Skip to content
PodSubs
ShowsPeopleGuestsJobsNewsEvents
More
BlogPodCred
Menu
ShowsPeopleGuestsJobsNewsEventsBlogPodCredNetworksSign in
Join nowSign in
  • Shows
  • People
  • Jobs
  • News
  • More
    • Guests
    • Events
    • Blog
    • PodCred
    • Networks
    Join nowSign in

Talking Science

Talking Science artwork

Vision – How the Eye Builds a Picture

Aug 25, 202624 minAudio

The sharp, seamless, full-colour picture you seem to be looking at right now isn't a direct recording of the world - it's a construction, built by your brain from a stream of messy and genuinely incomplete signals. This episode follows light on its journey from the cornea to the visual cortex, covering rods and cones, the blind spot, three-channel colour vision, and why optical illusions count as scientific evidence rather than mere party tricks. You'll come away understanding how the eye focuses light onto the retina, why rods and cones divide labour between night vision and colour vision, what causes the blind spot and why you never notice it, how just three types of cone can produce a near-limitless range of perceived colour, why illusions reveal the assumptions the brain relies on to fill in gaps, and what's actually going wrong in short-sightedness, long-sightedness, astigmatism and colour blindness. Talking Science is sponsored by Kinsoft. Visit www.kinsoft.com.au Further reading: American Academy of Ophthalmology - patient education on how the eye works; National Eye Institute (part of the US National Institutes of Health) - vision basics and common eye conditions; Scientific American - features on visual perception and optical illusions.

Audio and video play straight from the show's own hosting. PodSubs stores and serves no media.

Transcript

From the show. Tap a time to play from there.

I want you to try a little experiment right now, wherever you happen to be listening to this. Oh, this is a good one. Yeah, so just close one eye, hold up a finger at arm's length, just like slightly off to the outer edge of your peripheral vision. Make sure you keep your open eye staring dead straight ahead. Right, keep staring straight, but slowly sweep that finger further out to the side. If you do this right, you'll hit a very specific coordinate where your finger simply vanishes. It's totally gone. It doesn't blur

out of focus, it doesn't fade, the visual data is just deleted, and then it pops right back into existence as you keep moving your hand. It genuinely feels like a glitch in the matrix the first time you consciously notice it, you know? It really does. But that blind spot... It isn't an anomaly and it's not an optical illusion. It's a permanent physical gap in the visual field of every single human being. It sits in exactly the same anatomical location for all of us. Which is wild. We just go through our entire lives with our brains completely papering over it.

Yeah, we barely even notice it. And honestly, that shatters this massive illusion most of us carry around. Reveals this fundamental truth that we're exploring in our deep dive today. Yeah. Which is that the crisp, seamless, ridiculously high definition picture of the world you think you're looking at right now. It's not a direct recording of reality. Exactly. It is a highly edited construction built entirely by your brain. You are basically experiencing a simulation rendered in real time. Because the brain is taking this

stream of messy, deeply flawed, laggy, and frankly upside down electrical signals, and it's smoothing them out into the reality you experience. And unpacking how that construction pipeline works tells you just as much about the architecture of human consciousness as it does about the mechanics of the eye. Absolutely. So today we are taking a stack of sources, spanning everything from optical physics to evolutionary biology, and we're tracing that exact pipeline. Right, from the start of the process. Right. We're going to follow a single photon from the very front

of the eyeball all the way to the back of the skull. Our mission for this deep dive is to understand the actual mechanics of how light gets translated into an image, why our visual cortex is forced to act as this aggressive film editor, and how we end up with this remarkably stable picture of the world. And to appreciate the editing software in the brain, we really have to look closely at the hardware gathering the raw data first. That collection process starts at the front of the eye, which honestly operates on the exact

same physical principles as a mechanical camera capturing light. Okay, let's unpack this. Because if you look at the physics of it, the moment light hits your eye, it immediately hits the cornea, right? Correct, yeah. That clear, curved dome on the outside. And that dome acts as the primary heavy lifter for your vision. Because of its dense curved structure, the cornea is doing the vast majority of the actual refracting. Like it's bending the incoming light waves so they start converging toward a single focal point. Yeah, but that initial refraction is pretty crude, even if it's powerful. Once the cornea bends

the light, it passes right through the pupil. Which is just a hole, essentially. Exactly. The pupil is just negative space. It's an opening controlled by the iris, the colored part of your eye, which operates dynamically to manage the photon flood. Like a camera aperture. Exactly like an aperture. In glaring sunlight, it constricts down to a time. tiny pinhole so the visual system doesn't overload. And in the dark, it dilates wide open to catch every possible photon it can. And that mechanical gatekeeping is super vital because the lens, which is sitting right behind

the pupil, has a very delicate job. It really does. The cornea gives you a fixed static focus, but the lens gives you dynamic focus. It's this flexible, transparent. sack of proteins. And its entire purpose is to fine -tune the trajectory of the light through a biological mechanism called accommodation. It is such a phenomenal piece of engineering. The lens is suspended by these microscopic threads that are attached to a ring of ciliary muscles. So it can actually change shape. Yes. When you shift your gaze to, say,

a distant mountain, those muscles pull outward. They stretch the lens flat to bend the light just a tiny bit. Oh, wow. But then when you glance down at a book in your hands, the muscles contract inward, taking the tension off those threads. The lens physically bulges into this rounder, thicker shape, which bends the incoming light much more sharply to focus on the close -up text. And it's just smoothly, continuously altering his physical geometry dozens of times a minute. Without you even thinking about it. Right. But

here's where I kind of want to challenge the source materials camera analogy, though. Okay, go for it. If the cornea and lens are the camera optics and the pupil is the aperture, that means the light finally lands on the retina, right? That sheet of tissue lining the back of the eyeball. Right. The classical analogy always calls the retina the film. But isn't calling the retina film selling it entirely short? I mean, photographic film is just dead chemicals waiting to be exposed. Yeah, it's totally passive. Exactly. Yeah. But the retina... is wildly complex brain tissue

that just kind of got pushed out into the eye during embryonic development. What's fascinating here is that vision researchers actually despise the film analogy for exactly that reason. The retina is absolutely not a passive canvas. It is doing heavy computational work before the signal even leaves the eye. Yeah. By the time the visual signal actually leaves the back of the eye, the retina has already begun sorting for edges, detecting motion, enhancing contrast. The genuinely biological computational part of vision begins the millisecond light hits that

back wall. So let's look at the microscopic workers pulling off that computation. The retina is carpeted with specialized cells called photoreceptors, and they have the absurd task of taking physical light and turning it into electricity. Which is no small feat. No. And based on their microscopic shapes, they're divided into rods and cones. And the division of labor between them is severe. It's a very strict split. Well, cones require a massive amount of light energy to trigger. They're practically useless in the dark. But in exchange for requiring all that light, cones

provide our ability to perceive rapid changes, fine textures, and really sharp details. And they're hoarded in one specific location, right? Yeah, mostly in the center. You have this tiny central pit in the retina called the fovea. which is absolutely packed with cones, which explains a highly specific modern phenomenon. Like when you're scaring at a tiny high -resolution font on your phone, you have to look directly at the words to read them. Right, because you need the fovea. Exactly. When you lock your gaze on the text, you are mechanically aligning your eye,

so the light from those specific pixels lands dead center on the fovea. And if something happens outside of that? Well, say a bug flies by in your peripheral vision. it lands outside the fovea. You detect the movement instantly but you cannot resolve the detail. You literally don't have the biological hardware in your periphery to know if it's a housefly or a wasp until you rotate your eyeball and hit it with the fovea. Because the outer edges of your retina are dominated by the rods, not the cones. Rods are the evolutionary answer to surviving the night. They're the night

shift. Exactly. They are incredibly sensitive to low light levels. I mean, so sensitive that a single rod can theoretically detect a single photon. That is insane. They are the absolute workhorses of our night vision and peripheral motion detection. But the massive trade -off is their low resolution and their total inability to distinguish color. Okay, I want to apply this to a wildly frustrating human experience. Okay, let's hear it. You're spending a Saturday at the beach, right? Blazing intense sunlight. And

then you walk off the boardwalk and step into a deeply shadowed, windowless dive bar. Oh yeah, the classic dive bar blindness. Yes. For the first five minutes, you are functionally blind. You're stumbling over stools, apologizing to the air. Why does it take so excruciatingly long for your eyes to simply adapt to the dark? It's essentially a chemical bottleneck. So photoreceptors work using specialized light -sensitive pigments. When a photon strikes the pigment molecule, the molecule physically snaps into a different shape. And that shape changes what fires the electrical

signal to the brain. But once it snaps, that pigment molecule is what we call bleached. It is spent. The cell has to use metabolic energy to chemically rebuild it into its original shape. So in the glaring sunlight on the beach, you're... Sensitive rods are just completely overwhelmed. Their pigment is bleaching way faster than the body can possibly regenerate it. They basically clock out. They are entirely offline. You're navigating the bright beach using only your cones. Got it. So when you suddenly step into the dark

dive bar, there isn't enough ambient light to trigger the cones anymore. They shut down instantly. Now you urgently need your rods. But they're all bleached. Exactly. The specific pigment in the rods, which is called rhodopsin, is still totally bleached from the beach. Rebuilding rhodopsin is a slow, methodical chemical process. It takes roughly 20 to 30 minutes for your body to fully regenerate the rhodopsin supply. You are literally waiting for a chemical manufacturing plant in your eye to catch up to the lighting conditions.

Yeah, that's exactly what's happening. Now, the sources mentioned a brilliant hack for this, historically used by astronomers and sailors navigating by the stars. They use dim red lights in their cabins or on their instruments instead of normal white flags. Right, the red light trick. Because the red wavelengths of light are energetic enough to let the cones see the map, but they are too weak to bleach the rhodopsin in the rods. It basically allows you to read a compass without nuking the night vision you just spent 30 minutes chemically building. It perfectly isolates the two systems. You use the red sensitive cones

for the immediate task, and you keep the rods safely in reserve for looking into the dark horizon. Which perfectly transitions us to the cone's most famous job. The fovea is giving us that high -resolution central vision, but cones have to extract color from a universe that is just, you know, bouncing different wavelengths of electromagnetic radiation at us. Yeah, color is entirely constructed. And here's where it gets really interesting. Because you might intuitively assume that to see millions of distinct shades like magenta, teal, burnt orange, we would need millions of

specialized receptors, one for every conceivable color. which would require an eyeball the size of a beach ball just to house all those receptors, evolution found a vastly more efficient workaround, which is trichromatic vision. Tri, meaning three. We only have three variants of cones. That's it, just three. One is most sensitive to long wavelengths, which we perceive as red. One peaks at medium wavelengths, which is green. And one handles the short wavelengths, blue. But it is crucial to understand that these sensitivities overlap heavily. They don't just register one

rigid color. Right. They operate exactly like the pixels on a television screen. If you take a magnifying glass to a TV, there is no yellow light. There is no purple light. There are only red, green, and blue LEDs. Yep. And by dialing the intensity of those three lights up and down in varying proportions, the screen tricks your visual system into perceiving any color in the spectrum. And your brain is running the exact same algorithm in reverse. It calculates the color of an object by triangulating the relative signal strength coming from the three overlapping

cone types. So if the red cones are firing at 80%, the green at 50%, and the blue at zero, the brain computes orange. This triad system is elegant, for sure, but because it relies entirely on comparative ratios, it is highly fragile. Makes sense. If just one of those three cone types has a genetic mutation that shifts its sensitivity curve, the triangulation fails, the brain receives corrupted ratios, and this is the mechanical reality of colorblindness. You know, I always assumed colorblindness meant seeing the world through, like, a vintage black and

white movie filter, but... Based on this ratio system, it's more like a glitch in the color mixing software. Exactly. Total grayscale vision, which is called achromatopsia, is incredibly rare. Oh, really? Yeah. Most color blindness is simply a mixing error. If the red sensitive cone is altered, it's called protonopia. If the green is altered, deuteranopia. And what about blue? That's tryphonopia, though it's much less common. The red and green sensitivity curves naturally overlap the most. So when one of those

shifts, the brain completely loses the ability to distinguish between red and green hues. They just merge into a muddy brownish yellow. Now, the sources flag that this specific red -green mixing error is vastly more common in men. And this comes down to the genetic lottery. Yeah, it's all in the chromosomes. The blueprints for the red and green cones are located on the X chromosome. Because women typically have two X chromosomes, they have biological redundancy. If one carries a mutated code, the second X chromosome usually provides a healthy backup blueprint.

While men only have one X chromosome paired with a Y, there is no backup copy. So if that single X chromosome carries the shifted cone gene, The hardware gets built with the flaw. It's a fascinating vulnerability and is also exactly what clinicians exploit with the Ishihara test. Ah, the dot test. Yeah, those are the circles filled with hundreds of randomized colored dots with a number hidden inside. Everyone's seen those at the eye doctor. To a typical trichromatic system, the ratios compute and a neon green 12, for example, just

pops out against the red background. But if your cones are feeding the brain corrupted overlapping signals, the color contrast disappears entirely. The dots just look like random camouflage noise. It completely isolates the exact failure point in the hardware. But, you know, there is another massive computational hurdle the visual system has to clear. We have sharpness, we have color, but the retina is still a flat two -dimensional sheet of tissue. Right. And the universe we live in is three -dimensional. The incoming data completely lacks a z -axis. Exactly. So how does the brain

mathematically reconstruct depth from two flat sheets? That is the million -dollar question. It relies on having two cameras, stereopsis or binocular disparity. Because your eyes are spaced a few centimeters apart on your skull, they each capture a slightly different angle of the exact same scene. Which goes back to the jumping finger experiment from the intro. If you hold your finger close to your nose and rapidly blink alternating eyes, the finger violently jumps back and forth against the background. Because the brain is

actively overlaying those two mismatched images. By measuring the literal physical disparity, the exact pixel distance the object shifted between the left eye's feed and the right eye's feed. The visual cortex can instantly triangulate how far away the object is. So the closer the object... the more extreme the disparity. Spot on. Wait, I have to challenge this, though. Because the physics that triangulation mean the angles get incredibly narrow the further away an object is, the sources explicitly state that this two -eyed comparison stops being mathematically useful

after about 100 feet. Right, the angles just get too tight. The discrepancy between the two eyes is just too small to measure at that point. So if that is true, how do we successfully merge onto a highway at 70 miles an hour or, you know, land an airplane without totally misjudging the distance of the horizon? You are absolutely right to catch that. Stereopsis is really only for the immediate foreground, you know, threading a needle, catching a baseball, pouring your coffee. For everything else, the brain abandons the two -eye triangulation completely and relies heavily

on monocular cues. These are context clues that require only a single eye to decipher. Oh, like motion parallax. So if you look out the window of a speeding bullet train, the gravel right next to the tracks is an unrecognizable blur sweeping past you. But the mountains five miles in the distance barely seem to move at all. Exactly. The visual cortex analyzes that relative velocity. Fast movement equals close. Slow movement equals distant. It also calculates occlusion like if a tree blocks your view of a building, the tree must be closer. Makes total sense. It calculates

converging lines too, like a long highway narrowing to a point in the distance. Your brain builds a 3D model using logic and context rather than relying on raw binocular triangulation. Which brings us to the ultimate proof that the brain is doing most of the heavy lifting. All of this data, the color ratios, the depth calculations, the motion detection, it all has to actually get from the retina into the brain. And it faces a severe anatomical bottleneck to get there. Right. Every single one of those millions of

photoreceptors has an axon. Basically a biological wire. All of those wires have to route across the surface of the retina, bundle together into a massive biological cable called the optic nerve, and punch a hole straight through the back of the eyeball to reach the brain. Like thousands of Ethernet cables being forced through a single conduit pipe in a wall. Precisely. And at the exact coordinate where that optic nerve punches through, there is physically no room for any photoreceptor. It is a dead zone. Right. Any light that falls on that spot is completely unregistered.

It creates a literal black hole in your peripheral vision. This is the finger disappearing trick we started with. When the finger hits that specific angle, the light bouncing off it is landing squarely on the optic nerve bundle. There's nothing there to receive the signal. And if we connect this to the bigger picture, this anatomical flaw forces the brain to reveal its true nature. Because we don't walk around seeing two floating black orbs in our vision. Thankfully. Right. The visual cortex, this massive chunk of tissue at the back of the skull, is acting as a real -time rendering

engine. It analyzes the colors and patterns surrounding the blind spot, accesses your memory of what should be there, and literally paints over the hole with a fabricated texture. It is essentially a hallucination constrained by context. And it has to heavily rely on that hallucination software, because the raw video feed coming from the eyes is frankly nauseating. Oh, it's completely unwatchable. The sources detail a phenomenon called saccades. Our eyes are practically never still. To get that high resolution for via pointed at everything

in our environment, the eyeball violently jerks around in rapid ballistic movements several times a second. If you strapped a GoPro to a gimbal and violently whipped it around three times a second, the resulting footage would be an unwatchable motion blurred mess. But we don't experience reality as a motion blurred mess. To prevent us from getting motion sickness from our own eyes, the brain utilizes saccadic masking. And this is where it acts like an incredibly aggressive film editor. Exactly. During the millisecond, the eye is actually moving. The visual cortex

simply cuts the feed. It deletes the blurred transition frames entirely. It then seamlessly stitches the sharp, static shots together, slightly shifting your perception of time to hide the cuts. It's amazing. The audience, which is your conscious mind, only ever experiences a perfectly smooth, stabilized cinematic pan. And the processing power required to edit reality and manipulate time perception just to hide our own eye movements is staggering. And because the brain relies so heavily on these editing shortcuts, painting

over blind spots, masking motion, inferring depth, we can actually intentionally hack the software to prove it exists. This is where optical illusions come in. They aren't just parlor tricks. No, to a neuroscientist, an optical illusion is a targeted exploit of the brain's rendering engine. Like take the classic checkerboard illusion. You have a gray square in bright light and a gray square in the shadow of a cylinder. I know this one. Your brain insists the square in the shadow is a dramatically lighter shade of gray. But if you isolate them, they are the exact same

hex code of gray. The illusion works because your brain fundamentally does not care about the absolute number of photons hitting the retina. It cares about relative contrast. Exactly. In evolutionary terms, measuring absolute light is totally useless. Recognizing the dark silhouette of a predator against a slightly less dark grouping of trees is a matter of survival. Right. That's what keeps you alive. So the brain automatically compensates for the shadow, digitally brightening the square in your perception. The illusion just

catches the software overcorrecting. So what does this all mean for those of us who have to wear glasses or contacts? If my vision is a blurry mess when I wake up, is my visual cortex failing at its editing job? Not at all. That is a hardware geometry issue. It's entirely separate from the software. The brain is editing the signal perfectly. It's just being handed a corrupted file to start with. Oh, I see. It usually comes down to the physical dimensions of the eyeball throwing off the cornea's focal point. Let's look at the mechanics of that. If you have myopia, which is nearsightedness,

your eyeball grew just a millimeter or two too long from front to back, right? The cornea and lens bend the light. But the light rays converge and cross before they actually hit the retina. So the image has already started to spread back out into a blur by the time it hits the receptors. And conversely, hyperopia farsightedness means the eyeball is slightly too short or the cornea is too flat. The light rays physically hit the retina before they've had a chance to fully converge into a sharp point. And then there's astigmatism.

Instead of the eyeball being too long or short, the cornea itself is warped. I always think of looking at your reflection in a funhouse mirror that only stretches you vertically but leaves your horizontal proportions alone. That's a great way to picture it. The cornea bends the light sharply on one axis but weakly on another, creating conflicting focal points that just smear the image at any distance. But the genius of corrective lenses is how elegantly simple they are. They are just external pieces of math. A glass lens is custom ground to prematurely bend the incoming

light in the exact opposite direction of your eye's anatomical flaw. So it pre -bends it. Exactly. It corrects the math before the light even touches your cornea, ensuring the final focal point lands dead center on the retina. The moment the hardware gets a clean signal, the software in the brain processes it flawlessly. It is truly staggering when you take a step back and look at the sheer biological engineering happening for you every single second. You have a microscopic chemical plant regenerating rhodopsin so you can navigate the dark. You have a living protein -based lens

dynamically altering its geometry dozens of times a minute. You have a three -channel color matrix reconstructing a limitless spectrum from overlapping ratios. And doing the heaviest lifting of all, you have a visual cortex quietly painting over blind spots, analyzing relative velocity, and editing out shaky cam footage just to hand your conscious mind a stable reality. It is a symphony of physics, chemistry, and computation. And it's executed continuously from the moment you open your eyes until you fall asleep, requiring zero

conscious effort on your part. Which brings us right back to the finger we watched disappear at the start of this deep dive. It is a permanent physical reminder that a patch of your vision right now, wherever you are looking, is a total fabrication. An educated guess rendered by a lump of gray matter in the dark. And this raises an important question, one that really builds on everything we've uncovered in the architecture of the brain today. If the visual cortex relies so heavily on memory, evolutionary context, and expectation to fill in the gaps and literally

paint our reality, what happens when we encounter something utterly unprecedented? Like something we've never seen before. Yeah. If we cross paths with an object or a phenomenon that defies human physics and our brain has absolutely no context or memory to draw from, do we even see it for what it truly is? Or does the rendering engine just crash, edit it out entirely, and cover it up with a best guess just to keep the simulation making sense?

Open the transcript file

Hosts and Guests

Nobody is credited on this episode yet.

Credit the people on this episode

Be the first to rate this episode

More episodes

All episodes
  1. Nanotechnology – Building at the Scale of Atoms

    Nov 19, 202626 minAudio

    Nanotechnology works at a scale where a nanometre is to a metre what a single centimetre is to the distance across a football field roughly a hundred thousand times over. This episode explains what nanotechnology...

    Credit the people on this episode

  2. Animal Migration – Navigation Without Maps

    Nov 17, 202625 minAudio

    A tagged bar-tailed godwit once flew more than thirteen thousand kilometres from Alaska to Tasmania without landing once. This episode explores why animals migrate at all, from godwits and humpback whales to monarch...

    Credit the people on this episode

  3. Space Weather – When the Sun Attacks Our Technology

    Nov 15, 202625 minAudio

    The Sun is an active, magnetic star, and every so often it hurls bursts of radiation and clouds of charged particles toward Earth, a phenomenon called space weather. This episode explains solar flares and coronal mass...

    Credit the people on this episode

  4. Tsunamis – Waves That Cross Oceans

    Nov 12, 202623 minAudio

    An ordinary wave rolls onto the sand every few seconds; a tsunami can take tens of minutes between crests and cross an entire ocean at the speed of a jet aircraft while barely rippling the surface. This episode explains...

    Credit the people on this episode

  5. Symmetry – Physics' Deepest Idea

    Nov 10, 202620 minAudio

    A folded square of paper and a butterfly's wings share the same simple visual idea: a shape that looks the same after some transformation. This episode follows that idea from everyday geometry into the deepest...

    Credit the people on this episode