If you point a perfect lens at a tiny light source, it still does not draw a perfect dot. Instead, you get a small bright spot with faint rings around it. That pattern is the Airy disk, and it happens even when nothing is misaligned or out of focus.

A lens does not move one point of light onto one infinitely small point on the sensor. It spreads that point into a tiny pattern. That spread is what puts a floor under sharpness: two details can be close enough that their patterns blend together before the image gives you two clean peaks.

One useful way to picture it is as a pond-like wave pattern: imagine throwing pebbles into a pond so that each splash sends ripples across a circular opening, all overlapping and interfering with one another. In that metaphor, the center stays brightest because the contributions line up there most strongly, while the outer ripples cancel more and more as you move away from the middle. The water-ripple picture is not a literal lens model, but it gives the right first intuition: the size of the opening changes how the overlapping waves line up.

As the Aperture size slider moves, the bright center and the surrounding rings change size. A larger aperture pulls the central spot inward and concentrates more of the light near the middle. A smaller aperture lets the pattern spread outward. The important point is not simply that a bigger opening admits more light. The opening also changes the shape of the smallest spot the lens can make.

The rest of the article keeps that visual idea and makes it more precise. The real lens case is not water, pebbles, or surface ripples. It is light passing through a finite circular opening, then combining at the image plane. Because light is a wave, that combination can reinforce in some places and cancel in others.

A Point Becomes a Small Pattern

Start with the simplest possible subject: one distant point of light. In a ray diagram, every ray from that point might appear to meet at one perfect image point. That drawing is useful, but it leaves out diffraction. Diffraction is what happens when a wave is limited by an opening or edge. The opening cuts down the wavefront, and the remaining parts of the wave combine again after the aperture.

For a circular lens aperture, that recombination has a special shape. Most of the light lands in a bright central spot. A smaller amount lands in faint circular rings around it. The full spot-and-ring pattern is the Airy pattern, and the bright central region is commonly called the Airy disk. The pattern takes its name from George Biddell Airy’s 1835 circular-aperture analysis, On the Diffraction of an Object-glass with Circular Aperture.

This is the first important correction to the usual “better lens means smaller dot” story. Even a perfect circular lens has a smallest possible spot. Better manufacturing can remove defocus, astigmatism, coma, and other defects, but it cannot remove the basic diffraction pattern created by the aperture. That is why optical engineers call this the diffraction-limited spot: it is the limit that remains when the lens itself is otherwise ideal.

The Airy disk is also not an arbitrary blur shape. It has a predictable size. The radius from the center to the first dark ring is often written as:

$$r_1 = 1.22 \frac{\lambda f}{D}$$

This formula is easier to read if you connect each symbol to something physical:

• $\lambda$ is the wavelength of the light.
• $f$ is the focal length of the lens.
• $D$ is the diameter of the aperture.
• $r_1$ is the radius of the central bright spot out to the first dark ring.

The formula says three practical things. Longer wavelength makes the spot larger. Longer focal length makes the spot larger. A wider aperture makes the spot smaller. Those relationships are the reason blue light can resolve slightly finer detail than red light in the same optical system, and why a wider telescope aperture can separate closer stars.

Because photographers often talk about f-number instead of aperture diameter directly, the same relationship is also written as:

$$r_1 = 1.22\lambda N$$

Here $N$ is the f-number, which is $f/D$. At a higher f-number, such as f/16 or f/22, the aperture is small compared with the focal length, so the diffraction spot is larger. At a lower f-number, such as f/2.8, the diffraction spot is smaller. This is why stopping down a camera lens can improve some kinds of sharpness by reducing lens aberrations, then eventually reduce fine-detail sharpness once diffraction becomes dominant.

ZEISS gives a microscope-focused version of the same scaling in its discussion of fundamental aspects of Airy disk patterns: shorter wavelengths and larger numerical apertures produce smaller Airy disks, while longer wavelengths and smaller apertures make them wider. The camera, telescope, and microscope versions use different units, but the underlying tradeoff is the same.

The Size Controls Are Wavelength, Aperture, and Focal Length

The Airy disk explorer makes the formula visible. The left side shows the spot pattern on the image plane. The right side shows a radial profile: a graph of brightness as you move outward from the center of the spot. The orange marker shows the first dark ring, which is the $r_1$ distance in the formula above.

Increasing Aperture diameter D moves that orange marker inward. That means the central spot is getting smaller, so two nearby points have a better chance of staying separate. Increasing Wavelength moves the marker outward. Redder light produces a wider diffraction spot than bluer light under the same lens settings. Increasing Focal length while holding aperture diameter fixed also moves the marker outward because the f-number rises.

This is the core scaling law of the article. Diffraction blur is not random softness. It is a structured pattern whose size follows wavelength and aperture. Once that is clear, optical resolution becomes less mysterious: the question is how much these small patterns overlap.

Two Points Blend When Their Patterns Overlap

One point source makes one Airy pattern. Two point sources make two Airy patterns. If the points are far apart, the image has two separate bright spots. If the points are close together, the two patterns add together and can look like one wider spot.

That is what resolution means in this setting. It is not about magnifying the image until it looks larger. It is about whether the optical system preserves enough separation and contrast for two nearby sources to be distinguished. In astronomy the two sources might be stars. In microscopy they might be fluorescent emitters. In machine vision they might be tiny highlights or edges on a manufactured part.

A standard criterion is Rayleigh’s criterion:

$$\theta_R = 1.22\frac{\lambda}{D}$$

This version describes angular separation. It asks: how far apart do two distant points need to be, as seen from the lens, before the aperture can separate them? The answer gets smaller when wavelength is shorter or aperture diameter is larger. That is why a large telescope can resolve closer double stars than a small telescope at the same wavelength.

Rayleigh’s rule has a simple visual meaning. Two equally bright points are considered just resolved when the bright center of one Airy pattern lands on the first dark ring of the other. At that spacing, the combined profile has two peaks with a shallow dip between them. The dip is not huge, but it is enough to make the pair look like two sources rather than one.

In the Rayleigh explorer, watch the value sep/thetaR. When it is below 1, the separation is smaller than the Rayleigh limit and the two profiles tend to merge into one broad bump. Near 1, two peaks are visible but the valley between them is still shallow. Above 1, the peaks pull apart and the valley deepens.

The aperture and wavelength sliders change the rule itself. Making the aperture larger lowers $\theta_R$, so the same separation becomes easier to resolve. Making the wavelength longer raises $\theta_R$, so the same pair becomes harder to separate. The Separation slider then changes the scene: it moves the two point sources closer together or farther apart while the optical system stays fixed.

Rayleigh is useful because it gives a stable reference point, not because nature has a hard switch at exactly that value. Detection also depends on contrast, brightness, noise, pixel sampling, and processing. A high-contrast pair may be detected below the Rayleigh criterion; a low-contrast pair may be difficult even above it. The criterion is best treated as a physical baseline for comparing optical systems.

The Airy Disk Is a Point Spread Function

A point spread function, or PSF, is the answer to a simple question: if the real scene contains one perfect point of light, what shape does the image system record? For an ideal circular aperture, the answer is the Airy pattern. That makes the Airy disk more than a named spot. It is the basic blur fingerprint of a diffraction-limited lens.

Once you know the PSF, you can describe how the lens treats a more complicated scene. Imagine building the scene from many tiny points. Each point gets replaced by a copy of the PSF. Those copies overlap and add together. The recorded image is the result.

In image-processing language, this operation is convolution:

$$I_{image}(x, y) = I_{object}(x, y) * PSF(x, y)$$

You do not need the equation to understand the idea. It says the image is the original object after every point has been spread according to the PSF. In rendering and image processing, people often call the same kind of rule a blur kernel. The difference here is that the blur kernel is not chosen for artistic softness. For a diffraction-limited circular aperture, it comes from wave optics.

Real cameras and microscopes add more effects. Aberrations can make the spot asymmetric or larger. Sensor pixels sample the image in discrete steps. Demosaicing, sharpening, motion, and focus errors can all change the final result. Diffraction is still the clean baseline: the blur that remains even before those practical complications are added. The article on computational photography with HDR, burst denoising, and super-resolution picks up that software side: it shows how a camera can fuse measurements while still respecting optical and sampling limits.

The PSF convolution explorer applies that point-spread idea to repeated black-and-white detail. The upper signal is the ideal object. The lower signal is what remains after the diffraction-limited PSF has spread each point into its neighbors. As f-number or wavelength increases, the PSF gets wider and the lower signal loses contrast.

The Detail frequency slider changes how fine the repeated pattern is. Low-frequency detail has wide bright and dark bands, so it survives the blur more easily. High-frequency detail has narrow bands, so neighboring bright and dark regions smear into one another. This is why an image can look soft before a feature has completely disappeared. The detail may still be present, but with much weaker contrast.

For the image-processing side of the same idea, convolution and filtering in images and signals shows how a blur kernel changes repeated detail in a more general setting.

Fine Detail Fades Before It Vanishes

Optical resolution is often described with point pairs because the Airy disk makes that picture clear. But real images are not only pairs of points. They contain textures, edges, line patterns, and small repeated structures. For those scenes, contrast is often more useful than spot size alone.

This is where modulation transfer function, usually shortened to MTF, becomes useful. MTF describes how much contrast survives at each spatial frequency. A low spatial frequency is broad detail, like a large stripe. A high spatial frequency is fine detail, like tightly packed lines. Diffraction-limited blur behaves like a low-pass filter: broad detail passes through well, while very fine detail loses contrast.

For incoherent imaging with a diffraction-limited circular aperture, a common estimate for the cutoff frequency on the sensor plane is:

$$f_c \approx \frac{1}{\lambda N}$$

Here $f_c$ is the cutoff frequency and $N$ is the f-number. Above this frequency, the ideal diffraction-limited system transfers no contrast. Below it, contrast does not stay perfect until the cutoff. It falls gradually. That gradual fall matters because an image can lose useful detail well before a strict cutoff formula says “nothing gets through.”

This explains a common camera tradeoff. Opening a lens wide can reduce diffraction blur, but it may expose more aberrations. Stopping down can reduce aberrations and increase depth of field, but it also increases the diffraction spot. The sharpest aperture for a real lens is often a middle setting where aberration blur and diffraction blur are both reasonably controlled.

It also explains why megapixels are not the whole story. Tiny pixels can sample very fine detail, but the lens still has to deliver contrast at those fine scales. If the Airy pattern is wide compared with the pixel pitch, adding more pixels may record the blur more finely without recovering the missing optical contrast. That is why lens tests often use MTF: they are measuring contrast transfer, not just whether a line pair barely exists.

Evident Scientific’s discussion of cutoff frequency and Airy disk size frames the same relationship from the microscope side: the Airy disk in real space and the cutoff in frequency space are two ways of describing the same diffraction limit.

The Same Idea Across Cameras, Telescopes, and Microscopes

The units change from one field to another, but the idea stays stable. A telescope often talks about angular separation, because stars are effectively distant point sources. A microscope often talks about numerical aperture and nanometer-scale feature spacing. A camera lens often talks about focal length, f-number, pixel size, and sensor-plane blur.

All three are asking the same chain of questions. How large is the diffraction pattern? How much do neighboring point patterns overlap? How much contrast survives at the detail scale that matters for the task? NASA’s explanation of the resolving power of the Hubble Space Telescope is one familiar astronomy example: the practical question is whether nearby sources can be separated, not whether the telescope “magnifies” enough.

This chain is more useful than memorizing one formula:

• Diffraction sets the smallest ideal point pattern.
• The Airy disk describes the central part of that pattern for a circular aperture.
• The PSF describes how every scene point spreads into the image.
• Overlapping PSFs explain two-point resolution.
• MTF describes how much contrast survives for repeated fine detail.

Once those pieces are connected, the diffraction limit stops being a detached fact. It becomes a working model for choosing apertures, comparing objectives, interpreting lens charts, and building physically plausible camera blur in rendering. For a ray-optics contrast to this wave-optics limit, caustics from reflection and refraction shows how curved surfaces concentrate ray families into bright structures.

A Practical Way to Reason About It

When you need to reason about real optical resolution, move in this order. First, estimate the diffraction scale from wavelength and aperture. Use $r_1$ when you care about spot size on the image plane, and use $\theta_R$ when you care about angular separation. Second, compare that scale with the detector or sampling system. A sensor cannot recover detail that the optics never delivered, but poor sampling can also waste detail that the optics did deliver.

Third, decide whether diffraction is actually the main limit. At very small apertures, diffraction often dominates. At very large apertures, aberrations, focus error, or motion may dominate instead. Fourth, look at contrast for the detail you care about. Barely separating two ideal points is not the same task as reading a low-contrast texture, detecting a faint star, or measuring a small manufactured edge.

That order keeps the topic practical. You start with the clean physics, then add the messier parts of the real imaging system. If you want a broader signal-processing companion to the PSF discussion, convolution and filtering in images and signals explains the same blur-and-frequency relationship outside the specific optics setting.

Recap

The Airy disk is the smallest point pattern a perfect circular optical system can form. It exists because light diffracts through a finite aperture, not because the lens is defective. The spot gets smaller with shorter wavelength and larger aperture, and it gets larger with longer wavelength or higher f-number.

That one pattern explains the rest of the article. Two nearby Airy patterns overlap, which gives Rayleigh-style resolution. Many overlapping point patterns create image blur, which is the PSF view. Repeated fine detail loses contrast as the PSF gets wider, which is the MTF view.

The formulas are compact, but the mental model is simple: a finite aperture turns each point into a small structured spot. Resolution is what happens when those spots overlap.