers select filter types according to the spectral function required by the complete optical system. The category provides a starting point; the performance specification defines the component that must actually be designed and manufactured.
Understanding the different types of optical filter is therefore less about memorising a list of components and more about recognising the different ways in which an optical spectrum can be controlled. Once the required spectral behaviour is established, parameters such as wavelength, bandwidth, transmission and blocking can be defined around the needs of the system.
Related Guides
- What Is an Optical Filter?
- How Do Optical Filters Work?
- Types of Optical Filters
- Optical Filter Performance Characteristics
Types of Optical Filters – FAQs
What is the difference between a bandpass and a notch filter?
A bandpass filter transmits a selected wavelength region while rejecting wavelengths outside it. A notch filter does the opposite, rejecting a defined spectral region while allowing wavelengths on either side to pass.
What is the difference between a longpass and a shortpass filter?
A longpass filter transmits wavelengths above its spectral transition while rejecting shorter wavelengths. A shortpass filter transmits the shorter-wavelength region and rejects wavelengths beyond its transition towards the longer end of the spectrum.
Is a neutral density filter an optical filter?
Yes. A neutral density filter is an optical filter designed primarily to attenuate optical intensity rather than select a narrow wavelength region. Its attenuation is intended to remain relatively uniform across its specified operating spectrum.
Are IR-cut and infrared filters the same thing?
No. The term infrared filter covers a broad range of filters designed to control infrared radiation. An IR-cut filter has a more specific function: suppressing infrared wavelengths while transmitting the required shorter wavelengths, commonly the visible spectrum in imaging applications.
Can one optical filter have several transmission bands?
Yes. Multiband filters can be designed with two or more separate transmission regions and controlled blocking between them. They are used where an optical system needs to work with several defined wavelength bands.