You've Never Seen a Real Photo

[00:00] This video is sponsored by Incogn. I

[00:02] made this rubber band that changes color

[00:04] when you stretch it. It's cool, isn't

[00:05] it? It starts off blue, then goes green,

[00:08] and then red, and then when I let go, it

[00:11] returns back to blue. And it's the same

[00:13] mechanism that's used in this

[00:15] 135year-old photography technique that's

[00:18] halfway between a hologram and a

[00:20] photograph. It's the same chemistry as

[00:23] black and white photography. But look,

[00:25] when I change the viewing angle, it

[00:27] bursts into color. And this isn't just

[00:29] any color image. It's arguably the

[00:32] truest color image you'll ever see. And

[00:34] that's because standard color

[00:36] photography is a lie. That's a bit

[00:38] inflammatory, isn't it? Helps with the

[00:40] views. All right. Here's why standard

[00:43] color photography is a scam. So, when

[00:45] you take a picture of something yellow,

[00:47] let's say, then you've got yellow light

[00:50] hitting the sensor of the camera. And

[00:53] that light has a wavelength of about 580

[00:56] nanome. But when you look at the

[00:58] resulting image, you're not getting that

[01:01] wavelength coming from the image.

[01:04] Instead, you're getting some red

[01:05] wavelengths and some green wavelengths

[01:08] that your brain interprets as yellow.

[01:11] And that's because we don't have

[01:12] dedicated cone cells in our eyes for

[01:16] yellow wavelengths of light. Whereas,

[01:18] when I look at this image, all the

[01:20] wavelengths of light that were hitting

[01:22] the plate when the picture was taken are

[01:24] now hitting my eyes as I look at it. And

[01:27] that's because it's basically a hologram

[01:29] except it's encoding color information

[01:32] instead of depth information. For the

[01:34] vast majority of people, you'd never be

[01:35] able to tell, but we can use this

[01:37] spectrum analyzer to see it. So look,

[01:40] this is my phone screen through a

[01:41] spectrum analyzer. And as I move around

[01:44] the image, what you're seeing is just

[01:46] three peaks changing in height. That's

[01:49] red, green, and blue. And they match up

[01:51] with the three types of cones in our

[01:53] eyes. But when I point it at the Litman

[01:55] plate, you can see this complex

[01:57] distribution of wavelengths. And because

[01:59] there are flowers in this image, we're

[02:01] even getting a bit of ultraviolet there

[02:03] as well. Actually, that's a good point.

[02:05] Standard photography doesn't work for

[02:06] other animals, like bees, for example,

[02:09] because they have a different

[02:11] distribution of cones in their eyes. And

[02:14] actually, even for a tiny minority of

[02:16] humans that have four different cone

[02:18] cells and the neurological machinery to

[02:21] use the extra one, the yellow in this

[02:23] Litman plate will look like the yellow

[02:25] they see in real life, whereas yellow in

[02:27] standard photographs would look wrong.

[02:29] So, how is it that this black and white

[02:31] film can reproduce the full spectrum of

[02:34] colors from the real world? Well,

[02:36] they're called Litman plates, and they

[02:38] were made for me by an amazing

[02:39] photographer called John Hilty. I really

[02:42] hope you'll check out his stuff. I'll

[02:44] leave a link to it in the description.

[02:46] And once I had them in my hands, I just

[02:48] thought, well, I have to show this to

[02:49] the world. So, here we are. We're going

[02:51] to figure out together why color seems

[02:54] to magically emerge from this plate.

[02:57] Here's a clue to start with. Watch what

[02:59] happens when I breathe on this plate.

[03:02] See how the whole rainbow seems to shift

[03:05] to the left? Why the heck does it do

[03:07] that? Well, surprisingly enough, it's

[03:09] for the same reason that chameleons

[03:11] change color. Watch this.

[03:16] Yeah, that didn't work. I don't know why

[03:18] I thought that would work. Anyway, we'll

[03:20] get back to chameleons later, I promise.

[03:22] But let's use this rubber band of mine

[03:25] first. We'll call it synthetic chameleon

[03:27] skin because it too changes color for

[03:30] the same reason that a Lipman plate

[03:32] changes color when it's breathed on. So,

[03:34] the way I made this rubber band is I

[03:36] pulled the DVD apart and poured some

[03:39] silicon on this exposed surface inside.

[03:42] In other words, I used that surface as a

[03:44] mold. The surface looks really flat, but

[03:47] it's actually covered in tightly packed

[03:48] grooves. So, now my rubber band also has

[03:51] these tiny grooves in it. And it's these

[03:54] tiny grooves that bring the colors out.

[03:57] So, again, it's not color pigment like

[04:00] the silicon doesn't have any color of

[04:02] its own. It's transparent. Instead, the

[04:04] color comes from the structure of the

[04:06] surface. That's why it's called

[04:08] structural color. And that means if you

[04:10] can change the structure by stretching

[04:13] it, for example, then you can change the

[04:15] color. How cool is that? How a Lithuan

[04:18] play ends up with structural color is

[04:20] amazing. But first, what is structural

[04:22] color? Well, my favorite example is

[04:25] bubbles. Why do you get rainbows in

[04:27] bubbles? Well, let's zoom right in on

[04:30] the bubble film and see what happens

[04:31] when light shines on it. Some light gets

[04:34] reflected and some light gets

[04:35] transmitted from that top surface. Then

[04:38] some gets reflected and some gets

[04:39] transmitted from that bottom surface.

[04:42] And then some gets reflected and some

[04:44] gets transmitted from the top surface.

[04:46] Again, suppose this is red laser light

[04:48] we're shining here. And this is the

[04:50] wavelength. Now, it just so happens that

[04:52] this distance here is exactly one full

[04:55] wavelength of the red light. So you can

[04:58] see that the peaks and the troughs line

[05:00] up perfectly. So when you add them

[05:02] together, you get an even bigger wave.

[05:04] That's called constructive interference.

[05:06] But what if we switch to a blue laser?

[05:08] Well, now this distance is one full

[05:11] wavelength and a half. So now the peaks

[05:14] of this wave line up with the troughs of

[05:17] this wave and vice versa. So when you

[05:19] add them together, they exactly cancel

[05:21] out and you get no light coming out.

[05:24] That's called destructive interference.

[05:26] Now, for in between wavelengths, you

[05:29] wouldn't get perfectly constructive

[05:31] interference or perfectly destructive

[05:33] interference. So, you'd get a brightness

[05:35] in between for intermediate wavelengths.

[05:39] Except what's actually happening here is

[05:41] the light is bouncing up and down inside

[05:44] this soap film multiple times. What does

[05:48] that mean for those in between

[05:49] wavelengths? Well, when you add up loads

[05:52] of offset waves like this, it ends up

[05:55] looking like complete destructive

[05:57] interference anyway. In other words,

[05:59] when you have lots of reflections, the

[06:01] interference becomes really selective

[06:03] and it gets tighter the more reflections

[06:06] there are. So, you end up seeing only

[06:08] that one perfect wavelength of red light

[06:11] that exactly matches and a little bit

[06:13] either side. And the reason you see

[06:15] rainbows is because the angle that the

[06:19] light enters the film at changes this

[06:22] length. So here red light is reflected

[06:25] and here blue light is reflected and you

[06:28] get the full spectrum between these two

[06:30] points. So that also explains why DVDs

[06:33] have rainbows in them except that

[06:35] instead of the light bouncing up and

[06:37] down inside a thin film of soapy water,

[06:40] you've got light that's reflecting off

[06:41] closely packed ridges. So actually the

[06:44] same thing is true for my rubber band.

[06:45] You see different colors at different

[06:47] angles. The difference is that I can

[06:49] shift the rainbow by stretching it. And

[06:51] that makes sense because when it's

[06:53] stretched those ridges become more

[06:55] spread out and so the wavelengths that

[06:57] constructively interfere become longer.

[07:00] And the same thing happens when a

[07:02] chameleon changes color. So a chameleon

[07:04] has these special cells called

[07:05] iridaphors that are packed full of these

[07:09] tiny regularly spaced crystals of

[07:12] guanine. When the crystals are tightly

[07:14] packed, they cause constructive

[07:15] interference for blue light and

[07:17] destructive interference for all the

[07:20] other wavelengths. But a chameleon is

[07:21] able to engorg these cells by changing

[07:24] the osmotic pressure. When the cells

[07:26] expand, the crystals move further apart

[07:30] from each other. This causes the

[07:32] crystals of guanine to spread out so

[07:34] that they constructively interfere

[07:37] longer wavelengths of light. That's not

[07:40] the whole story though. Like chameleons

[07:42] also have a pigment layer. So the final

[07:44] color that you see is a mixture of

[07:47] structural color that the chameleon can

[07:49] tune and this pigment color. Real

[07:52] chameleon skin differs from my synthetic

[07:54] chameleon skin in a couple of ways. The

[07:56] first is that with my synthetic

[07:58] chameleon skin, the interference is

[07:59] happening because of surface ridges.

[08:01] Whereas with real chameleon skin, the

[08:03] diffraction is happening through deeper

[08:05] and deeper layers into the skin. The

[08:07] other difference is that real chameleon

[08:09] skin isn't iridescent, which is actually

[08:11] really impressive. Like, if there's

[08:12] diffraction going on, why don't I see

[08:14] the color change when the viewing angle

[08:17] changes like it does with this mother of

[08:20] pearl or this peacock feather? It's

[08:23] actually a bit easier to see how moro

[08:25] butterflies do it. That incredible blue

[08:28] is also structural color. And just like

[08:31] the chameleon, it looks blue from all

[08:33] angles. To see why, you have to look

[08:36] under an electron microscope. This is a

[08:38] cutaway through the wing. Here's a

[08:40] diagram of it for simplicity. So, the

[08:43] interference comes from these ridges

[08:45] here. So, if light comes directly down,

[08:48] it interferes with this stack of

[08:50] branches, and they're spaced for blue

[08:53] light. Now, if all the branches from the

[08:55] neighboring trees lined up like this,

[08:58] we'd get iridescent. Because look, when

[09:01] light comes in at an angle, it can

[09:03] bounce off this top branch, then the

[09:06] second from top branch on this

[09:08] neighboring tree and so on. And each gap

[09:10] is the same distance. So you've got

[09:13] constructive interference for say red

[09:16] when viewed at this angle. But cleverly,

[09:18] the moro butterfly has all these trees

[09:21] offset randomly at different heights

[09:23] from each other. So when light comes in

[09:25] at an angle, it doesn't see regularly

[09:28] spaced reflectors. it sees a random

[09:30] mess, so constructive interference is

[09:33] impossible. So, you only get reflected

[09:35] light when it's coming straight down,

[09:36] and that's always blue. Chameleons do a

[09:39] similar thing with those guanine

[09:41] crystals, but it's not as easy to

[09:43] visualize. But anyway, we finally come

[09:45] full circle back to the Litman plate.

[09:48] So, when I breathed on the back of this

[09:49] Lipman plate, the gel on the back

[09:51] absorbed some of the moisture from my

[09:53] breath and expanded. And so just like

[09:55] the crystals inside the chameleon skin,

[09:58] the structures inside this plate

[10:00] expanded and separated when engorged

[10:03] with water. And that separation causes

[10:06] the color to shift. But what are these

[10:08] structures inside the Litman plates? And

[10:10] how did they get there? To figure it

[10:12] out, let's look at how one of these

[10:14] pictures is taken. First, take a sheet

[10:16] of glass and coat it in a photographic

[10:19] emulsion. This emulsion is made of

[10:21] gelatin with lots of teeny tiny crystals

[10:24] embedded in it. These are tiny grains of

[10:27] silver haloid, meaning a latis of silver

[10:30] and something like broomemide. You then

[10:32] put the plate face down in a bath of

[10:35] mercury. I don't have any mercury, so

[10:37] you'll have to use your imagination.

[10:38] Now, let's see what happens when the

[10:39] plate is put in a camera and exposed to

[10:42] light. In particular, we'll zoom right

[10:44] in and see what happens when yellow

[10:46] light hits this little section of the

[10:49] plate. First, it passes through the

[10:51] emulsion from the front and then it gets

[10:53] reflected off the mercury bath and then

[10:56] it passes back through the emulsion from

[10:59] the back. And crucially, the reflected

[11:01] light interferes with the incoming light

[11:04] and that creates a standing wave. You

[11:06] know, standing waves, right? The

[11:08] important thing is you've got light

[11:09] energy here at the antinodeses and no

[11:12] light energy here at the nodes. And so

[11:15] these crystals here at the antinodeses

[11:18] start to absorb photons. And eventually

[11:21] as a result, some of the atoms in those

[11:23] crystals break away from the halite

[11:25] atoms and form metallic silver flakes.

[11:28] Just over here, the plate is maybe being

[11:31] exposed to blue light instead. Blue

[11:33] light has a shorter wavelength. So these

[11:35] bands of exposed crystals are closer

[11:38] together. In other words, the color of

[11:40] the light is encoded into the spacing

[11:42] between these bands. After the image is

[11:44] exposed, the plate is developed. And

[11:46] this bit is really clever. So you've got

[11:48] this emulsion with silver haly crystals

[11:51] in it. And some of these crystals have

[11:53] little silver flakes in them from the

[11:55] exposure. And these silver flakes act

[11:58] like a catalyst during development. so

[12:01] that any crystal that contains even a

[12:03] tiny amount of metallic silver becomes

[12:06] entirely metallic silver by the end.

[12:09] There's a final step that flushes away

[12:11] all the silver haly that didn't get

[12:14] converted into metallic silver and what

[12:17] you're left with is lots of tiny silver

[12:20] mirrors embedded in gelatin. Then to

[12:23] view the plate, you just shine white

[12:25] light onto it. And every time the light

[12:27] hits one of the mirrors, some gets

[12:29] reflected and some gets transmitted down

[12:32] to the next mirror and so on. And just

[12:34] like with the soap bubble and the

[12:37] chameleon and the butterfly, the

[12:39] resulting interference produces a

[12:41] specific color that depends on the

[12:43] spacing of the mirrors. In other words,

[12:45] the Litman plate contains a defraction

[12:48] grating that encodes all of the spectral

[12:52] information from the original scene. So

[12:54] why does it only work when it's

[12:56] illuminated from certain angles? Well,

[12:59] with normal pigments, you have light

[13:02] scattering. So white light can hit the

[13:04] pigment from any direction. Some of it

[13:07] gets absorbed, some of it gets

[13:08] scattered, and some of that scattered

[13:10] light reaches your eyes. But when light

[13:13] scatters off a Litman plate, you don't

[13:15] get interference. So you see it as black

[13:17] and white. It's only when the light

[13:19] reflects off the surface that you see

[13:22] the color. Which is why when I shift the

[13:24] focus of my camera lens here, you can

[13:26] see that the light source is reflected

[13:29] there in the plate. You know, Litman

[13:31] plates have their drawbacks. Reprints

[13:34] are basically impossible. Exposure times

[13:38] run to minutes. the restricted viewing

[13:41] angle is less than ideal and they can

[13:45] get a bit washed out if you don't get

[13:47] the process exactly right which is

[13:50] probably why pigment-based RGB

[13:53] photography eventually came to dominate.

[13:56] So, Lipman plates might feel like a

[13:58] historical culde-sac, but it turns out

[14:01] that holograms like this one that can be

[14:04] viewed with reflected white light are a

[14:07] direct descendant of Litman plates. So,

[14:11] you know, thank you, Gabriel Litman. You

[14:14] deserved that Nobel Prize. Right,

[14:16] massive tangent. I love scams. What I

[14:19] mean is, well, I'm interested in

[14:21] everything, including scams. So, when I

[14:24] get a scam phone call, I try and keep

[14:26] them on the phone as long as possible

[14:27] because I'm interested to know like what

[14:29] are their tactics and stuff like that. I

[14:30] don't particularly recommend it cuz you

[14:32] could end up giving away something

[14:34] unintentionally. But let's have a look

[14:36] at an example. This is a phone call from

[14:38] my bank. So, you're

[14:42] >> Yes, you

[14:46] wigs. I thought you'd know my address.

[14:48] >> No, I can send you over a text message

[14:50] now.

[14:52] Yeah. Okay. It's interesting to see how

[14:54] they socially engineer their way around

[14:57] various safety features built into the

[14:59] banking app. For example, there's a big

[15:00] banner in the app that basically says

[15:03] we're not calling you. So like if the

[15:05] bank isn't calling me, who the hell are

[15:07] you?

[15:07] >> Yeah, it's just Okay, I'm just telling

[15:09] you now. It says here, you've never

[15:11] spoke we've never spoken to you on the

[15:12] phone.

[15:13] >> Yeah, that's correct, sir. That's what

[15:14] we've got added onto your account cuz if

[15:16] somebody else has logged in, they can

[15:18] see when's the last time you've been on

[15:19] the call with us here, sir. Maybe you're

[15:21] seeing some red flags here. But the

[15:23] thing that gave it away perhaps more

[15:24] than anything was just how frustrated he

[15:27] was getting with me. Anyway, carry on.

[15:28] So, what what what do you need me to do?

[15:31] >> Are you recording the call, you

[15:34] >> I'm not recording it. Are you genuinely

[15:38] >> Are you sure you're not recording?

[15:41] >> This guy that wasn't very convincing. I

[15:43] just don't think my bank would speak to

[15:45] me like that. It became clear from the

[15:47] call that they had my bank details,

[15:49] which was worrying. And I actually

[15:51] figured out how they must have got it

[15:52] because while he was trying to persuade

[15:54] me of his legitimacy, he read out my

[15:56] email address, but the email address was

[15:58] wrong. It's not the one from the

[16:00] account. It's a random one that I hardly

[16:01] ever use. But I remember I did use it

[16:04] recently for a specific transaction on a

[16:07] camping website. So my guess is that the

[16:09] caller must have used something like SQL

[16:11] injection on that camping website to get

[16:13] my car details. But anyway, every time

[16:15] you put your details into an online

[16:17] form, you're hoping that the company

[16:18] behind it has good security. Obviously,

[16:22] that's not always the case. So, it's a

[16:24] good idea in general to try and minimize

[16:28] how many companies have our data. The

[16:30] good news is that our data has been

[16:32] consolidated to a large extent by these

[16:34] middleman companies. So, if you can get

[16:36] your data off their databases, it can't

[16:38] trickle down to all the others. Not only

[16:40] would you be less vulnerable to data

[16:42] breaches, you'd also get fewer spam

[16:44] calls. The bad news is there are

[16:47] hundreds of these middleman companies

[16:49] called data brokers. And they all have

[16:51] different procedures for requesting data

[16:53] removal. Honestly, if one person could

[16:55] just sit down and do all the hard work

[16:58] of figuring out how to interact with all

[17:01] of these different companies to remove

[17:02] their data, well, I'd happily pay them a

[17:05] little bit of money to do the same thing

[17:06] for my data since they've already done

[17:09] all the research. And of course, that's

[17:11] exactly what I've done. I'm talking

[17:13] about Incogn, the sponsor of this video.

[17:16] It's super easy. You give them your

[17:18] details, you give them permission to act

[17:19] on your behalf, and then they just go

[17:21] away and do it. They also do people

[17:23] search sites, which is reassuring. And

[17:25] on the unlimited plan, you can do custom

[17:27] removal requests if you find your data

[17:30] anywhere out there in the wild. I did it

[17:33] because it turns out that past Steve was

[17:35] doxing future Steve. So, while you might

[17:37] not be able to protect yourself from

[17:39] every camping website that hasn't

[17:41] sanitized its inputs, you can protect

[17:43] yourself from data brokers and people

[17:45] search. If you're interested, the offer

[17:47] on this one is really good. Go to

[17:48] incognite.com/science

[17:50] and use code science at checkout to get

[17:53] 60% off an annual membership. The link

[17:56] is also in the description, so check out

[17:58] Incogn today. I hope you enjoyed this

[18:00] video. If you did, don't forget to hit

[18:01] subscribe and the algorithm thinks

[18:03] you'll enjoy this video next.

[18:07] What's that?