Curiosities and unusual trivia about light and color
A graphically illustrated guide to various oddities about color and light. You'll find various amusing contradictions about the perception of color. Have you ever wondered how many pictures (or movies!) you could ever ever ever see? Did you know that blue and yellow actually makes grey/white (well, it's true!). Then there's the 'Mirror Timebomb' - see how a sphere in the shape of a mirror can explode if lit from the inside. Also, the shocking news - "Is your monitor displaying a fake green?". You'll not only find it probably is, but have a chance to see what /real/ green looks like. All this and lots more! Enjoy...
There have been various ways of measuring the speed of light - from using precision engineered rotating wheels... to the 'nudges' of the stars caused by stellar aberrations... to Galileo's classically inaccurate lamp test (visit http://www.what-is-the-speed-of-light.com for more info), but there's a spectacular visual more fascinating than any of these which will now be revealed!
As much as have all come to accept that the speed of light travels at 186,000 miles per second, there will still be some confusion as to what the ramifications of this might be. Just stop and think a moment. What would everything look like if the speed of light was slowed down to - ooooh... right down to 1 metre per second ?Weird things would happen. Such as these: a: If you switch on a light bulb, nothing would happen at first, but then after 2 to 3 seconds, you would then see a bright light bulb hanging in mid air - everything else would remain pitch black. b: Very shortly afterwards, the ceiling shows a circle of light which would grow and grow... then soon after this, the nearest parts of the room would illuminate and eventually the whole room too, but... c: Shadows would remain pitch black for a little longer! Remember, what usually stops the shadowy parts of the room from remaining pitch black is that the light bounces off the other parts of the room and reaches these shadowed areas. If the light was slowed down to 1 metre per second, these shadowed areas would gradually lighten up as more and more parts of the room reach them. d:
If you tried suddenly moving your hand, an area in the shape of your hand would momentarily be pitch black! This is because the volume of your hand is taking up space that light could have used instead. Also if you tried moving, your own shadows on the walls would drag behind you, and try to 'catch' you up. The further the shadow is away from you, the longer it would take for it to catch up. Lots of swirly effects possible =) e:
Finally, if you then tried switching off the light, the shadows have their revenge as they are the last pieces of light to reach your eyes!
NB. I found out that one of these five effects is actually wrong! See if you can guess which one it is. Post up to Skytopia Forum once you think you know - or post up some extra effects that I haven't thought of!
Anyway, all this sounds very interesting, but it is... entirely theoretical. Light will never travel that slow (though see the on/off MPEG video near the bottom here for a CGI demo). For things to be visibly slowed down, things need to be done on a grand scale!
Take a look over to the diagram on the left. 2 mirrors are facing each other - simple huh? As you probably know, this has the effect of dozens of mirrors appearing inside each other. But imagine if these mirrors were 100 miles wide/long and were distanced at about the same length. You're there - right in front of the blue mirror - looking directly into the red mirror.
The first thing you would see of course is the red mirror itself.
Then you would see the blue reflection inside of this.
Click the picture for a full screen version! If you would like to see this picture in 3D stereo, then visit the Skytopia Stereoscopic 3D gallery
Next we have the reflection of the original red mirror's reflection (2) and inside of /that/ we have the blue reflection again. At this stage, we're now 4 times smaller than the original. Light has travelled 400 miles to reach you... and in just one 500 hundredth of a second. This is not yet long enough for one to notice a 'building' up effect... yet... =)
Here we go... it would take light:
1600 miles and 16 reflections in 1/116 second
6400 miles and 64 reflections in 1/29 second
25,600 miles and 256 reflections in 1/7 second !
This is just about big enough to be seen and slow enough for some weird stuff to happen. Try rotating one of the giant mirrors slightly or shooting across a large object in between the mirrors at high speed!
After you're done, the next task is to create a 100 mile diameter mirrored sphere and go inside. Hours of fun as your reflection 'multiplies' in the sphere.
If you don't have access to two giant 100 mile wide mirrors, or you don't have the ability to slow down time, then one way you can see this effect is to try pointing your webcam at the monitor screen, or video camera connected live to a TV. With any luck, you should see images appearing one by one on the screen. Sometimes the effect can be beautifully dazzling, like a kaleidoscope (at least with my video camera). You may however need to use an old Cathode Ray Tube TV though as modern LCD displays have a degree of lag to them. [Back to top]
The Mirror Timebomb
Imagine if you will a mirror..., and not just any mirror, but a specially designed spherical hollowed out mirror. The mirror is the surface of the sphere but on the inside. Now add some kind of light source inside (a light bulb will do for now). Turn it on and what do you think would happen?
Inside the mirrored sphere. The moment in time just before the whole thing 'exploded' - caught myself by a special heat sensing camera (wide-focal lens) - located outside and looking into a tiny 'pin-hole' in the mirror's surface. I only had one chance at this, considering the cost of the experiment, so I was glad it went to plan! It was important to get the image straight to the PC (milliseconds) before the heat melted the camera! Click here for a higher res.*
* Joking of course, this image was actually rendered with a 3D program
As the light rebounds inside the sphere, the intensity grows and grows - as light is rebounded for the umpteenth time - and
more light is continually being pumped out from the bulb.
It has nowhere to go! What happens next???
Assuming the light really cannot escape, two things will start to happen... a: The inside will become brighter and brighter as light builds up - and countless more photons are swimming inside. b: The air and surface of the sphere will become hotter and hotter - as light is turned into heat.
Now what happens?
To keep all this energy from escaping somehow, we need perfect materials which can reflect almost all light. Perhaps if it was filled with air, and (near) ideal materials were used, we can either expect it to melt away over a period of a few seconds, or the thing suddenly explodes in a great fury - as particles, photons and other subatomic particles all rush to escape.
[Back to top]
Yellow + Blue = green" ?????? (bzzzzzz). False.
A few basics on colour
Primary colours: Red, green and Blue
Secondary colours (ink primaries): Magenta, Yellow and Cyan. White should also be classed as a secondary color, because it is still based on the primary colors (all three of them).
Additive mixing (combines spectral reflectances):
Examples: red+green=yellow, magenta+yellow=half saturated red, red+yellow=orange.
The brightness (value) will vary according to whether you're using chess board style colour splitting (average), translucency (average) or monitor style 'mixing' (addition). But the main thing to remember is additive mixing will always keep the hue and saturation constant.
Multiplicative mixing (combines spectral transmissions): Multiplicative mixing is more commonly known as subtractive mixing unfortunately - see this site).
Examples: red*green=black, magenta*yellow=red, red*yellow=red. Light filters are a perfect example of multiplicative mixing. So are most dyes and inks (think of felt tip mixing).
What type of mixing does paint use?
Mixing paint is not so simple, and neither multiplicative or additive mixing will do. It turns out that the mega-complicated Kubelka-Monk formula is needed to precisely find out the mixed color. It can very roughly be approximated to something like the geometric mean or 'minimum' function, but this a dirty shortcut, as it won't take into account things such as light scattering and the precise absorption spectrum of the paint etc.
It as a long held myth among the general population that blue and yellow equal green.
The concept is ingrained from primary school and is often assumed even at college/university level... It is of course incorrect.
There are two main ways of mixing colour - additive mixing, and multiplicative mixing (This is more commonly known as 'subtractive' mixing, but mathematically, this is incorrect). In either case, blue and yellow will always make white/grey or black.
There are three potential reasons why people think that blue and yellow equal green. These are:
1: When people see the colour Cyan - they think they're seeing 'light blue'. They're not. If you try mixing a more cyan-like blue, then naturally a certain amount of green will be produced.
This is light (or pale) blue
This is 'Cyan' (or as cyan as you're ever going to see on a CRT or LCD display - see later for comments about the poor cyan and green on monitors)
2: There's no such thing as pure 'blue' paint (i.e. no such pigment that only stimulates the eye's 'blue' cone). Or rather we don't have the technology to create it yet. In other words, what's seen as blue paint contains many wavelengths of light around the green part of the spectrum, and obviously that's going to produce a muddy green if it was mixed with yellow.
This is almost blue, except it's got a small amount of green/cyan in
... if the 'almost blue' was mixed with yellow, it would produce a hue something near to this muddy green
3: On top of both those factors, there's also the possibility that complex chemical reactions within the paint would distort the hue. See this site for a quick overview of the impurities in inks and paints.
All of these factors combine to make it confusing, but the fact remains that real blue and yellow can only ever make a hueless colour - white/grey/black.
To demostrate further, why not experiment with translucency in your favourite paint package.
Likewise, Green + Magenta = black/white, Red + Cyan = black/white. The color wheel in the diagram to the right will show these colors are opposite each other.
If you're interested in mixing colors and light generally - visit this excellent site which clearly explains color mixing, and how the three ink primaries are Cyan, Magenta and Yellow, NOT Red, Yellow and Blue: http://home.att.net/~RTRUSCIO/COLORSYS.htm. Here's another site giving the advantages of the CMY model.
Also interesting is this discussion on mixing colours using inks such as quinacridone magenta, hansa yellow, and the rarely used 'phthalo cyan'. [Back to top]
Can we really see in ?? False.
We cannot see in 3D at all - but 2D. Sure, the information we receive is from a 3D world, but the image projected onto each eye is only 2D. If we could see in true 3D, then we'd actually see
/behind/ stuff - even opaque objects.
The image projected onto our eyes is a 2D image, and we have 2 eyes, so it's (x*y)+(x*y), not (x*y*z). No doubt it'd be a case of information overload if we really did see things in 3D.
Having said all this, the 2 images combine in a very special way, to form a pseudo 3D image in our minds eye. The brain somehow automatically converts differing levels of 'blurred-ness' and the gap of the dual (cross-eyed) images to represent distance. A stunning piece of brain 'calculation' if you ask me... [Back to top]
How much information are our eyes receiving every second?
Let's give a value of 10,000 'pixels' for each dimension (which is roughly as good as we could wish to see), that'd mean we see:
10,000*10,000 pieces of information for each eye. Multiply this by a factor of two for each eye and then further multiply it by 24 (8 bits for Red, green and blue) to represent the 16 million
colors we see. So far we've reached a number of 4,800,000,000.
Then multiply that by 100 for the bits of info we receive every second and we end up with the not too small number of 480,000,000,000. Not even the most powerful super-computers or the latest optic technology can match the intricacy of our eye.
How many different hour 'movies' could we ever possibly see?
Ooooh, now you're talking. We go into powers here - absolutely stacks. Let's just start with how many pictures we could possibly see. Say if our eyesight was worse than a fly - say if we could only see a visual area of 2 times 2 pixels. And let's just assume we could only see 2 'colors' - black and white. This would mean there are 16 possible pictures we could ever see:
Now let's increase this to a visual area of 3 by 3 pixels - and a slightly less miniscule palette of 3 colors (one more than before!). I'm not going to draw all the possibilities, but there are loads - 729 to be precise (or 93). One only has to now increase the amount of colors possible from 3 to 4 and you get 6561 possible combinations (94)!
The formula is quite simple: colors to the power of pixels (or c^p) Now is where the big numbers come in. How many unique colors could we actually see?
At this point we've got to be careful - we can distinguish between a lot more than '16 million' colors if we include the various brightnesses that are not available on the average monitor. For the sake of this test, I'm going to give a number about 4 times larger than the 256 amount of brightnesses available on a monitor... - say about 1000*1000*1000 (RGB).
So we've got 1 billion (1,000,000,000) colors and 100 million (100,000,000) pixels to shape our vision. Now all we have to do is calculate what 100,000,000 to the power of 1,000,000,000 is.
Simple if you know how. Most calculators can't handle numbers as big as this, so there's a trick involved. You just work out the logarithm of 1,000,000,000 - which is 9 'digits' (looks like 10 but we use 9 in our calculations). Then we multiply this amount (9) by 100,000,000.
So, the amount of digits that our number of possible pictures contains is.....(drum roll)...
This covers every single picture you could ever possibly see. From a picture of you - to a picture of your own house, to a picture of the ocean or sky - every possible sight is contained within this number!
Note that is only the amount of digits the number contains. To print the number out in full would take up millions of pages and around 2 CDs!!!
But we haven't finished yet! Oh no... We still haven't taken into account the factor of time/animation! :)
OK, let's say, we watch a movie - lasting an hour. How many movies could you ever possibly see?
This number is horrifically big - going into powers of powers. Here goes...
We want to calculate that 900,000,000 digit number to the further power of 3600*100. That's 3600 seconds in an hour and 100 frames per second to convince us we're seeing actual smooth animation rather than just a load of stuttering pictures.
So the sum we have to do is:
1,000,000,000 to the power of: (100,000,000 multiplied by 360,000)
This means the amount of possible movies we could ever see is a number containing..... (drum roll)........
Ooooh, I bet you won't see a number bigger than that. It would be gigantic enough even if it wasn't referring to the number of digits. Bearing in mind that if you were to add even one digit to make 324,000,000,000,001; this would make the number ten times as big! To store all these pictures onto CDs would in fact require a number of CDs not much greater than the number we had in the first place..... Well, it /is/ much greater, but only by a few thousand digits - which is kinda nothing at this level.
Every number that has gone into that calculation is printed in this extra special long formula.
Of course, this number doesn't take into account the sound/music and speech alongside the film, so..... right, I think I'll stop there =D I'll leave that for another time. [Back to top]
Is the green on your monitor display a fake imitation of the real thing?
Yes it's true - the green you've been so used to may well be a poor quality imitation! It's a common problem with many TVs and monitors and as far as I know, it could well apply to all cathode ray tube and LCD displays. What actually makes the green so poor is a fair amount of red (and a bit of blue) contamination. This makes everything that's coloured green on your monitor white-yellowish. Because cyan is made up of the poor green element, this also tends to be a very pale imitation of the real cyan too.
To test how good your monitor is, there are two simple ways. The most efficient way requires you to have a special green plastic see-through filter (you might have something like this without realising - get those 3D glasses - one side has a green filter!). If you do have something like this, then skip to the sub-heading "So what does this 'real green' look like then?" - otherwise carry on reading.
OK, you'll need to wait until late so that the room will be dark once you switch off the light. Then refer back to this doc.
Now get a CD (preferably a recordable CD - as the silver sides tend to be shinier). Now in your favourite paint package, fill the whole screen with green. Make sure there's nothing except green on the screen (a black border doesn't matter of course if it can't be helped). Once you've done this, use the silver side of the CD and tilt it at various angles to reflect the screen. You should be able to see a disc of light on the silver side of the CD. In the middle, is the colour of the green screen reflected. But on the edge of the colour band ring, you should see a faint red aura. On the inset of the main rim you should also be able to see a tinge of real green. If so, this proves your monitor is as faulty as the rest of them :) If on the other hand, you can only see one tone of green - then lucky for you! - your monitor is an exception to the rule.
"So what does this 'real green' look like then?"
If you have a green filter (get those 3D glasses - the green see-through plastic in them is ideal!), then you'll be able to see what it will look like. Otherwise, well..., I can't magically produce a true green for you, but the best I can do is show you a table showing what could be next.
The "What next?" box represents what should be the true green (yes - it really is this much of a jump!)
Stop press: There is a brilliant way you can see true green and cyan without resorting to filters or CDs. Visit the Optical Illusions page and look for the "Eclipse of Mars" illusion.
For those who have the green filter, take a look at this below. I have adjusted the shade of the color to be as dark as the filter will show. Unfortunately, the brightness of the green isn't marvellous (because you're using a filter), but at least you'll be able to see a (fairly dark) /real/ green. Put the filter over the white portion on the right.
55% green (0,140,0) A close match in terms of brightness (not saturation!) to the green filtered version
(for those looking via green filter) - Now, that's how real green should look!! In comparison, the 'green' on the left looks very muddy.
Shocked? You should be. As far as I know, the saturation level of green is about as bad with both TFT/LCD and CRT monitors. Apparently, the future of display technology - OLED - has much better color saturation.
Maybe some of you are going to say "ah... - all you have to do is brighten/darken the green shown, and then I bet it will appear as green as how the filter shows". For you lot, here's a color chart - showing dark 'green' (0,0,0) to bright 'green' (0,255,0). Nowhere will you find a green as pure as the filter shows.
For the ultimate proof; put a red filter over the bright screen square shown to the right. If the square was truly green, then the filter overlay should make it black. Since, it isn't, there must be some red light coming from the fake green, and therefore through the red plastic filter.
Here's another graduation table should you wish to check out the purity of the red element (red filter needed of course). Red is actually very good and much better than monitor green, but by no means perfect. [Back to top]
This box represents the next jump for real red.
Magenta, Ultraviolet light, Infrared radiation and beyond
You might well know that the colors that we see reside in a relatively tiny portion of the electromagnetic spectrum - from 400 to 700 nanometres. At the limit of these points (ultraviolet and infrared), they decrease in intensity until nothing can be seen at all.
At the fine range of the visible light spectrum - exists violet. This is interesting because red is needed to mix with blue to get a similar color to violet. At least /some/ red is needed... but hang on a minute... red is all the way near the beginning of the light spectrum! How can this red mix with blue (to produce the aforementioned violet) if they're miles apart?!
This seems to imply that if we /could/ see further along the invisible electromagnetic spectrum, the colors would wrap round and we'd see red, green and blue, red, green, blue again and again and again, but the truth is a little more subtle than this.
To understand what's really happening, I first need to explain that the eye uses three types of cones: red, green and blue. Wavelengths of light around 700 nanometres stimulate the red cone. For the green cone, it's around 550 nm, and for the blue - about 450 nm.
But our eyes have a slight quirk about them - when a wavelength near the blue end of the spectrum (approx. 450nm) activates our blue cone, it also (very slightly) activates our red cone too! Note that this doesn't happen the other way - red wavelengths don't stimulate the blue cone.
Most diagrams on the net tend to show three curves in a spectral response diagram such as this lot:
See where the red curve gets 'second helpings'? It shows how wavelengths around the blue area will unexpectedly activate the red cone slightly. In other words, we'll never be able to see what true blue really looks like!
Idiosyncracies, unusual artefacts; the science of light is full of them. In fact, more and more is being discovered about light and color even today. [Back to top]
Light bulbs - why cream?
Ever since Edison invented the light bulb, every household up and down the globe has had to put up with the same cream light bulb. It's a pity, because white would accurately reflect the true colors of objects. Is it because of technical difficulties why white light isn't common-place - or is it because people prefer the 'warmer' tones that cream light provides? Incredibly, not one scrap of information (apart from what you're reading now) is on the web regarding this. Must be one big conspiracy ;)
Also annoying is the general dullness of bulbs. I don't know about you, but even the 150 watt bulbs that are available are way too dull. It's almost like being asleep...
300 watt, 500 watt and even 1000 (!) watt would be invigorating in comparison. What would be cool is if it was as bright indoors as it is outdoors (when the sun is out). Only a personal opinion, but also probably very true =P Oh, the electricity bill? Never mind that - I am talking about in an ideal world...
Anyway, back to the cream bulb business. Exactly how cream are they?
I put the theory to the test. For the sake of the experiment I have to assume that the monitor white really is true white.
First, I got a piece of paper - a white piece of paper. I then used the RGB (red, green, blue) sliders to try and match the shade of the paper. What RGB values will produce the apparent 'white' of the piece of paper?
Here's a quiz to demonstrate the idea. I realise your light bulb might have a different brightness, or that your monitor may be darker or lighter, or even that you might be holding the piece of paper at a different distance or angle, but give it a shot - I think you might be surprised. Make sure of course it's evening and the curtains are closed to block any light from outside.
Which box is closest (in terms of hue /and/ brightness) to the shade of a piece of white paper?
It's a total shock at first. That dull dark brown really is approximately as bright as the brightest shade in your dull-lit cream room!! Depressing huh?
In comparison, full spectrum light bulbs would contain all 3 primary colours in equal measure - ~33.3%
Worse still is the amount of blue in the average light bulb - a very miniscule 14%!
The answer to the quiz in case you haven't already found out is "color number 2". To get the shade closer to the actual paper shade that you've got, try altering the color in Deluxe Paint or Paint Shop Pro. You'll find that even if the brightness varies to the figures I've got, you should still get the same color hue ratios...
To verify the find, I used the same proportions (53%, 33%, 14%) and made red the maximum brightness (100% or 255/255). I calculated the values for green and blue easily enough:
Green: 157/255 (33%) and
Blue: 68/255 (14%).
Now was this very suspicious orange color (shown below) going to match the theory?
Surely it looks a bit too orangey.....
I eagerly moved the paper closer to the light - and using trial and error, adjusted the angle so that it would be darker/lighter to try and match the shade on the screen. Was it the same hue???
Very close indeed - I was right.
The thing you must remember is that stuff looks less apparently orange when everything around it is also tinted with the same color. The eye 'adapts' and creates a new standard for 'white' according to what it sees. So even though this orange looked too orangey for the bulb's light, it's only an optical illusion (in the same way it's an optical illusion that it looks darker on screen than in real life). For another real-life example of what I mean, visit this page here. Colours can easily appear to be something they are not.
Anyway, it worked - closing the paper on the light source made it brighter and it was identical to the bright 'orange' hue on the screen. We really are seeing everything Orange-ey in the evening!
Incredibly, it's this exact hue of orange (51% red, 32% green, 17% blue - (255,163,84))....(only much, much brighter) that 60 watt light bulbs use to radiate stuff. Pathetic - I know.
100 watt light bulbs don't fare much better either. It's this exact hue (49% red, 33% green, 18% blue - (255,169,93))....(only much, much brighter) that it uses to light up everything. Also pathetic.
A proper white light bulb would need about 3 (!)
times as much blue and over 1.5 the amount of green. Halogen bulbs fair a bit better but not much:
Now try finding the brightest red you can find. Perhaps you'll find a bright red cover off a magazine or food item. What RGB color shade from the monitor do you think it's closest to?
You'll find it to be at a brightness of about 90/255 which is unbelievably - this bright:
The reason why it looks so dark on the screen is mostly due to the way it's contrasted against the bright white background. If that red was instead contrasted against the wallpaper in your room, it would be quite bright! [Back to top]
UPDATE: THIS SECTION WAS BASED ON A 6500K MONITOR COLOUR TEMPERATURE WHICH IS 'PRETTY WHITE'. HOWEVER, THERE'S REASON TO BELIEVE OUR PERCEPTION OF 'REAL WHITE' TO BE CLOSER TO AROUND 5000K ACCORDING TO THE EVIDENCE PRESENTED AT THIS PAGE, WHICH IN TURN WOULD SLIGHTLY ALTER THE RESULTS ABOVE, BUT NOT BY A GREAT DEAL
If the info on this site has been of sufficient interest, a small donation would be appreciated:
Dynamic Optical Illusions - - - View a collection of old, new and exclusive optical illusions, and see the best voted illusion of all time - The Eclipse of Mars.
Mike's Electric Stuff: Some very interesting experiments. Fun with Plasma, Xenon flash tubes, Neodymium magnets, Lasers and lazer reflections... and the effects of microwaving a CD! Good science fair project ideas... ;-)
Robert's Colour Mixing page. A clear introduction to the science of colour mixing - explaining additive/subtractive color and featuring cool color charts. It even taught me a few things...
Cool Physics movies Animations of jets breaking the speed of sound, and what you would see if you travelled past Saturn at 99% the speed of light...