Why depth cues exist: the image is flat
This wants examples and reasons for depth cues. Light lands on the retina as a flat, 2-D image, but the world is 3-D, so depth is not received — the brain works it out from cues. A depth cue is visual information used to judge distance. Five of six are monocular (superimposition, relative size, linear perspective, texture gradient, height in the plane); only stereopsis is binocular. Cues are redundant, so a flat picture still looks deep.
The six cues: feature plus depth inference
Every cue definition has two halves — the feature in the image and the depth inference; a describe loses a mark for only the feature. The six: superimposition, relative size, linear perspective, texture gradient and height in the plane (all monocular), plus stereopsis, the only binocular cue, from each eye's different image. Two cues can fit one object (a distant tree looks smaller and higher up), so name the one asked.
Three illusion types (bullets g-i)
Visual illusion — perception does not match physical reality; the brain is tricked ("seeing things that aren't there" is too vague). Three types: fictions — an edge seen where none is drawn (Kanizsa triangle); ambiguous figures — one image, two readings, so perception flips (Necker cube); distortions — size or length misjudged (Muller-Lyer). An illusion is the normal process on misleading input, not a fault of the eye.
Visual constancies: shape, colour, size
Visual constancy — objects keep the same properties despite a changing retinal image: shape (a door stays rectangular), colour (an apple stays red in lamplight) and size (a receding friend looks unchanged). The brain steadies it from what it knows. Size constancy bridges to distortions — a misleading distance cue throws the size judgement. Constancy is NOT a depth cue: a cue answers how far away, a constancy what it really is.
Drawn from real examiner reports.
Superimposition is covering, not lines
Linear perspective is the magnet answer here — three cues get answered as "converging lines", and superimposition is one. Superimposition is one object partly covering another, so the covered one is perceived as further away — nothing to do with lines meeting at a point. Self-check: if converge appears, you've described a different cue.
Superimposition confused with linear perspective (June 2023 P2 Q22).
Texture gradient is grain, not lines
Texture gradient is the grain of a surface getting finer, more densely packed with distance, so finer = further away. It is answered as linear perspective, but a texture is a surface pattern, not a line, and never meets at a point. Write "grain gets finer with distance", not "lines converge to a point".
Texture gradient confused with linear perspective (June 2022 P2 Q22).
Height in the plane is position, not size
Height in the plane — objects further away appear higher up in the visual field, nearer the horizon. The trap: distant things also look smaller, but size is a different cue (relative size), scoring nothing here. For height in the plane, use higher up/nearer the horizon, never smaller.
Candidates wrongly discussed size when height in the plane was asked (June 2023 P2 Q23).
Relative size needs same actual size
Relative size — of two objects known to be the same actual size, the one giving the smaller image is perceived as further away. The condition is essential: a close mouse gives a bigger image than a distant elephant. Never "small objects are further"; only "of two equal-size objects, the smaller-looking is further".
Wrong answers claim small objects are always further away (June 2019 P2 Q23).
Visual constancy is not a depth cue
Visual constancy is not a depth cue. It answers a different question: a cue asks how far away, a constancy asks what is it really like — its real shape, colour or size despite a changing retinal image. Examiners record it answered as linear perspective. Define it as objects keeping shape, colour or size despite environmental change, not as a distance cue.
Visual constancy confused with linear perspective (June 2023 P2 Q21).
Stereopsis is not stereotype
Two errors. The word: stereopsis is not stereotype; "stereo" means two. The mechanism: it is the only binocular cue, so one-eye answers fail. Eyes a few cm apart give each a slightly different image; the brain fuses them and the size of the difference gives distance (bigger = nearer). Close one eye: distance still comes from monocular cues.
Stereopsis widely answered as stereotype (June 2022 P2 Q21).
A visual illusion, defined precisely
"Seeing things that aren't there" is too vague — there is a real stimulus; what is wrong is the interpretation. Naming Muller-Lyer is not a definition — say what happens first. An illusion is not a fault of the eye: the eye reports correctly and the brain applies its normal rules, so illusions are consistent for everyone.
"Seeing things that aren't there" is too vague and naming Muller-Lyer alone is not a definition (June 2024 P2 Q22).
Two unique points on a 2-mark describe
A 2-mark describe needs two unique points on a depth cue — the feature in the image, then the depth inference. Superimposition: one covers another, so covered = further. "In front, so in front" is one point twice, scored once.
AO3 needs AO2: apply the scene
AO3 cannot be awarded without AO2: no scenario link scores nothing. Application names the specific thing in the scene — road edges, sand grain — and its cue. Naming the character or repeating the stem isn't applying.
'Difference' needs a connective
"Describe the difference" needs a connective — two definitions side by side score one mark. Use whereas: "relative size is how big it appears, whereas height in the plane is how far up it sits". Prose only; the 7-9 band is missed mainly for weak AO3 and balance.
The image on the retina is flat. The world is not. Everything in this statement is about how a brain turns a two-dimensional image back into a three-dimensional world — and what happens when it gets it wrong.
Depth cue — a feature of the visual information that the brain uses to judge how far away something is.
Monocular depth cue — a depth cue that works with one eye only. Five of the six here are monocular.
Full notes, flashcards, Q&A and the topic quiz for every premium subject.
Premium plans are US$8.99/month or US$49.99/year — first month free.
Studying with a parent's blessing? Show them this.
