Is PSF (Point Spread Function) Tied to a Camera and Lens Pair for All Exposures?
If you are attempting image deconvolution or forensic sharpening, you might hope that once you measure the PSF of your Canon R5 and 85mm f/1.2, you have a universal key to "unlock" detail in all future photos. Unfortunately, the PSF is not a static constant. It is a state-dependent function that fluctuates based on the specific physics of each individual exposure.
1. The Myth of the "Fixed" PSF
In a perfect vacuum, a PSF would only represent the diffraction limit of the lens. But in real-world photography, the PSF is the sum of several variables. To say a PSF is tied to a camera/lens pair is like saying the "speed of a car" is tied to its make and model—it ignores the gear, the RPM, and the road conditions.
2. Why PSF Changes Between Exposures
There are four primary "points of failure" where a pre-measured PSF will fail to match a new exposure:
A. Aperture (The Diffraction Factor)
The size of the hole light passes through radically alters the PSF. At wide apertures (e.g., f/1.4), the PSF is dominated by lens aberrations (spherical, coma, astigmatism). As you stop down to f/11 or f/16, the PSF is reshaped by diffraction, growing larger and more circular (the Airy Disc). A PSF measured at f/2.8 is useless for a photo taken at f/8.
B. Focus Distance (The Geometric Factor)
Most lenses are "optimized" for a specific distance (usually infinity). As you move to the Minimum Focus Distance, the internal lens elements shift, often introducing "field curvature" or changing the shape of the blur. This means the PSF for a landscape shot is physically different from the PSF for a macro shot using the exact same lens.
C. Position in the Frame (Spatial Variance)
A lens is rarely "isoplanatic" (meaning the PSF is the same everywhere). The PSF at the center of the sensor is typically sharp and symmetrical, while the PSF in the corners may be stretched into "cat-eye" shapes or smeared due to lateral chromatic aberration. If your subject moves from the center to the edge, the PSF effectively changes.
D. Wavelength (Chromatic Factor)
Light of different colors (wavelengths) refracts at different angles. This results in Chromatic Aberration, where the PSF for the Red channel might be slightly larger or shifted compared to the Blue channel. If your lighting temperature changes from daylight to tungsten, your effective PSF shifts as well.
3. The Exposure Triangle and PSF
| Setting | Does it affect PSF? | Why? |
|---|---|---|
| Aperture | Yes (Primary) | Changes diffraction and aberration balance. |
| Shutter Speed | Technically No | Does not change the optical PSF, but adds Motion Blur (a different kernel). |
| ISO | No | Affects noise floor, but not the optical spread of light. |
Note: While shutter speed doesn't change the optical PSF, camera shake during a long exposure creates a "composite PSF" that is a convolution of the optical PSF and the motion path.
4. Practical Implications for Deconvolution
If you are using software like FocusMagic or AstroArt for deconvolution, you cannot use a single "calibration" file for everything. For high-precision work, you must:
- Measure a "PSF Library" for various f-stops.
- Use Blind Deconvolution, which estimates the PSF from the image data itself rather than relying on a pre-set pair.
- Account for the Anti-Aliasing (AA) Filter on the sensor, which adds a permanent, fixed layer of blur regardless of the lens.
Conclusion
The Point Spread Function is tied to the optical state of the system at the moment of capture, not just the hardware names on the box. While the camera and lens provide the "base" for the PSF, the aperture, focus distance, and even the color of the light are what ultimately define the blur. To get the best results, treat your PSF as a unique attribute of every single frame.
Keywords: Point Spread Function photography, PSF camera lens pair, image deconvolution PSF, lens aberrations PSF, diffraction limit Airy Disc, spatial variance PSF, optical transfer function, PSF vs focal distance.
