Coming soon
Video walkthrough
We’re producing a short video for every module — check back soon.
Exam category: Image Production → Image Acquisition & Evaluation
Why it matters: This module is the "put it all together" module — contrast, resolution, distortion, and noise are the four things you're actually judging when you look at a finished image, and the registry tests each one as both a definition and a cause-and-effect relationship.
A good-quality image accurately represents the anatomic area of interest, and it is judged on five characteristics: brightness, contrast, spatial resolution, distortion, and noise.
Notice what is not on that list. kVp, mAs, SID and filtration are exposure factors — they influence quality but are not quality characteristics. Pixel pitch, fill factor and bit depth are detector parameters. Anode angle, focal spot and grid ratio are equipment and geometry factors. The five above describe the finished image itself; everything else describes what produced it.
Equipment QC exists to make sure the machine isn't quietly degrading any of the five. Two more checks round out what actually gets graded on a finished image: whether an exposure error is so severe it's unsalvageable (saturation), and whether the required identification markers are actually present.
Subject contrast comes from the anatomy itself — how differently adjacent tissues attenuate the beam. The chest (air, bone, soft tissue) has naturally high subject contrast; the abdomen (mostly similar soft tissues) has naturally low subject contrast.
Film images are described by scale of contrast: few density differences but large jumps between them = high contrast / short scale; many density differences with small jumps = low contrast / long scale. Raising kVp increases beam penetration, reduces differential absorption between tissues, and lowers contrast (longer scale).
Differential absorption is the whole mechanism underneath subject contrast: some of the beam is absorbed by the tissue and some transmits through it, and different tissues do not absorb to the same degree — bone absorbs far more photons than air-filled tissue. Those differences are what structurally represent the anatomy. Anything that makes tissues absorb more alike flattens subject contrast.
Added filtration is aluminum placed in the beam at the tube port. The beam leaving the tube is polyenergetic, and its low-energy photons cannot penetrate the anatomy at all — they contribute nothing to image formation and only add patient dose. Filtration's job is to remove them.
Doing that has a predictable side effect on the image. Increasing beam filtration decreases the quantity of the beam and increases its quality — the average photon energy goes up. A higher-energy beam is absorbed more uniformly by different tissues, which is exactly the differential-absorption argument above: less difference in absorption means less subject contrast.
The one-line version: added filtration strips out the low-energy photons that were only dosing the patient, and the harder beam it leaves behind reduces subject contrast. It is the same trade you make when you raise kVp.
Spatial resolution = the smallest object the system can still show as distinct — essentially, sharpness. Some unsharpness is unavoidable in any imaging system; the goal is minimizing it.
Motion is the single biggest enemy of spatial resolution. Any movement during the exposure smears the edges of detail — motion unsharpness — and no amount of processing recovers it. There are two kinds:
Memory cue — time is the fix for motion: you can talk a patient out of voluntary motion, but the only lever for involuntary motion (peristalsis, heartbeat) is a short exposure time — raise the mA to keep the mAs.
Modulation transfer function (MTF) measures how well a system preserves contrast as object size shrinks (as spatial frequency increases), scored 0 to 1.0. Large objects are easy (MTF close to 1); small, fine detail is hard, and most systems' MTF drops well below 1.0 there. A higher MTF at a given spatial frequency = better detail visibility at that size.
We draw the x-ray source as a point, but it isn't one — x-rays come from an area on the anode called the focal spot. Because the source has width, every edge in the image is drawn with a slightly fuzzy border rather than a clean line. That fuzzy border is penumbra: the unsharp edges of objects in the projected image. Some penumbra is always present.
Three geometric levers control how much:
Use of a small focal spot, an increase in SID, and a decrease in OID result in less geometric unsharpness and increased resolution.
That single sentence answers most geometry questions on this topic, in either direction — a larger focal spot, a decreased SID, or an increased OID each make unsharpness worse. Most tubes have dual focus (a large and a small focal spot), so switching to the small focal spot for a fine-detail exam like a wrist is the everyday application: a smaller source area means less penumbra and sharper edges.
The trade-off is heat. The small focal spot concentrates the same energy into a smaller area, so it has lower heat-loading capacity — high-mAs techniques need the large focal spot. Note what switching focal spots does not do: it doesn't change radiation output per mAs, doesn't change patient dose, and doesn't affect contrast or brightness. It buys sharpness and nothing else.
Penumbra can be calculated. The same quantity is called geometric unsharpness (Ug) or focal-spot blur, and the geometry is a simple similar-triangles ratio:
Ug = (focal spot size × OID) ÷ SOD
where SOD is the source-to-object distance — not the SID. (SOD = SID − OID.) Keep OID and SOD in the same units so their ratio cancels out; the answer then carries the units of the focal spot, i.e. millimeters.
Worked example: a 1.0 mm focal spot, OID 18 cm, SOD 120 cm →
Ug = (1.0 × 18) ÷ 120 = 18 ÷ 120 = 0.15 mm.
Three ways this calculation goes wrong, all of which show up as answer options: dividing OID by SOD and forgetting to multiply by the focal-spot size; using the SID in the denominator when the question gave you an SOD; and inverting the fraction so SOD ends up on top. Write the formula down before you touch the numbers.
X-rays leave the anode in all directions, but those emitted toward the anode side have farther to travel through the anode material itself, and some are absorbed by it — the "heel". The result is a beam that is not uniform along the tube's long axis:
X-rays are more intense on the cathode side of the tube; their intensity decreases toward the anode side. The difference between the two ends can be as much as 45%.
That is a defect you can exploit. Put the thicker end of the body part under the cathode, where the beam is most intense, and the thinner end under the anode — the result is a more even exposure across the receptor. Imaging a femur, the thick proximal end goes toward the cathode. Imaging a thoracic spine, the patient's head goes under the anode end, so the more intense radiation reaches the larger lower vertebrae.
Don't confuse it with its neighbors. The line-focus principle is a different anode concept — a steeply angled target makes the effective focal spot smaller than the actual one, which improves resolution. Off-focus (extrafocal) radiation comes from outside the focal spot and degrades resolution. Focal spot blooming is the effective focal spot growing at high mA. Only the heel effect describes intensity varying from cathode end to anode end.
Distortion = the image misrepresenting the true size or shape of the anatomy. It comes in two types:
Memory cue — TIRE: Tube, IR, and part alignment errors cause Elongation.
Quantum noise (called quantum mottle on film) is visible brightness/density fluctuation caused by too few photons reaching the receptor. Fewer photons = more visible noise — this is why a digital system can "fix" the brightness of an underexposed image but not the noise baked into it.
Signal-to-noise ratio (SNR) compares the strength of the actual exposure signal to the amount of noise present. Higher SNR = cleaner, higher-quality image; low photon counts (quantum noise) directly drag SNR down.
Detective quantum efficiency (DQE) measures how efficiently a receptor converts the x-ray exposure it receives into a usable image, scored 0 to 1.0 (no system reaches 100%). A higher-DQE receptor needs less radiation to produce an equivalent-quality image — DR generally has higher DQE than CR.
Routine QC checks that the equipment itself isn't degrading any of the above, independent of technique:
A diagnostic-quality image can still be read badly, and the features that decide it split into two groups.
The monitor's own properties. Alongside luminance, a primary display is judged on its matrix size and pixel pitch — the same two quantities that set spatial resolution on the acquisition side (see Digital imaging), now applied to the display. A monitor cannot show detail finer than its own pixels, so an image acquired at high resolution can still be displayed at low resolution. Contrast resolution depends on the monitor being able to render the grayscale range the image actually contains: a system producing 14-bit data needs a display capable of a wide grayscale range, or the extra shades are simply not shown.
The room the monitor is in. Viewing conditions are a graded feature of the display, not an afterthought — principally ambient lighting and monitor placement. Ambient luminance is the room light reflecting off the display surface, measured with the monitor switched off, and it is added into the monitor's other luminance measurements. That's the mechanism worth holding: bright room light doesn't just make the screen harder to look at, it raises the floor of the display's black level, compressing the darkest shades. Keeping ambient lighting low improves soft-copy viewing for exactly that reason.
Memory cue — the room gets added to the screen. Ambient luminance is measured with the monitor off and then adds into its other readings. Turn the reading-room lights up and you have raised the monitor's black floor, losing dark detail that the receptor captured perfectly well.
An artifact is anything on the image that isn't the patient. Two are worth knowing by their appearance, because both are routinely misread as an exposure problem.
Off-center grid cutoff (lateral decentering). A focused grid's lead strips are angled to match the divergence of the beam. If the central ray is not aligned side-to-side with the center of the grid, that match is lost across the whole grid at once, and the strips absorb primary radiation everywhere. The image shows an overall loss of exposure across the entire receptor — which looks exactly like too little mAs, but no amount of extra technique fixes it.
Its three siblings fail differently and are the usual distractors: an off-level grid is tilted relative to the beam; off-focus means the SID used is outside the grid's recommended focal range; and an upside-down focused grid cuts off severely at the periphery while the center is comparatively spared.
CR ghost image. A faint image of previously imaged anatomy superimposed on the current image means the imaging plate was not completely erased before reuse. The residual latent image from the prior exposure is still on the plate. It is usually correctable by running the plate through additional erasure — which is why "erasure thoroughness" is a standing QC test for CR.
Both are pattern-recognition questions. An overall dark-or-light shift with no anatomy problem points at the grid; a second, unrelated anatomy faintly present points at the plate. Neither is quantum mottle (that's random speckle from too few photons) and neither is a windowing error (that would change the whole displayed image, not add a ghost).
The exposure index (EI) reports the amount of radiation exposure received by the image detector — not the kVp used, not the patient's entrance skin exposure, and not the signal-to-noise ratio of the displayed image. It is a receptor-side number, which is exactly why it can flag a technique problem the displayed brightness hides. (The Digital imaging module covers how EI values are derived and how deviation is reported.) That handles routine over/underexposure; there's a more extreme failure mode graded separately: saturation. Every digital receptor has a finite dynamic range — a maximum signal it can register before its response curve flattens out. Push the exposure far enough past that ceiling and the receptor simply can't record any more differences: every pixel in the saturated region reads the same maximum value.
The visible result is a region that's uniformly bright with no contrast or detail — and unlike routine over/underexposure, this isn't something brightness/contrast post-processing can fix. The receptor ran out of room to record the data needed to show that detail, so it was never captured in the first place.
Memory cue — EI catches "off-target," saturation catches "off the chart": a bad exposure indicator value means the technique needs adjusting. Saturation means the exposure was so far past the receptor's ceiling that no processing can bring the anatomy back.
An anatomic side marker (R/L) and patient/exam identification are both evaluated, required parts of every finished image — not paperwork attached afterward.
Memory cue — markers are exposed, not typed: if the beam didn't put it there, it doesn't count as a marker.
Students conflate "the computer can fix an underexposed image's brightness" with "the computer can fix its quality." It can't — the noise from too few photons is baked into the raw data before any processing happens. A second trap: assuming higher kVp always means a "better" image — it increases penetration but lowers contrast, which is not automatically desirable depending on the exam. A third: assuming a digitally-typed "R"/"L" label satisfies the marker requirement — it doesn't; only a marker physically exposed by the beam, before processing, counts.
Source: The display-monitor block in section 5 (viewing conditions, matrix size and pixel pitch as monitor features, contrast resolution needing a display that can render the image's grayscale range, and ambient luminance as room light reflecting off the display surface measured with the monitor off and added into its other luminance measurements, with low ambient lighting improving soft-copy viewing) is Fauber, Radiographic Imaging and Exposure, 6th ed., Ch. 5, as are the display-monitor luminance figures (minimum 1.0 cd/m², maximum 350 cd/m², luminance ratio recommended above 250, and an LR of 350 giving improved image contrast). Resolved 2026-08-02: this module previously taught a single average-luminance range cited to Fauber 5th ed.; that edition is superseded and the 6th specifies a minimum/maximum pair plus a luminance ratio instead, so the figures above were updated to the 6th and the superseded range removed. The module's other Fauber citations were re-verified against the 6th ed. on the same pass: Ch. 3 for image formation and radiographic quality (including dynamic range) and Ch. 4 for the digital receptor material; Ch. 2 for the anode heel effect — intensity greatest on the cathode side and falling toward the anode, by as much as a 45% difference, confirmed present in the 6th ed. — and for beam filtration and compensating filters; Ch. 7 for off-center grid cutoff; the five image-quality characteristics are Fauber's own list ("brightness, contrast, spatial resolution, distortion, and noise"). Focal spot, penumbra and the geometric-unsharpness factors per Bontrager, Ch. 1 ("use of a small focal spot, an increase in SID, and a decrease in OID result in less geometric unsharpness and increased resolution"); the Ug/focal-spot-blur formula and its worked example per Bushong, Radiologic Science for Technologists, 12th ed., Ch. 21; the CR incomplete-erasure ghost artifact per Bushong, Ch. 23. Motion unsharpness (voluntary vs. involuntary; short exposure time as the control) per McQuillen Martensen, Radiographic Image Analysis, 6th ed., Ch. 1, and Bontrager, Textbook of Radiographic Positioning and Related Anatomy, 10th ed., Ch. 6. Saturation and dynamic range per Bontrager, Ch. 1 (Fig. 1.156); marker placement/legal requirements per Bontrager, Ch. 1. ARRT Content Specifications, Image Production §F (Criteria for Image Evaluation).
Ready for the module check?
5 questions, freshly drawn each attempt. 85%+ to pass. Retries are unlimited and immediate.