Aperture, Equivalence, and Diffraction on Medium Format

Three concepts decide how every photograph actually renders. Circle of confusion sets where a focus decision becomes visible as blur. Equivalence governs how that behavior changes when you swap sensor sizes. Diffraction is the hard physical ceiling no glass can beat. They interact, and the interaction matters more at 100MP than at 24MP. A high-resolution sensor reveals what a lower-resolution one would hide.

What this page answers:

• What circle of confusion actually is, and why your hyperfocal calculator is making a hidden choice on your behalf.
• Why the same f-stop produces different depth of field, bokeh, and noise across sensor sizes, and why f/9.5 was right in Tuscany.
• Why diffraction takes over earlier on a 100MP body than most DoF apps assume.
• The actual sweet-spot range for the XCD V-series, the Sigma 300-600/4 Sport, and the Leica APO-Summicron 43.

Start with circle of confusion. Everything else builds on it.

Circle of Confusion: What Changes on a 100MP Sensor

The sensor doesn't move. That sounds obvious, but most explanations of focus and depth of field skip past it, and the resulting confusion is exactly what gives circle of confusion its name. When you turn the focus ring, you're not shifting the sensor inside the camera. You're shifting where light from a particular subject distance lands behind the lens. Whatever distance you've focused on converges to a point on the sensor. Everything else converges in front of it or behind it, and what hits the sensor is a disk.

That disk is the circle of confusion.

If you focus on a person standing at five meters, light from the person converges exactly at the sensor's surface and is recorded as a point. Light from a flower at two meters has its convergence point behind the sensor, so it hits the sensor mid-converge and is recorded as a disk. Light from a mountain at infinity has already converged in front of the sensor and is diverging again by the time it hits, also as a disk. Three subjects, three different-sized disks, one fixed sensor. The lens chooses which subject becomes the point.

Why "confusion"

The eye is forgiving. Below a certain disk size, you read the spot as a point even though it's still a small smear. Above that size, you start to see it as a blur. The threshold tracks the limit of human visual resolution at typical viewing distance, roughly five line pairs per millimeter on a print viewed from about 30cm with normal vision. Pick a disk size on the sensor that, after enlargement and viewing, lands below that threshold, and the photo reads as sharp.

The circle of confusion is exactly that tolerance. It's the maximum disk diameter you're willing to accept on the sensor while still calling the result "in focus."

Where the 1500 comes from

The standard convention is sensor diagonal divided by 1500. The number isn't arbitrary. It falls out of the same assumptions above: a print roughly 25cm wide (about 30cm diagonal) viewed from 30cm, with the visual system at five line pairs per millimeter. Run that backward through the enlargement factor from sensor to print, and you land on a constant of about 1500.

Plug in real sensors:

• Full frame (43mm diagonal): 0.029mm, usually rounded to 0.030mm
• APS-C (28mm diagonal): 0.019mm
• Micro Four Thirds (22mm diagonal): 0.015mm
• X2D II 44×33 (55mm diagonal): 0.037mm

Smaller sensors get a stricter CoC because they need more enlargement to reach the same print size. The same 0.030mm disk that's invisible on a full-frame print becomes a visible smear on a print made from a Micro Four Thirds frame, because the MFT print enlarges further. The CoC compensates: you tolerate less blur on the sensor in the first place.

The X2D II runs the math in the other direction. It needs less enlargement, so its tolerance for sensor-side blur is bigger. This is part of why medium format produces a more relaxed-looking image at base sharpness. You're enlarging less, so smaller imperfections don't propagate.

Traditional CoC vs Pixel-Level Sharpness on 100MP

The d/1500 number is calibrated for prints. For pixel-level sharpness it's wrong by a factor of about five.

A 100MP X2D II sensor has 3.77µm pixels. A modern CoC convention for pixel-peeping is 2× pixel pitch, which is 7.5µm. That's 4.8× stricter than the d/1500 traditional value. The numbers diverge sharply when you set hyperfocal distance: roughly 8.5m at 50mm f/8 by traditional CoC, roughly 41m at 50mm f/8 by modern CoC.

If you set focus to the traditional hyperfocal value and then open the file at 100% on a 5K display, the foreground will be mush. The traditional CoC was right for the print viewer of 1960. The modern CoC is right for the pixel-peeper of 2026. Both are valid. They just answer different questions. Pick the one that matches your actual output.

What this means at the camera

Circle of confusion is a tolerance, not a property of the lens. Pick traditional (d/1500) when you're optimizing for prints. Pick modern (≈ 2× pixel pitch) when you're optimizing for pixel-level sharpness on screen. Whatever your hyperfocal calculator is using, it's making this choice for you, often silently. Find out which.

The Light Cone

§1 described what happens. This section is why it happens, geometrically. If you want the cone-and-disk derivation, keep reading. If §1 was enough, skip ahead to the Tuscany section.

The diagram below isolates the light cone behind a lens for a single point source. The cone converges to a point at one specific distance behind the lens, set by the focus distance and the focal length. In a real camera, the sensor sits at exactly that distance for whatever subject the lens is currently focused on. The diagram instead lets you slide a virtual sensor along the optical axis to see what happens at every position.

This is not how a real camera works. The camera's sensor is fixed. What moves, when you focus, is where the cone converges behind the lens. The diagram inverts that: it freezes the cone and slides the sensor, because that makes the geometry visually obvious. Same physics, easier to see.

Two things drop out of the diagram:

The disk size scales linearly with the sensor's distance from the convergence point. Defocus your lens by a small amount and the disk is small. Defocus by a lot and the disk is large. The relationship is geometric, not optical.

A wider aperture makes a wider cone. The aperture diameter is the base of the cone. A smaller f-number means a larger physical opening, which means a wider cone, which means the same sensor offset produces a bigger disk. That's the geometric reason fast lenses deliver shallow depth of field. It's not a special property of the lens. It's the unavoidable consequence of having a wider light cone hit a sensor that isn't at the convergence point.

The next section uses that fact to explain why the same f-stop behaves differently across sensor sizes.

The Tuscany Equivalence: Equivalent Aperture on Medium Format

I joined a Tuscany landscape workshop with Andrea Livieri in 2026. He runs the same workshop in 2027 for anyone who wants to chase the same light. Every full-frame shooter on the trip had landed on f/8 for our morning sessions. I was at f/9.5 on the X2D II. A few of them looked over and politely suggested I was wrong. Same scene, same depth of field target, same lens decision tree. Why was I a third of a stop narrower?

Equivalent aperture is why.

The X2D II's 44×33mm sensor has a linear ratio of about 0.79 to a full-frame 36×24mm sensor. To compose the same shot from the same spot, the X2D II needs a longer focal length than full frame. Specifically, full-frame 50mm corresponds to X2D II 63mm for equal framing. That's the crop factor running in the opposite direction from the APS-C and Micro Four Thirds case most photographers are used to. Sub-FF systems have crop factors greater than one. Medium format has crop factors less than one.

That focal-length difference matters because of one definition: aperture diameter equals focal length divided by f-number. A 63mm lens at f/8 has a physical aperture opening of 7.9mm. A 50mm lens at f/8 has an opening of 6.25mm. Bigger physical aperture, wider light cone behind the lens, bigger out-of-focus disks for the same amount of defocus. Run the cone geometry from §2 and the result drops out: at equal framing, the X2D II at f/8 produces a wider light cone than a full-frame camera at f/8, so it gives shallower depth of field and more bokeh, not less.

The shorthand for this is "equivalent aperture": multiply your f-number by the crop factor (0.79 for X2D II) to get the full-frame number that produces matching depth of field and bokeh.

• X2D II at f/2.8 ≈ FF at f/2.2
• X2D II at f/4 ≈ FF at f/3.2
• X2D II at f/8 ≈ FF at f/6.3

Run that backward and you get the Tuscany answer. The full-frame shooters at f/8 had a particular depth of field. To match that DoF on the X2D II, I needed f/8 / 0.79 ≈ f/10. The exact click I took depends on the camera's aperture-step setting: in third-stops that's f/10, in half-stops that's f/9.5. I had the X2D II on half-stops, so I shot f/9.5. They were at f/8. Same final-image DoF.

The four equivalence rules

There are four questions you can ask about whether two cameras behave equivalently at a given f-stop. They do not all answer the same way.

Depth of field: equivalent. Multiply your f-number by the crop factor (0.79 for X2D II) to get the FF-equivalent. To match a particular FF DoF on medium format, divide instead: f/8 / 0.79 ≈ f/10 in third-stops, or f/9.5 in half-stops. This is the rule that put me at f/9.5 in Tuscany.

Bokeh quantity: equivalent. Same calculation. Bigger physical aperture diameter at the same framing produces bigger out-of-focus disks. The X2D II at f/2.8 delivers the same quantity of background blur as a full-frame camera at f/2.2 from the same shooting position. Bokeh quality (the shape and texture of out-of-focus regions, the smoothness of the transition from sharp to soft) is a lens-design problem, not a sensor-size problem. Smooth bokeh comes from apodization, aperture blade count, and corrected spherical aberration. It is not a medium format perk.

Total light gathered: equivalent in the math, and medium format wins. The X2D II sensor is roughly 1.7× the area of full frame. Even at f/9.5 to their f/8, my sensor caught more total photons than theirs did. The quiet advantage of medium format is photon collection at base ISO. Cleaner shadows. Richer tonal gradation. The "equivalent ISO" framing some photographers use is this same fact restated.

Exposure: not equivalent. F-number is a measure of light intensity at the sensor plane, regardless of sensor size. My f/9.5 received about a third of a stop less light per unit area than their f/8. I compensated with shutter speed or ISO and we all walked home with usable exposures. The equivalent-aperture math does not extend to exposure. Use the f-number on the lens as marked.

Equivalence handles the cross-sensor math. Diffraction is the physical ceiling no aperture trick gets past, and it lands harder on a 100MP body than on a 24MP one.

What this means at the camera

For matching depth of field across systems, multiply your f-number by the crop factor (or divide, depending on which direction you're matching). For exposure, use the f-number as marked. Total light captured tells the noise story, and medium format has more sensor area to gather it. Forget any rule that conflates these.

Diffraction on Medium Format: The Physics Ceiling

Diffraction is pure physics. Light passing through any finite aperture spreads as a wave, and the smaller the opening, the more it spreads. The result on the sensor is not a perfect point but a small disk, the Airy disk, surrounded by a faint set of rings. The size of that disk depends on two things: the wavelength of the light, and the f-number of the lens. Nothing else.

The math, written out:

Airy disk diameter ≈ 2.44 × λ × f-number

For green light (λ ≈ 550nm, the wavelength the eye is most sensitive to and the one most sharpness measurements assume), the disk grows with f-number on a fixed schedule: about 10.7µm at f/8, 14.7µm at f/11, 21.5µm at f/16, 29.5µm at f/22. There is no glass on Earth that can shrink any of those numbers.

That's the hard ceiling. A Leica APO Summicron at f/8 produces an Airy disk identical in size to a kit zoom at f/8, because diffraction does not care about lens design. The world's most expensive piece of glass and the cheapest one in the same f-stop bin are equally diffraction-limited.

What glass quality changes is whether you ever reach that ceiling.

Aberrations vs diffraction

A real lens has aberrations. Spherical, chromatic, coma, astigmatism, field curvature: each is a way that light from a single subject point fails to land on a single sensor point. They all blur point sources, on top of the diffraction limit.

At wide apertures, aberrations dominate. The lens is using the full diameter of its glass, including the optically worst outer regions, so spherical and coma aberrations are large. The Airy disk is small (because the f-number is low), but it is swimming in a sea of aberration blur, and what you see on the sensor is the aberration, not the diffraction.

Stop down. The lens uses progressively smaller central regions of its glass, where the optics are better corrected, so aberrations shrink. The Airy disk grows, slowly. At some point the two cross. That crossover is what photographers call the lens's "sweet spot."

Past the crossover, you are trading aberration improvement (which you have already spent) for diffraction softening (which only gets worse). Stopping down further makes the image less sharp, not more.

Better glass loses to diffraction earlier

This is the part most photographers get backward.

A cheap lens has lots of aberration to clean up by stopping down. It might not reach its sweet spot until f/8, because aberrations are still significant at f/4 and f/5.6. By f/11 or even f/13 it is still aberration-limited and the slowly-growing Airy disk is not yet large enough to take over. So a cheap lens has a wide sweet spot, centred around a relatively narrow aperture.

A diffraction-limited lens is essentially aberration-free from wide open. There is nothing to clean up. As soon as you start stopping down, you are spending Airy disk for nothing. The XCD V-series primes (38V, 55V, 90V) are this kind of lens. So is the Leica APO Summicron 43 inside the Q3 43. Their sweet spot is wide open through f/5.6, with f/8 already costing per-pixel detail on a 100MP sensor. Stop them down to f/11 and you are worse off than wide open.

The shorthand: the better the glass, the narrower the sweet spot, and the earlier diffraction takes over.

Why the 100MP Sensor Is the Measuring Stick

A sensor does not cause diffraction or aberrations. It just resolves them, or fails to.

A 24MP full-frame sensor has roughly 6µm pixels. The difference between a 7µm Airy disk (around f/5.6) and a 12µm Airy disk (around f/9) is invisible at that pixel pitch. Both are at or below the sensor's two-pixel resolution limit. The image looks equally sharp at f/5.6 and f/9 because the sensor cannot resolve the difference. Photographers shooting a 24MP body see no penalty for stopping down to f/11 or f/13, and conventional wisdom on those bodies reflects that.

A 100MP X2D II sensor has 3.77µm pixels, roughly half the linear size, four times the pixel count over the same area. Now the 7µm and 12µm Airy disks are cleanly distinguishable. The sensor can see diffraction and aberrations both, and the cost of stopping down past the sweet spot becomes visible at 100% on screen and on a large print.

Same lens, same physics, same diffraction. The difference is whether the sensor can resolve it. A 100MP medium format body brings diffraction back into the conversation in a way that 24MP and 36MP bodies historically muted.

What this means at the camera

For the X2D II with V-series primes, your sweet spot is wider than you think and starts at wide open. Most well-designed lenses on a 100MP body already cost per-pixel detail by f/8. For a long zoom with more aberrations to stop down out of, push further: the Sigma 300-600 on the SL3 stays sharp deeper into the f-stop range than any XCD lens. The 100MP sensor is the measuring stick that makes both effects visible. A 24MP body would have hidden them.

Lens Sweet Spot and the Diffraction Threshold

Diffraction is universal. Aberrations are per-lens. The practical question is: where does each of your lenses cross over from aberration-limited to diffraction-limited? That answer is the only one that matters for shooting decisions, and it is not on the lens barrel.

Here are three lenses I shoot, with the crossover point spelled out for each.

XCD V-Series Sweet Spot (38V, 55V, 90V) on the X2D II

The Hasselblad XCD V-series of primes are essentially diffraction-limited from wide open on the X2D II. Photography Life's lab review of the 90V called it "an extremely sharp lens," with peak performance close to wide open. Jim Kasson's Siemens star measurements on the 90V show diffraction "seriously affecting sharpness at f/11" on the X2D II's 100MP sensor.

Operationally:

• Wide open through f/5.6: the sweet spot. Aberrations are already negligible. The Airy disk is small relative to the pixel pitch.
• f/8: still excellent, but diffraction is taking a measurable bite at 100% on a 100MP file.
• f/11 and beyond: noticeable softening at pixel level. Kasson's data flags f/11 as the threshold where diffraction becomes the dominant softener.

The 38V and 55V follow the same pattern. They are the same family, designed to the same target. Stop them down past f/8 only when DoF demands it. You are trading sharpness for depth, knowingly, not buying lens performance.

Sigma 300-600/4 Sport Sweet Spot on the SL3

A different story entirely. The Sigma 300-600/4 Sport is a complex zoom design with substantial aberrations to clean up by stopping down. Dustin Abbott's MTF analysis finds it "very sharp lens in the center and mid-frame areas, with a typical corner drop-off at wide apertures" that improves on stopping down. The Digital Picture's testing notes good corner performance at 300mm f/4 with steady improvement on closing the aperture.

The aberration profile shifts the operational sweet spot later in the f-stop range:

• f/4 to f/5.6: good central sharpness, but corners and mid-frame are still aberration-softened. Aberration-limited.
• f/5.6 to f/8: the sweet spot. Abbott identifies f/8 as peak performance for centre and mid-frame at 600mm. Aberration improvement continues into this range, masking the slowly-growing Airy disk.
• f/11 and beyond: the lens has finished cleaning up. Diffraction takes over as the dominant softener.

For a wildlife or distant-subject shooter on the SL3, this is still the lens to push deeper into the f-stop range than any XCD lens. The aberration headroom you spend by stopping down buys real sharpness gains, more than offsetting the slow Airy disk growth, until you cross f/11.

Leica APO Summicron 43 Sweet Spot (Q3 43)

The Q3 43's fixed lens is a 43mm f/2 APO-Summicron ASPH. The "APO" designation is the load-bearing word: the lens is corrected to produce minimal chromatic and spherical aberration, and reviewers consistently describe it as sharp across the frame from wide open. Tahusa Camera Reviews calls it "properly sharp, even wide open, across most of the frame." Amateur Photographer's review highlights the seven aspherical surfaces specifically for "cross-frame sharpness."

The crossover pattern matches the XCD V-series:

• Wide open through f/5.6: the sweet spot. Aberrations are negligible from the start.
• f/8: still very sharp, but diffraction is on the move.
• f/11 and beyond: stop here only when DoF demands it.

Three Sensors at 3.77µm: What Differs Is the Lens

All three of these cameras happen to sit at roughly 3.77 to 3.78µm pixel pitch (X2D II 100MP at 44×33, SL3 60MP at full frame, Q3 43 60MP at full frame). At this resolution, the sensor is sensitive enough to resolve diffraction penalties on any of these lenses past about f/8 with green light. Across the three lenses, the sensor resolves diffraction equally. The variable is how much aberration headroom each lens has to spend on the way to the diffraction floor. The XCD V and APO Summicron have almost none. The Sigma 300-600 has a lot.

No DoF app I have used models this per-lens reality honestly. The "diffraction warning" feature in PhotoPills and DoFMaster treats diffraction as a single sensor-driven threshold, typically f/11 or f/13. For an XCD V lens on a 100MP X2D II, that warning fires too late. For the Sigma 300-600 on the SL3, it fires too early. The honest answer is that diffraction onset is a lens-driven point on the f-stop dial, and it has to be tested per combination.

What this means at the camera

Forget any single "diffraction starts at f/X" rule. The crossover from aberration-limited to diffraction-limited depends on the lens and is revealed (or hidden) by the sensor. For XCD V-series glass on the X2D II, the sweet spot is wide open through f/5.6. For the Sigma 300-600 on the SL3, push to f/8 and accept f/11 if depth demands it. For the Q3 43, treat it like the XCD V-series. Test your specific combinations, write the answer down, and stop trusting the apps.

tl;dr for working shooters

1. Circle of confusion is your tolerance for blur, not a property of the lens. Pick traditional or modern based on your output.
2. To match depth of field across sensor sizes, multiply your f-number by the crop factor. To match exposure, do not.
3. Larger sensors gather more total light. That is why medium format looks cleaner in shadows even when stopped down for equivalent DoF.
4. Diffraction is physics. Aberrations are engineering. Better lenses meet the diffraction limit sooner.
5. Your sweet spot depends on the lens, not the body. Test your specific lenses and remember the result.