Narrow Bandpass, Dark Image: Why 5 nm Can Cost You 70% Transmission
Narrow Bandpass, Dark Image: Why 5 nm Can Cost You 70% Transmission
Interference bandpass filters are popular in machine vision for a simple reason: they're exceptionally good at passing only a narrow slice of wavelengths while rejecting everything else. Built from thin-film stacks—metal or dielectric films with roughly λ/4 thickness—they extinguish unwanted wavelengths through interference, producing precisely defined transmission bands.
That's exactly what you want when you're isolating your illuminator wavelength or fighting uncontrolled ambient light. But there's a critical trade-off hiding in the physics.
The Narrower You Go, The More Light You Lose
Here's what catches many engineers off guard: the narrower the bandwidth, the larger the transmission loss—even within the pass range itself.
A 5 nm bandwidth filter can mean roughly 70% loss of transmission. Narrow down to 10 nm and you're still looking at about 50% loss. Even a relatively wide 20 nm bandpass typically costs you around 10% of your in-band light.
In other words, "narrow" doesn't just mean "selective." It often means "hungry for photons."
When Photon Hunger Meets Exposure Constraints
This photon hunger collides directly with the reality of machine vision constraints. Your exposure time may be bounded by motion blur. Your f-stop may be limited by depth-of-focus requirements. You can't always adjust freely.
Consider what happens when you push toward short shutter times and small apertures. Going from a relaxed f/4 at 1/200 second to a demanding f/16 at 1/3200 second requires roughly 64 times brighter lighting just to maintain the same image brightness.
Now add a narrow bandpass filter on top of that. The system's only escape valves are the same three levers: increase lighting power (if thermally and spatially possible), relax exposure time (if motion blur permits), or open the lens wider (if depth of focus allows).
The Real Nature of the Decision
Choosing a filter bandwidth isn't purely a spectral decision. It's fundamentally a light budget decision that ripples through your entire system design.
Before committing to a narrow bandpass, ask yourself: can my lighting, exposure, and optics absorb the transmission hit? If the answer is uncertain, work the numbers before ordering the filter.
This is part of KUPO's technical resource series on optical filters for machine vision. For custom filter solutions, contact KUPO CO. LTD.
Frequently Asked Questions
https://www.kupooptics.com/en/blogs/application-notes/mv_narrow_bandpass
Why do narrow bandpass filters produce darker images?
Narrow bandpass filters offer precise wavelength selection but cost significant transmission. A 5 nm FWHM filter can reduce transmission by 70% or more compared to a wider bandpass.
How does bandwidth relate to light loss in interference filters?
The narrower the bandwidth, the larger the transmission loss—even within the pass range itself. This is a fundamental trade-off in thin-film interference filter design.
What constraints does narrow bandwidth create for machine vision systems?
Your exposure time may be bounded by motion blur, and your f-stop by depth-of-focus requirements. You can't always compensate for the light loss with camera settings alone.
What is the fundamental nature of the bandwidth decision?
Choosing a filter bandwidth isn't purely a spectral decision—it's fundamentally a light budget decision that ripples through your entire system design.
When is a narrow bandpass filter worth the transmission cost?
When spectral selectivity is critical—laser line isolation, fluorescence imaging, or suppressing a specific interference source. The precision is worth the photon penalty when your application demands it.
What should you verify before choosing a narrow bandpass filter?
Confirm your light source has enough power, your camera has sufficient sensitivity, and your exposure time is not already at its maximum. If any of those constraints are tight, a wider bandpass may be the better engineering choice.