draft · v0.1.226
💬Comments welcome. To leave a note, select any text and click the note / highlight button that pops up — or open the panel with the tab at the top-right (‹). Notes are visible only inside our private review group.
💡 In a hurry? Jump to this chapter’s 8 big lessons ↓

2.7 Lens Image formation⧉

A pinhole forms an image by throwing away almost all the light; only the rays through the tiny hole survive. That is a serious problem: a pinhole sharp enough to make a crisp image passes so little light that exposures take seconds, and (as the previous part's diffraction limit warned) shrinking the hole eventually blurs the image anyway. We want to collect a wide cone of light from each scene point, far more light than a pinhole, and still focus it to a single image point. That is exactly what a lens does, and the price it charges is a single plane of focus and a finite depth of field.

fig-lens-imaging
Figure 2.7.1. The third imaging scenario, and the sum of the first two. Where the bare sensor (Figure 2.7.1 of the previous chapter) averaged every direction and the pinhole (Figure 2.7.1b there) admitted a single ray per point, a lens gathers a whole cone of rays from each scene point, canopy (red), mid-trunk (blue), base (gold), and focuses them back to one photosite. The image is the same inverted one as the pinhole's, but now bright and sharp, because (almost) all of each point's light is used instead of thrown away. The bill (one plane of focus and a finite depth of field) comes due over the rest of this chapter.

The two image-formation playgrounds from the pinhole chapter (Figures 2.4.4 and 2.4.5) return here, now opened on the thick lens rather than the bare sensor. They carry the same four scenarios as before (bare sensor, pinhole, thin lens, thick lens); flip between them and notice how much brighter and sharper the lens images are than the pinhole's, and far more so than the bare sensor's flat average. The lens earns that by gathering a whole cone of rays from each point instead of the single ray a pinhole keeps.

fig-image-formation-lens
Figure 2.7.2. The 2-D image-formation playground again, opened on the thick lens. Same four scenarios as in the pinhole chapter (bare sensor, pinhole, thin lens, thick lens): switch among them and watch the image go from a flat average (bare sensor) to dim but sharp (pinhole) to bright and sharp (lens), because the lens collects a whole cone of rays per point instead of one. Click a pixel to see the cone it gathers, or a scene point to see where its rays land. Notice how much more light reaches each pixel once a lens replaces the pinhole.
fig-image-formation-3d-lens
Figure 2.7.3. The same experiment in 3-D, on the thick lens: orbit the scene, click the sensor to see the cone of rays one pixel now gathers, and read the recorded image in the inset. Compare the lens against the pinhole and the bare sensor with the scenario buttons, and notice how much brighter and sharper the lens image is while it stays in focus.

Recall the framing from the pinhole chapter: image formation is about choosing which rays contribute to each image point (→ see Big lesson L2.20). The pinhole answered with one ray per point; the lens gives a richer answer, and it is worth stating as its own lesson.

💡 Big lesson (L2.27) — a lens makes every ray from an in-focus point reconverge to one image point

Where a pinhole keeps only one ray, a lens gathers the whole cone of rays leaving a scene point and bends them back to a single image point, but only for points at the focus distance. That is the lens's bargain: vastly more light (a bright image) in exchange for a single plane of sharp focus, with everything nearer or farther spread into a blur disk (the circle of confusion). The aperture sets how wide that cone is, and hence the depth of field. Every later idea about focus, bokeh, refocusing, and light fields is a variation on "which cone of rays reconverges where."

This chapter follows that bargain in three steps. First, a single lens bending light by refraction: Snell's law at two curved surfaces, with the Fermat "equal optical path" view alongside. Then the thin lens: the small-angle linearization of that law that hands us the focus equation, conjugates, magnification, and the f-number in a few clean lines. Finally depth of field: the range of distances a single focus plane keeps acceptably sharp, the circle of confusion that defines it, and the surprising fact that (at fixed framing) it barely depends on focal length at all. Throughout, we are using the paraxial model; what it throws away (aberrations, color) points forward to the OPTICS part.

2.7.1 Single lenses: refraction and Snell's law⧉

A real lens is a piece of glass bounded by two curved surfaces, that is, two interfaces: air-to-glass and glass-to-air. It works by refraction: the bending of light as it crosses between media of different refractive index, introduced in Light and physics. At each interface a ray bends according to Snell's law (Descartes' law, if you are French):

$$ n_1 \sin\theta_1 = n_2 \sin\theta_2, $$

where $n_1, n_2$ are the refractive indices on the two sides and $\theta_1, \theta_2$ the angles measured from the surface normal. Drawn on a deliberately thick lens, with the two interfaces far apart, you can watch the bending happen twice (once entering the glass, once leaving) and the net effect of the two curved surfaces is to steer a parallel bundle of rays so they converge to a focal point (Figures 2 and 3). A curved interface bends each ray by an amount that depends on where it strikes the surface, and the curvature is chosen so that all the rays from one scene point are redirected to meet at one image point. That is the whole job of a lens: bend a wide cone back to a point.

fig-single-lens-refraction
Figure 2.7.4. A single lens focusing by refraction at two interfaces. A lens is glass with two curved surfaces. Parallel incoming rays refract by Snell's law n₁ sin θ₁ = n₂ sin θ₂ at the first (air→glass) interface and again at the second (glass→air) interface; the two bends together converge the parallel rays to a focal point. The thin-lens idealization (collapsing both bends into one plane) is overlaid for comparison.
fig-snell-two-interface
Figure 2.7.5. Snell's law at both surfaces of a thick lens. A ray arriving parallel to the axis refracts at the curved front surface (air→glass), crosses the glass, and refracts again at the curved back surface (glass→air); the angle of incidence and angle of refraction are marked at each interface, measured from the local radial surface normal, with the indices n₁ (air) and n₂ (glass). The two refractions add up to bend the ray toward the optical axis, where it meets the focus. A faint second ray shows the convergence. This is the accurate two-surface picture that the thin lens later collapses into a single plane.
fig-xkcd-1791
Figure 2.7.6. xkcd #1791 "Refractor vs Reflector" — Randall Munroe, xkcd.com/1791 (CC BY-NC 2.5).

A second, equivalent view: Fermat and equal path lengths. There is a wave-side way to see why a lens focuses, worth holding alongside the ray picture because it explains why the curvature has to be just so. A lens focuses because every ray from the object to the focus travels the same optical path length, geometric distance weighted by the refractive index of each medium it crosses (Figure 2.7.10). A ray through the thick center of the lens takes a longer geometric path but spends more of it in slow glass; a ray skimming the thin edge takes a shorter geometric path, almost entirely through fast air. The lens's shape is exactly what makes these trade off perfectly, so all the rays arrive at the focus in phase and interfere constructively, building a bright point. Snell's law (rays bending) and Fermat's principle (equal optical paths) are two descriptions of the one fact; either predicts the same focus.

fig-fermat-equal-path
Figure 2.7.7. The Fermat (equal-path) view of focusing. Several rays leave one object point, pass through different parts of the lens, and meet at the focus. The central ray's longer geometric path is spent partly in slow glass; the edge ray's shorter path is almost all in fast air. The lens shape equalizes the optical path length (geometric length × refractive index) of every route, so all rays arrive in phase and interfere constructively. The wave counterpart of the Snell's-law ray picture.

How strongly a lens bends light (its focal length) is set by the curvature of its two surfaces and the glass index, through the lensmaker's equation:

$$ \frac{1}{f} = (n - 1)\left(\frac{1}{R_1} - \frac{1}{R_2}\right), $$

with $R_1, R_2$ the radii of curvature of the two surfaces. More curvature, or a higher index, means a shorter focal length (stronger bending). We can watch this happen directly by solving the wave equation: send a point source's diverging wave into a slab of glass and let the wave evolve (Figure 2.7.8). Because the glass is thickest on the axis, it slows the middle of each wavefront most, and that is exactly enough to bend a diverging wavefront into a converging one, which collapses to an image point on the far side. The focus lands where the lensmaker's and imaging equations say it should, and tightening the curvature or raising the index pulls it inward, just as $1/f=(n-1)\,2/R$ predicts.

fig-thick-lens-wave-sim
Figure 2.7.8. A thick lens focusing, simulated. A live 2-D solution of the wave equation: a point source on the left radiates circular waves that cross a biconvex glass slab in which the wave travels slower (index $n$). The glass is thickest on the axis, so it retards the center of each wavefront most, turning the diverging wave into a converging one that collapses to an image point, the Fermat/equal-path view made literal. The focus obeys the lensmaker's equation $1/f=(n-1)\,2/R$ and the imaging equation $1/v=1/f-1/u$, both shown live on the figure. Interactive: sweep the surface curvature $R$, the glass index $n$, and the source distance $u$ and watch the focus move. The lens sits in an opaque mount so all light goes through the glass; absorbing borders, so nothing reflects off the window edges.

This whole picture, though, already assumes the rays make small angles with the axis: it is paraxial. At large angles the rays from a single point do not all meet at exactly one place, and the image degrades; those departures are aberrations, the subject of the OPTICS part, where compound lenses (many elements) are designed to cancel them. Our own eyes are refracting systems too: the cornea and the lens are the two main bending elements, doing for the retina what a camera lens does for the sensor.

2.7.2 Thin lens optics⧉

The two-interface, Snell's-law lens is accurate but cumbersome. For everything in this book except the dedicated optics chapters, we replace it with a linearization called the thin lens, and two assumptions do all the work. First, the rays stay near the axis and make only small angles, so Snell's law can be linearized: for small $\theta$, $\sin\theta\approx\theta$, the angle and its sine agree and the bending becomes proportional to the angle. Second, the lens is thin enough that we can ignore how far the light travels inside the glass, so both refractions collapse into a single bend at one plane.

Be candid about what that buys and what it costs (Figure 2.7.9). The small-angle (paraxial) swap is exactly why the focus equation $\tfrac{1}{f}=\tfrac{1}{u}+\tfrac{1}{v}$ (below) comes out so clean, and why "every ray from one object point converges to one image point" is true in the model: with $\sin\theta\approx\theta$ a single object point images to a single, perfectly sharp point (stigmatic imaging), and a parallel bundle from infinity converges to one point in the focal plane. A real lens only approximates this. Rays that wander too far from the axis miss the ideal point, the departures we call aberrations (spherical aberration, coma, and the rest). And the model drops color: the real index $n$ varies with wavelength (dispersion), so red and blue bend by slightly different amounts and focus in slightly different places (chromatic aberration), while the thin lens pretends every wavelength shares one $n$ and one focus. So the thin lens is a convenient lie, linearizing Snell's law and throwing away color, accurate enough that the whole rest of the book leans on it and wrong in precisely the ways the OPTICS chapters spend their time correcting.

The image-formation figures at the start of this chapter make this payoff tangible: put Figure 2.7.2 or Figure 2.7.3 on the thin or thick lens and click a point to watch its whole cone of rays reconverge onto a single pixel, far brighter than the pinhole's lone ray.

💡 Big lesson (L2.28) — when a quantity is small, linearize: keep the first Taylor term

What we just did to Snell's law is one of the most powerful moves in all of physics and engineering, and it is worth naming as a general tool, because the book will reach for it again and again. Replace a messy nonlinear function by its tangent line near the operating point, keep only the first-order (linear) term of its Taylor expansion, and it is an excellent approximation as long as the relevant quantity stays small. Here the small quantity is the angle, $\sin\theta\approx\theta$, and the reward is the entire clean, linear machinery of paraxial optics. The same reflex linearizes brightness change into the optical-flow constraint (a chapter away), drives the Gauss–Newton steps of lens optimization and bundle adjustment, and propagates small noise through a pipeline. Two disciplines keep it sound. Know when it breaks: the approximation is only as good as the terms you dropped are negligible, so always ask "how small is small?" And keep the next term in your pocket: restoring it is not a footnote but a whole theory: for optics that next term is the cubic $-\theta^3/6$, and restoring it is the study of aberrations (next big lesson, and the OPTICS part).

💡 Big lesson (L2.29) — thin-lens optics is a linear (first-order) approximation of Snell's law when angles are small (the paraxial regime)

The clean thin-lens equation $1/f = 1/u + 1/v$ is not a separate law: it is Snell's law (L2.4) linearized, kept to first order under the small-angle approximation $\sin\theta\approx\theta$, the regime known as paraxial (rays near the axis, gentle angles, one wavelength). Because it keeps only the linear term, it tells you exactly when the idealization holds and exactly how it breaks: restore the next term $\sin\theta\approx\theta-\theta^3/6$ and the Seidel aberrations appear; let $n$ vary with wavelength and chromatic aberration appears. "Thin lens," "first-order optics," "a linear (paraxial) lens," and "a lens with no aberrations" are all the same approximation, and the OPTICS part is the story of paying back what it borrowed.

fig-thin-lens-approximation
Figure 2.7.9. The thin lens is the small-angle linearization of a real, thick, Snell's-law lens (interactive). A biconvex glass lens is traced two ways: the correct optics (refract by Snell's law at each curved surface, for three wavelengths whose index differs by the dispersion) and the thin approximation (bend once at the center plane, one index, dashed). Near the axis and for one color they agree; away from the axis the correct rays focus short of the ideal point (spherical aberration) and the colors split (chromatic aberration), exactly what the thin lens throws away. Sweep the thickness, curvature, index, dispersion, and incoming tilt.

Under these two assumptions the geometry collapses to bending at a single plane, and two construction rules locate any image:

  1. A ray arriving parallel to the axis leaves through the focal point $F$, at distance $f$.
  2. A ray through the center of the lens passes undeviated, straight through, as through a pinhole. (This second rule is why a thin lens has the same perspective as a pinhole: the central ray is the chief ray, so the projection geometry of the previous chapter carries over unchanged. The lens only adds light-gathering and focus on top of pinhole projection.)
💡 Big lesson (L2.30) — the two thin-lens ray rules

Parallel rays through a thin lens converge to a point at the focal length, and rays going through the center of a thin lens do not deviate. The first rule is what focal length means (where a far subject's parallel bundle meets) and carries the lens's focusing power; the second carries its perspective, identical to the pinhole's. Between these two construction rules and similar triangles, you can derive all of thin-lens optics.

fig-thin-lens-two-rules
Figure 2.7.10. The two thin-lens ray rules (interactive), in the order of the lesson. Left (rule 1): two bundles of parallel rays, an on-axis bundle meeting at the focal point F and a tilted bundle meeting off-axis, each converging to a single point on the focal plane a distance f behind the lens: the focusing power, and the meaning of focal length. Right (rule 2): a bundle of rays passing through the center of the lens, every one undeviated (straight through, as through a pinhole): the lens's perspective, identical to a pinhole's. Tilt the second bundle and set f on the left; highlight a center ray on the right. Together with similar triangles these two rules give all of thin-lens geometry.
Margin note — "linearization" and "no aberrations" are the same approximation

The thin lens is precisely first-order (paraxial) optics: replace $\sin\theta \approx \theta$, keeping only the linear term of the sine. Keep the next term, $\sin\theta \approx \theta - \theta^3/6$ (third-order optics), and out pop the classical Seidel aberrations: spherical aberration, coma, astigmatism, and the rest. So "the thin-lens linearization" and "a lens with no aberrations" are literally the same assumption, dropped at the same place. The OPTICS part keeps the cubic term.

The focus equation. Place an object at distance $u$ in front of the lens; its sharp image forms at distance $v$ behind, and the two are tied by the thin-lens (conjugate) equation:

$$ \frac{1}{f} = \frac{1}{u} + \frac{1}{v}. $$
fig-thin-lens
Figure 2.7.11. The thin-lens model and its focus equation (interactive). The two interfaces of a real lens are collapsed into one plane; small angles are assumed. An object at distance u images to an inverted picture at distance v, located by two rays: a ray parallel to the axis exits through the focal point F at distance f, and a ray through the lens center passes undeviated; their intersection is the image tip. The undeviated central ray means the thin lens shares the pinhole's perspective. Distances are in millimeters over a normal-lens focal range (35 to 70 mm); the bars underneath show 1/u + 1/v = 1/f made visual, the top bar (1/u then 1/v, end to end) always exactly as long as the 1/f bar below it, on a fixed scale set by the shortest focal length.
💡 Big lesson (L2.31) — the focus equation is symmetric: object and image are interchangeable conjugates

The thin-lens equation $\tfrac{1}{u}+\tfrac{1}{v}=\tfrac{1}{f}$ is symmetric in $u$ and $v$: swap them and it is unchanged, so object and image are interchangeable. Send the light back the other way and the rays retrace exactly the same path (the reversibility of light, made arithmetic). Two consequences fall straight out: an object at infinity images at $v=f$, which is what focal length means, and focusing a camera is nothing but moving $v$ until the chosen object's conjugate lands on the sensor.

💡 Big lesson (L2.58) — the simplest way to focus is to move the sensor: full (unit) focusing

The most basic way to focus follows straight from the equation: move the whole lens relative to the sensor, changing the lens-to-sensor distance $v$ until the chosen subject's conjugate lands on the sensor. This is full or unit focusing, the entire optical unit sliding with its focal length fixed. A bellows, the helicoid focus ring on a simple prime, and a bare enlarger lens all work this way, and it is the mental model to keep. It is only the simplest option, though. A compound lens has many elements, and moving just some of them (front-cell, internal, or rear focusing, sometimes with floating groups) can focus while extending the barrel less, holding aberrations in check across the focus range, or even nudging the effective focal length. Those techniques, and why designers reach for them, come in the OPTICS part's focus section.

Read it back: the reciprocals of the object and image distances add to the reciprocal of the focal length. An object at infinity ($1/u \to 0$) images at exactly $v = f$: that is what "focal length" means for a lens, the image distance for a far subject. As the object comes closer, $1/u$ grows, so $1/v$ must shrink: the image moves farther back. Focusing a camera is nothing more than adjusting $v$ (moving the lens, or the sensor) until the image of the subject you care about lands exactly on the sensor (Figure 2.7.12). Each object distance has exactly one image distance that focuses it (they are conjugates) and an object placed right at the focal point $F$ sends out rays that exit parallel, imaging at infinity (the reverse of the first case).

Symmetry and reversibility. Notice that the focus equation is symmetric in the two distances: writing it $1/f = 1/s_o + 1/s_i$, with $s_o$ the object distance and $s_i$ the image distance, nothing distinguishes the two: swap them and the equation is unchanged. This is no accident; it is the reversibility of light paths made arithmetic. Object and image are conjugate: send light the other way, and the rays retrace exactly the same path, so the old image plane now acts as object and the old object plane as image. The two are conjugate planes, and the magnification swaps with them: $m = -s_i/s_o$ simply inverts ($m \to 1/m$) when you exchange the roles. This object↔image reciprocity runs all through optics: it is why a lens images perfectly well either way round, and (a case we meet in the OPTICS part) why a lens's entrance and exit pupils are themselves conjugates: images of the same aperture seen from the two sides.

fig-thin-lens-conjugates
Figure 2.7.12. Thin-lens conjugates. For object distance u and image distance v, 1/f = 1/u + 1/v. An object at infinity images at v = f; a far object just beyond f; a close object far behind the lens; an object at the focal point F sends rays out parallel (image at infinity). Each object distance pairs with exactly one image distance: focusing a camera moves the sensor/lens to put the chosen conjugate on the sensor.

The image is inverted and resized; the magnification is

$$ m = -\frac{v}{u}, $$

the ratio of image to object distance (the minus sign records the flip).

💡 Big lesson (L2.32) — magnification is the ratio of projected size to real-world size

Magnification is the size of a subject's image divided by the subject's actual size in world units (meters): $m = -v/u$ is exactly that ratio (image height over object height). Photographers quote it as a reproduction ratio: 1:1 ("life size") means the image on the sensor is as large as the subject itself; 1:2 means half life size. Crucially, what fraction of the frame a subject then occupies depends on the sensor size: the same magnification fills more of a small sensor than a large one. On a full-frame ($24\times36$ mm) camera a $1{:}2$ reproduction images a subject about $48\times72$ mm to exactly fill the frame; put that same subject on a smaller sensor at the same magnification and it overflows. So magnification is a property of the optics-and-distance alone, while "how big it looks in the picture" also needs the sensor size.

Finally, the lens has an aperture (the diameter $D$ of the opening that admits light) and photographers describe it not by $D$ directly but by the f-number

$$ N = \frac{f}{D}, $$

the ratio of focal length to aperture diameter. A small f-number ($f/1.4$) is a wide opening; a large one ($f/16$) is a narrow opening: the number is under the divider, so it runs backwards from how much light gets through. The f-number controls two things at once, exposure and blur. We meet the blur in the Depth of field chapter next; the aperture and the f-number get a fuller treatment at the end of this chapter (why the ratio makes brightness portable, and the $\sqrt2$ stop series), and the exposure practice built on it is taken up in Exposure.

The same focus equation quietly explains why macro photography is hard. Rewrite it as $1/v = 1/f - 1/u$: to focus a close subject, $u$ shrinks, so $1/u$ grows, so $1/v$ has to shrink, which means $v$, the lens-to-sensor distance, has to grow. High magnification is the same statement, since $m=-v/u$: getting to 1:1 (life size) requires $v=u=2f$, twice the infinity-focus distance, and going beyond 1:1 pushes $v$ larger still. An ordinary lens simply cannot rack its focusing helicoid out that far. The fix is to physically move the lens farther from the sensor, which is exactly what an extension tube (a hollow, glassless spacer) or a bellows (a continuously adjustable one) does (Figure 2.7.13): they add nothing optical, they just buy the large $v$ the equation demands. This is also why a lens's closest focus distance and its maximum magnification are the same spec seen two ways, and why dedicated macro lenses build long focusing extensions into the barrel.

fig-extension-tube
Figure 2.7.13. An extension tube: spacing for close focus. A glassless spacer inserted between lens and camera body increases the lens-to-sensor distance $v$. Because $1/u+1/v=1/f$, focusing close (small $u$) demands a large $v$, and the extra $v$ raises the magnification $m=-v/u$ toward 1:1 and beyond — which is precisely why macro work needs extension tubes or bellows rather than glass alone. (Author's photograph.)

2.7.3 Aperture and the f-number⧉

We met the f-number $N = f/D$ a moment ago; it is worth a section of its own, because it is one of the few numbers a photographer touches every time they shoot, and because why it is written as a ratio is genuinely illuminating. The aperture is a physical thing, the diameter $D$ of the opening that admits light. Yet lenses are never marked in millimetres of opening; they are marked $f/2.8$, $f/5.6$, $f/11$, the focal length divided by that diameter. The slash is literal: $f/2.8$ means "the focal length, split into 2.8," an opening $D = f/2.8$ wide. Because $D$ sits in the denominator, the scale runs backwards from intuition, a small f-number is a wide opening (lots of light, shallow depth of field) and a large f-number is a narrow one. Once you internalize that the number is a divisor, "$f/1.4$ is bright and $f/22$ is dim" stops feeling backwards.

Why quote the opening as a fraction of focal length at all, rather than its raw diameter? Because the fraction, not the diameter, is what fixes the brightness of the image. The light delivered to a point on the sensor arrives as a cone: the rays that clear the aperture and reconverge there. The wider that cone, the more light. Its half-angle $\theta$ satisfies $\tan\theta = (D/2)/f = 1/(2N)$, so the cone depends only on the ratio $N$, never on the focal length by itself (Figure 2.7.14). Double the focal length and double the diameter, and the opening is physically twice as big, yet the cone, and therefore the brightness, is unchanged, because $N$ is unchanged. This is the whole reason the ratio exists: it makes a brightness setting that transfers. Set any lens to $f/2.8$, a $24$ mm or a $600$ mm, and each throws the same irradiance onto the sensor, even though the $600$ mm's opening is enormous by comparison. A raw diameter could never do that; the ratio nets out the focal length exactly. (The transfer is only mostly exact: vignetting darkens the corners, oblique rays fall off toward the frame edge, and at close focus the effective f-number opens up to $N_\text{eff} = N(1+m)$ with magnification $m$, the "bellows factor" behind macro's light loss. The geometry above is the clean first-order truth those are corrections to.)

fig-aperture-fnumber-cones
Figure 2.7.14. Why the f-number is written as a ratio, read off the light cone. Left: the same f-number gives the same cone, and hence the same brightness, at any focal length. A short lens (aperture $D$ at distance $f$) and a long lens (aperture $2D$ at distance $2f$) have the identical $N = f/D = 4$, so their converging cones share the very same half-angle $\tan\theta = 1/(2N)$; the long lens's opening is physically twice as wide yet pours no more light onto the sensor point, because only the ratio matters. Right: at a fixed focal length, opening up (a smaller f-number) widens the cone. The light collected follows the opening's area, $\propto D^2 \propto 1/N^2$, so $f/2, f/4, f/8$ pass $1, \tfrac14, \tfrac1{16}$ of the light; halving the area (one stop) takes a $\sqrt2$ step in $N$, which is the whole reason the aperture series climbs $f/2, 2.8, 4, 5.6, 8, \dots$

The same cone picture explains the aperture scale's strange-looking numbers, $f/1.4, f/2, f/2.8, f/4, f/5.6, f/8, f/11, f/16$, a march in steps of $\sqrt2$. Light gathered goes as the opening's area, which goes as the diameter squared, so at a fixed focal length the light scales as $D^2 \propto (f/N)^2 \propto 1/N^2$. A stop, photography's unit, is a factor of two in light. To halve the light you must halve the area, and halving an area means shrinking the diameter by $\sqrt2$, i.e. multiplying $N$ by $\sqrt2$. That is the entire origin of the series: each full stop is one $\sqrt2$ step in f-number and one factor of two in light (the figure; and panel B of Figure 2.7.14). It is why the numbers look irregular but are really a clean geometric progression, and why aperture, shutter, and ISO can all be counted in the same stops (the exposure triangle of Exposure).

💡 Big lesson (L2.60) — the f-number is the aperture diameter as a fraction of the focal length

$N = f/D$: the f-number tells you the width of the opening measured in focal lengths, not in millimetres. $f/4$ means the opening is one quarter of the focal length across. Writing the aperture this way is what makes it a portable brightness setting: because the cone of light converging on each sensor point has half-angle $\tan\theta = 1/(2N)$, the ratio alone sets the image brightness, so the same f-number gives the same exposure on any lens, whatever its focal length or physical size. The denominator is why a small number is a big opening, and the area-goes-as-diameter-squared law ($\text{light}\propto 1/N^2$) is why the stop series climbs by $\sqrt2$.

Finally, a caveat that matters when light is precious. The f-number is a purely geometric quantity: it assumes a perfect, lossless lens and predicts the light such a lens would gather. Real glass is not lossless, every air-to-glass surface reflects a little, the elements absorb a little, and a complex zoom with twenty elements can quietly swallow up to a stop before any light reaches the sensor. The T-stop (transmission stop) folds those losses in, the f-number of an ideal lossless lens passing the same light, $T = N/\sqrt{\tau}$ for transmittance $\tau$, so a lens engraved $f/2.0$ may test at $T/2.3$. Stills photographers can ignore the difference; cinematographers, who must intercut shots from different lenses at a matched brightness, buy lenses marked in T-stops. The full treatment, and why it matters for metering, is in Exposure (f-stops versus T-stops).

Recap: big lessons of this chapter

(L2.27) — a lens makes every ray from an in-focus point reconverge to one image point

Where a pinhole keeps only one ray, a lens gathers the whole cone of rays leaving a scene point and bends them back to a single image point, but only for points at the focus distance. That is the lens's bargain: vastly more light (a bright image) in exchange for a single plane of sharp focus, with everything nearer or farther spread into a blur disk (the circle of confusion). The aperture sets how wide that cone is, and hence the depth of field. Every later idea about focus, bokeh, refocusing, and light fields is a variation on "which cone of rays reconverges where."

(L2.28) — when a quantity is small, linearize: keep the first Taylor term

What we just did to Snell's law is one of the most powerful moves in all of physics and engineering, and it is worth naming as a general tool, because the book will reach for it again and again. Replace a messy nonlinear function by its tangent line near the operating point, keep only the first-order (linear) term of its Taylor expansion, and it is an excellent approximation as long as the relevant quantity stays small. Here the small quantity is the angle, $\sin\theta\approx\theta$, and the reward is the entire clean, linear machinery of paraxial optics. The same reflex linearizes brightness change into the optical-flow constraint (a chapter away), drives the Gauss–Newton steps of lens optimization and bundle adjustment, and propagates small noise through a pipeline. Two disciplines keep it sound. Know when it breaks: the approximation is only as good as the terms you dropped are negligible, so always ask "how small is small?" And keep the next term in your pocket: restoring it is not a footnote but a whole theory: for optics that next term is the cubic $-\theta^3/6$, and restoring it is the study of aberrations (next big lesson, and the OPTICS part).

(L2.29) — thin-lens optics is a linear (first-order) approximation of Snell's law when angles are small (the paraxial regime)

The clean thin-lens equation $1/f = 1/u + 1/v$ is not a separate law: it is Snell's law (L2.4) linearized, kept to first order under the small-angle approximation $\sin\theta\approx\theta$, the regime known as paraxial (rays near the axis, gentle angles, one wavelength). Because it keeps only the linear term, it tells you exactly when the idealization holds and exactly how it breaks: restore the next term $\sin\theta\approx\theta-\theta^3/6$ and the Seidel aberrations appear; let $n$ vary with wavelength and chromatic aberration appears. "Thin lens," "first-order optics," "a linear (paraxial) lens," and "a lens with no aberrations" are all the same approximation, and the OPTICS part is the story of paying back what it borrowed.

(L2.30) — the two thin-lens ray rules

Parallel rays through a thin lens converge to a point at the focal length, and rays going through the center of a thin lens do not deviate. The first rule is what focal length means (where a far subject's parallel bundle meets) and carries the lens's focusing power; the second carries its perspective, identical to the pinhole's. Between these two construction rules and similar triangles, you can derive all of thin-lens optics.

(L2.31) — the focus equation is symmetric: object and image are interchangeable conjugates

The thin-lens equation $\tfrac{1}{u}+\tfrac{1}{v}=\tfrac{1}{f}$ is symmetric in $u$ and $v$: swap them and it is unchanged, so object and image are interchangeable. Send the light back the other way and the rays retrace exactly the same path (the reversibility of light, made arithmetic). Two consequences fall straight out: an object at infinity images at $v=f$, which is what focal length means, and focusing a camera is nothing but moving $v$ until the chosen object's conjugate lands on the sensor.

(L2.58) — the simplest way to focus is to move the sensor: full (unit) focusing

The most basic way to focus follows straight from the equation: move the whole lens relative to the sensor, changing the lens-to-sensor distance $v$ until the chosen subject's conjugate lands on the sensor. This is full or unit focusing, the entire optical unit sliding with its focal length fixed. A bellows, the helicoid focus ring on a simple prime, and a bare enlarger lens all work this way, and it is the mental model to keep. It is only the simplest option, though. A compound lens has many elements, and moving just some of them (front-cell, internal, or rear focusing, sometimes with floating groups) can focus while extending the barrel less, holding aberrations in check across the focus range, or even nudging the effective focal length. Those techniques, and why designers reach for them, come in the OPTICS part's focus section.

(L2.32) — magnification is the ratio of projected size to real-world size

Magnification is the size of a subject's image divided by the subject's actual size in world units (meters): $m = -v/u$ is exactly that ratio (image height over object height). Photographers quote it as a reproduction ratio: 1:1 ("life size") means the image on the sensor is as large as the subject itself; 1:2 means half life size. Crucially, what fraction of the frame a subject then occupies depends on the sensor size: the same magnification fills more of a small sensor than a large one. On a full-frame ($24\times36$ mm) camera a $1{:}2$ reproduction images a subject about $48\times72$ mm to exactly fill the frame; put that same subject on a smaller sensor at the same magnification and it overflows. So magnification is a property of the optics-and-distance alone, while "how big it looks in the picture" also needs the sensor size.

(L2.60) — the f-number is the aperture diameter as a fraction of the focal length

$N = f/D$: the f-number tells you the width of the opening measured in focal lengths, not in millimetres. $f/4$ means the opening is one quarter of the focal length across. Writing the aperture this way is what makes it a portable brightness setting: because the cone of light converging on each sensor point has half-angle $\tan\theta = 1/(2N)$, the ratio alone sets the image brightness, so the same f-number gives the same exposure on any lens, whatever its focal length or physical size. The denominator is why a small number is a big opening, and the area-goes-as-diameter-squared law ($\text{light}\propto 1/N^2$) is why the stop series climbs by $\sqrt2$.