Lecture 11 - Color and Artifacts Video Lecture | Exposing Digital Photography by Harvard Extension School

hello everyone welcome to digital media
e 10 so I hope last week was awesome and
instructive I certainly enjoyed watching
the the guest lecture and I want to
thank again Shelly and John for for
throwing that together and presenting
that to you today is is color and
artifacts and of course before we get
started just a couple of quick reminders
this is the first week in a while where
we don't have something due today in
terms of the projects or the problem
sets of course the critiques are always
or do every week but we do have the
problem set 5 to next week and again
this was a three week problem set
because some of the material from
today's class is actually applicable to
that problem set so we want to give you
time to actually use in order to
synthesize the information from today
and then the final project proposal is
due on the is due next week as well and
so do start thinking about that that
proposal it's not - you don't have to
write too long in the document really
one of the most important things that we
want to understand is if the scope of
the project that you have in mind is
going to be sufficient is it actually
going to be of approximately two of the
numbered projects in terms of quantity
of work and in terms of the scope of
that of that project and that's really
what we're looking for and do feel free
to get started on your final project
even as you work on your proposal
because we may it's extremely rare that
we just outright reject the entire idea
for a proposal in fact I don't think
we've ever done that but instead we work
with you to try to adjust the scope to
fit our expectations either ask you to
tone it down a little bit or beef it up
depending on the amount of work that you
that you described to us in the proposal
and so do get started on that and even
if you do miss the proposal deadline
please do send one to us anyway because
we want to know what it is that you're
up to in fact that's a requirement for
submission of the final project at some
point before you actually send us the
implemented final project we have to
look at and approve a proposal even if
it's a few
Slate's so do be sure to take a look at
that and hopefully in sections we've
been talking about some ideas about
about final projects and sort of
nurturing that that discussion as well
so today we're going to talk about color
and artifacts and there's been a fair
amount of discussion about color at
least about the color perception of the
eye and by a fair amount I mean not very
much at all but today we won't get to
beef it up a little bit when we talk
about the way that digital cameras use
color that computers use color and the
way that we have to actually pay some
extra attention to it to make sure that
the color is consistent across all of
your devices and all of the other things
that come with this but first before we
get actually into that the reason that
this is important is that the eye that
we perceive light the way we perceive
color is very different than is
mathematically or scientifically true
and so in other words the way that we
perceive it is biased based on a number
of things and this it's very hard to
enumerate all the things mostly because
I don't know them but also because
there's so much that our visual system
is trying to achieve to try to represent
the world in some way that's not just
accurate but useful to us and a lot of
things that have sort of come to play in
the past few maybe millennia worth of
evolution is trying to make things make
visual sense to us so just as a couple
of examples I have some optical
illusions here how many of you how many
of you have seen this one before
nobody awesome cool so this is like a
tiled floor and there's an object with
the light source off to the right and
the light source head has some light
that hits the object and cast a shadow
across the floor now you'll notice that
there's two tiles that are in fact
labeled there's one at the very top
labeled a there's one in the middle
labeled B and the great thing about this
illusion is that even when you know what
the what it is it's still it still
tricks you so in this case a and B the
tiles a and B are exactly the same color
so even if you know that you might be
looking at me like no that doesn't sound
quite right this doesn't think you're
making this up well like I'll prove it
to you okay so here's the alright here's
the slide and I have this app on my
computer which comes with Macs called
the digital color meter what it does is
it basically allows us to display color
color information in this case red green
and blue values from 0 to 256 just 0 to
255 and 8 bits just like we've been
talking about computers will actually
display the color information so here
you'll notice that if I hover my mouse
over this black area of this slide it
looks black and so we see values of 0
for red green and blue over here the
kind of grayish area or light grayish
areas this very light gray at 256 so on
ok so let's take a look now at this so
we have push this out of the way let's
look at the color information for tile a
notice that it's this kind of it'll grow
139 139 139 now if I move my mouse down
to over B notice that's exactly the same
value now you still may not be convinced
you may be thinking okay this is very
carefully orchestrated plot to destroy
my perception of color so I also have
this same image in Photoshop no where
was it out here it is same image in
Photoshop and what I'm going to do is
I'm using the eyedropper tool to pick up
the color this is the same value that we
just picked up a moment ago and I'm
going to use a brush to actually paint
that color from A to B and now hopefully
you can actually see that it is in fact
the same color once we actually remove
the shadow out of the way then all of a
sudden does this optical illusion become
a lot more apparent pretty crazy right
but this kind of implies some issue this
that we might actually have with color
perception within a scene when we're
actually trying to take a photo and we
assume that it's actually some color and
then we might look at the image later
and perhaps it actually appears slightly
differently than we might have wanted or
expected it so that's this is really
just shows the context through which
because our brain Percy
this area as in shadow it actually is
supposed to our visual cortex actually
provide some signal to us that says that
this color information is actually
different because it's actually in
shadow then then colors that are outside
of the shadow range but there's also
other things that happen to our visual
system as well so I'm going to show you
this image and what I want you to do is
to stare at this very center that the
black dots and just do it for about
thirty seconds and I'll talk to you
while this is happening and I promise
what's going to happen is nothing scary
is gonna jump out no monsters nothing
like that but in fact what I'm going to
do in the in about twenty more seconds
is I'm going to change this slide and
the slide is going to look very similar
in fact the image itself is going to be
identical except for the fact that all
of the color information will be removed
and when I do that slide change I want
you to keep staring at the very center
the black dot it'll be all aligned so
that it's actually exactly the same and
and without blinking see if you can see
how this image actually changes in
appearance okay ready I'm about to
change it in three two one
do you see blink and then like look away
and then look back did you see what
happened to that image what happened to
this grayscale image after you had
stared at that inverse color for about
thirty seconds I'm gonna try it again
maybe not quite for thirty seconds but
again stare at the very center black
dots for a few seconds really focus
channel your your your energies into
this and and it's kind of a really neat
effect that will happen as a result okay
I'm going to change it in three two one
do you see the inverse color so what you
actually see once I change this slide is
the original color of the image sort of
superimposed on top of this grayscale
image and once you blink and look away
and then return your eyes you actually
see that this image is in fact grayscale
so what's going on here well in this
case we have our eyes get used to
specific stimulus and so very much like
our
get used to the fact that there's a
we're in a bright room and it adjusts
likewise so that we can actually see the
the the light within that room a similar
thing happens with color it gets used to
the stimulus that you're providing to it
in terms of the color and it provides
like this negative image of it so that
when I actually go to this black and
white version of the image we actually
see those inverted colors and then once
we like look away and blink and all that
then then our eyes or the the cells in
our eyes reset so that we no longer have
that negative image that persists and
there's of course many many illusions
that that exist here's another great one
this is just the this is an illusion
called the truck slurs fading and if you
stare at the very center crosshairs and
sort of end in your peripheral vision
don't directly look at this blinking
gray dot that's moving around but if you
look at the center crosshairs for long
enough what you will notice is that all
of the color actually just disappears
you see that it takes like just a few
seconds then all you see is this inverse
cyan or green dot that that is moving
around the image do you do you notice
that pretty neat stuff and in fact this
is so prevalent the fact that the tricks
that our visual system play on us are so
prevalent that every year there is a bus
best illusion of the Year contest and
there are significant numbers of entries
every single year and every single one
tends to blow my mind so this is the
winner for this year called the oh dear
the dynamic effing house so let's take a
look at this one this one's just a
couple of seconds I think maybe a minute
long or so but it is also it has some
text along the bottom so do take a look
at that so this is standard sort of
visual illusion where things look
different this based on proximity to
their size this is not necessarily
related to color but this is kind of a
knee
a neat thing all right so we have a
circle that appears to change size
wild stuff
there's a few others as well one of the
other ones that I that I really liked
back when I first saw it released was
this which is that it's it actually
makes it's it's it's very difficult to
describe things that appear to be the
same are in fact not so this is sort of
an issue with our perspective perhaps or
our the way that we would perceive
relative perspective in this particular
image pretty interesting stuff so anyway
let's come back to this idea now with
all of this in mind we have to try to
overcome these issues and try to
represent our color accurately within
our photos so color of course is made up
from the electromagnetic spectrum
wavelengths of light and there's various
ways of representing the full spectrum
of color that we can actually perceive
and this is just one of them this is
this is a graph that that is a
chromaticity graph that shows us the
possible types of colors are the
possible colors that we can actually see
now along the very outside of this graph
from this number 380 along the outside
going clockwise through 525 60 the
greens and then over here to 700 this is
month this is supposed to represent
monochromatic light meaning that there
are wavelengths of light that if there
was some source that was able to admit
photons at precisely those wavelengths
these are the colors that we would see
it is essentially the purest form of
color that we can see there's no mixing
of color to actually achieve that
particular wavelength of light and then
everything in the middle are colors that
we can perceive because it is a mix of
one or more other colors so for example
purple what we would consider purple or
violets is in the middle here between
red right up here and blue down here and
so that is what the color purple is it's
actually a mix of these two colors it is
there's
no single source of monochromatic light
that can actually produce the color
purple that we would find in the middle
here along this along this linear part
of the graph down here now this is
actually pretty interesting when you
think about it because this implies that
in order for us to see certain
wavelengths of light in their maximum
saturation that we need some source that
can actually provide just that single
wavelength of light to our to our to our
eyes and that's precisely one of the
issues that we have is that there are a
great deal of things that try to produce
color such as our computer screens for
example or our camera phones or what
have you or smartphones but what they
end up doing because as you might recall
they have LEDs that are red green and
blue only they cannot in fact reproduce
the full amount of colors that we can
actually perceive so we can try to
define the space through which these
devices will actually display accurate
color and there's a wide variety of
these spaces some of them are
theoretical some of them are practical
in terms of representative of what the
artists our devices can actually display
to us but one of these is for example
the sRGB color space so imagine that we
have this this graph that has all of the
possible colors that we can perceive and
we're going to have three points on that
graph representing a red a green and a
blue because that is going to be
representative of the monochromatic
light the the light that our computers
can actually show us our displays can
actually show us so that means that
those LEDs are really the the red the
green and the blue LEDs at full
brightness are really the full amounts
of saturation that we're going to
receive for that color so really they
can't go they can't exceed the bounds of
their of their design limitations really
so there's a subset of all of the colors
that we can actually see that are beat
that are able to be reproduced by
devices now every device is actually
different the fact that we have this
sRGB color space which is really used
it's it's extremely common it's really
meant to try to unify all of the
different devices with different
capabilities in terms of color
reproduction but not all devices are
actually able to show the entire srgb
space some of them will only be able to
show a limited portion of even this
limited amount of the full color gamut
some of them can show a little bit more
but will shoehorn the colors that that
srgb can actually display into that into
their gamut and then just not show other
colors so the way that we would get a
mix of light is of course to shine these
LEDs had specific illuminations to mix
them into the various colors that we
might actually perceive there's a couple
of and so with this idea of an sRGB
color space we have this we have this
notion of mixing all of the colors
together to get a specific value of
white so if we turn on all of the green
the red and the blue for example then we
we know that we have some white value
but realize that white is it's kind of
subjective as you know if you were to if
you were to take a look at if you were
to read a book for example in
candlelight that light tends to be very
red but this would we would still
perceive the page of the book as being
white and if we go outside during the
middle of the day for example the color
of the the apparent color of the white
would be very different even though we
would still perceive it as white and so
there's a similar thing that we'll get
to in more detail in just a little bit
when we talk about white balance but
there is a specific point in srgb and
other color spaces that defines white
for that color space the point has a
specific name in this case it's d65
which is in the middle there and it
really is just representative of a white
value they've just picked a white value
and they run with they say this is the
white value
this colorspace this is when your
computer receives the value 255 255 255
representing white this is the precise
color value that you should actually
display and just to make this graph a
little bit more concrete realize that we
have an X and an y value Y is
approximately the luminance notice that
the greens are at the very top blue is
at the very bottom and red is somewhere
in between and the x value so along the
bottom here is the chromaticity so
approximately how saturated so varies in
a similar way that we can think of
different colors as having a specific
brightness if we were to remove that
color can we think of colors as having a
certain amount of saturation just by
looking at them and red to our eyes
tends to look extremely saturated so
that's why we have this this particular
style here this is trying to replicate
or trying to acknowledge the human
visual system on top of the way that
that that colors are actually exist in
the world
so anyway coming back to this idea of
having these color spaces we have the
red the green and the blue outer points
defined and anything that we can
replicate within those values is only
contained within this triangle it's
actually not possible to replicate
colors outside of that triangle within
this particular color space so you'll
notice if we take a look at this there's
perhaps quite a lot of area that's
missing from this color space there's a
lot of the scions the greens in
particular are really missing from this
color space and that's kind of a
limitation perhaps of this particular
definition now with that said realize
that srgb is the most common color space
and we'll talk more about what what that
is and what it means but even though I'm
going to show you another color space
realize that this one is sort of the
de-facto standard by default so you can
assume that your your phone your
computer or your printer or operating in
some form similar to the sRGB color
space but coming back to the side
of having very few greens represented by
this color space there is another color
space called Adobe RGB that tries to
extend a little bit into this green area
to try to represent more colors but
realize what's happening here is that
what we are saying is that we still have
values in our image from 0 to 255 but
basically the way that we can think of
these values is that it's it's kind of
like a paint by number system what the
color space is doing for us is telling
the computer that when it receives some
specific value it knows precisely what
color it should actually output to the
user so what that means is if I have a
value of 255 for green and 0 for red and
0 for blue I should see just the purely
saturated green and those greens should
technically be different in Adobe RGB
compared to srgb so we're not changing
the values we still have 8 bits worth of
value in our images but by applying a
color space to those values
we're basically changing the palettes in
a way of what the colors that will
actually be displayed by the computer
that's how it should nominally work is
that the computer should know precisely
given the color space information to
know precisely what a value should look
like and will display that to the user
but there's a lot of issue with this and
there's a lot of additional complexity
that actually goes into this but before
I get into that I do want to show you
the vast array of color spaces that we
actually have so Adobe RGB and srgb are
extremely useful to us as photographers
but there's in fact a wide variety of
other color spaces that exist as well so
here I'm going to bring up a a this
other application called the color sink
utility which is also included with Macs
and we have a graph down here which we
can actually change
to yxy to approximate what it is that's
the same sort of graph that we were
looking at just the second go it's not
quite zoomed out enough for us here but
we can sort of get a sense of what's
happening with the color space compared
to the full gamut of colors that we can
actually see and if I change the colored
space to look at some other ones for
example let's find srgb in this list so
I scroll down here's srgb we can see
that it in fact has a smaller gamut
compared to or a smaller number of
possible representations compared to
that Adobe RGB but let's take a look at
some other ones as well
NTSC is an old television standard and
it was supposed to represent far more
colors in its color space now that might
or might not be true whether or not it
was actually able to achieve that but
that was in fact the color space that
was applied to it so let's see if I have
something else like the color LCD this
one is actually pretty interesting
because this implies that my computer it
has a color LCD has a display that is
able to produce a certain number of
colors and so we could represent those
possible colors with the color space so
that's kind of what we're seeing here is
the possible color space represented by
my laptop screen in this case this is
not the calibrated version the
calibrated version in fact looks like
this so the way that my computer came
initially was with this color space with
a wide array of colors but that was
found to not actually be true when I
used the hardware device called a color
a meter to actually ask a I used that
that device along with some software and
asked it to try to figure out what
precisely were the colors that my
monitor was able to represent we noticed
that it is able to represent a far
smaller space than originally intended
by the manufacturer and in fact monitors
change over time even if you get a
monitor that's a fantastic quality and
has a wide color gamut as the as the
monitor ages
it in fact is able to reproduce fewer
and fewer colors but we'll get to more
on that in just a minute so just sort of
make this idea a little bit stronger
here though there's the color LCD which
is the original color profile that came
with my computer I'll hold that for
comparison and compare it to the actual
proper color calibration of my machine
which of course is much smaller still
even compared to the sRGB color space so
there's a wide variety of profiles and
color spaces that might exist for a a
computer for our purposes we care mostly
about those two color spaces that I had
mentioned before srgb and if you are
printing perhaps you care about Dobey
RGB perhaps if you're doing professional
printing or if you have a printer that
has more than three inks then you might
use even a different color space or even
a like a CMYK style color space where
there would be four points instead of
three because you have four inks to
actually represent the various colors on
your machine so what does all of this
mean for us practically speaking how do
we actually use this to make our photos
better well realize that there's a whole
chain of items of devices that are used
when you are actually creating a digital
digital photograph one of them of course
is a camera and the other one is most
likely a computer so if we simplify the
whole chain justit's just to these two
devices this is essentially what is
happening we capture a photo with our
camera and hopefully it's properly
calibrated by the by the manufacturer to
actually know precisely what color
information was read from the sensor and
then it will convert that into those
values those zero to 255 values for all
of the different colors and embed within
it a profile saying that this value
represents this color
and because our device is talking in a
specific color space we know when we
read this file in particularly when our
computer reads this file like a
Photoshop or a preview or whatever
application you're using to actually
view this image it knows that those
values represent some specific color
that it wants to display on the screen
but realize that there's even more to
this chain here because we have this
paint-by-numbers sort of system where we
have an image file that has a specific
color in the very first pixel and
another specific color in the second
pixel and so on and the computer is able
to interpret that to know that that
specific colors supposed to represent an
actual color that it should display to
the user but it still has to tell the
monitor it still has to tell to tell the
display that it actually wants to
present that color accurately and so
we're this is where the idea of color
profiling becomes really really
important is that your computer probably
understands given some software that is
properly able to understand color space
information and the embedded profile
provided by your buyer camera it knows
precisely what color it's supposed to
display but it may not actually know if
the monitor is correctly displaying that
color value and so what color profiling
does it excuse me is that it notifies
the computer when it sends a specific
signal to the monitor for a given color
it knows precisely what's that will
produce and so there's a wide variety of
techniques that you can actually use to
just to profile your monitor there are
some purely software based solutions
where you sit in front of your computer
and it displays some colors and it's
displaced some touch lines and it tries
to get you to change some sliders and
then you have just visually profiled
your monitor and that works it's perhaps
better than nothing but it's not really
that accurate because especially when
you're using a laptop or a
especially nowadays now that most of our
monitors are flatscreen they're not the
CRT style changing your angle to the
display actually slightly changes the
color that you perceive from from that
display so changing your viewing angle
or changing the amount of ambient light
will actually alter your perception of
the color that's being given by your
monitor so unless you're able to
reproduce that exactly every single time
doing it purely visually is not
necessarily going to give you the best
quality out of your profile so you can
use some hardware so I had mentioned
before that happen to use a I happen to
use like a hardware device that actually
allows you to profile your monitor
specifically unfortunately these devices
tend to be somewhat expensive they tend
to run a couple of hundred dollars in
order to get a good calibration system
one of them is the one that I used
before was an X right color display
color profiler and the this is actually
like a little device that you rest on
top of your monitor and then it has some
software that displays color on the
monitor and the device actually captures
the color and records precisely the
value that was output by your monitor
does that for a wide variety of colors
and then it knows it's able to tell the
computer through a profile which tells
your computer precisely what your your
your display is capable of it then the
computer is then able to finish the
change then able to take an image that
has some values and a given profile and
it will know the value of the color that
it's supposed to display at every single
pixel and then it will know how to
translate that into a signal that your
display can then show properly and so
this way is really the only way that you
can be sure that the colors that you're
looking at on your computer are
absolutely correct and once you've
started calibrating your displays you'll
actually start to notice that while many
displays are actually pretty good I mean
they're not soo
terrible they're actually not that great
you'll notice that in grayscale images
for example certain shades of grey will
look a little bit more green or a little
bit more red you'll notice that some
some of the contrast especially at the
very dark regions of images are very
bright regions of images you may just
you may you may not able to be to
differentiate between the darkest shades
or the lightest shades compared to a
properly calibrated monitor another such
things and these actually end up being
pretty important when you start working
on these images for print or for
actually displaying them in some sort of
professional or even somewhat
professional setting so calibrating your
monitor can be really really important
for this purpose but if you're just
gonna send your photos to your family
for example and you're just gonna
display them on the internet chances are
most people that view your images aren't
going to have a properly calibrated
display so maybe it's not quite as
important in that context and using
something like a software color
calibrator would actually be just fine
for your purposes getting at least
something out of it would perhaps be
better but maybe saving the money and
not getting a full-blown several hundred
dollar colored color calibration system
isn't necessary unless you really feel
like it's necessary for your particular
workflow any questions on this okay so
we have a color space that defines a
sort of a paint-by-numbers system where
we have zero values of 0 to 255 and the
color space defines what those values
should be what colors those values
should actually represent the computer
is able to interpret that and understand
it but it needs a monitor profile
it needs profiles to actually understand
I so the computer will say okay I want
to display this color but I need a
profile in order to actually properly
display it on whatever attached monitor
there is so there's multiple steps in
this chain that are actually somewhat
important when you're actually working
with photographs like this and it's also
important to realize that like you saw
with the
the color sync information from my
laptop for example it wasn't able to
actually reproduce all of the available
all of the possible colors for srgb and
that's especially true for color spaces
like Adobe RGB there's I don't know if
there's any reasonably priced monitor or
if there's even any monitors period that
can actually display the full color
gamut provided by Adobe RGB perhaps
there are but but it's actually somewhat
difficult to to achieve the reason that
you would want to use it though is that
you would want your your your
photographs of a landscape scene for
example they have a wide array of green
colors to actually display in as
accurate way of a possible as possible
on a print that might be one reason why
you actually want to use this so there's
this notion of needing to be able to
have color spaces and actually convert
between them perhaps you have an image
in a specific color space and you
actually want to convert to another
color space so perhaps you set to your
cameras color space through Adobe RGB
for example and by the way many cameras
actually do have the option of embedding
a specific color profile within the
image file most the time by default its
srgb but sometimes you can actually
change it to Adobe RGB if you want to
capture that additional color
information that's somewhat it that's
somewhat incorrect to say if you
actually want to more properly replicate
the possible green colors out of the out
of the scene you might want to switch it
to Adobe RGB but then if you decide
later on that you actually want to
display this image online that perhaps
it would be better to actually convert
this color space from Adobe RGB to srgb
so that you can then accurately show
some colors on the screen so there's
this idea of gamut mapping which does
precisely this and there's oh there's a
variety of techniques of gamut mapping
but for us as photographers we really
care about - and those are these the
perceptual gamut mapping and
colorimetric gamut mapping and for the
vast majority of cases when we're
dealing with photographs we
going to want to use perceptual because
this maintains this this maintains the
same perception or as much as possible
the same perception of gradients and
colors from one color space to the next
in other words if we have a large gamut
like we would and say in Adobe RGB color
space we want to convert it to a smaller
gamut like in say srgb we would want to
shoehorn all of the the gradients into
the possible gamut for srgb so this is
almost all the time
what you want in terms of of digital
photography but there's also
colorimetric where if it's actually
important that you preserve the color
values as they are you can essentially
throw away that extra information you
can think of it sort of like clipping
clipping that color information away and
preserving the portions of that gamut
that actually exists that actually
exists in both of those and so that's a
lot of the colors when we're talking
about srgb and Adobe RGB in particular
the red and the blue values will largely
be identical I believe that the that
technically that red and the blue are
exactly the same point and excuse me
srgb as they are an Adobe RGB but really
this is probably more of a scientific
endeavor if you really want to preserve
the raw true color information when
mapping from one gamut to the next now
why is this actually important we'll
realize that when we're dealing with
color management we are actually
displaying most of the time our images
in srgb and for the longest time for
many many years web browsers up until
probably honestly last year or maybe
this year I think now most of the major
web browsers actually have proper color
management support where when they're
looking at an image it will in fact take
into account the embedded color profile
for many years what we would see in in
just about every popular web browser you
name it it probably didn't support
proper color information so what this
meant was that if you had an image in
Adobe RGB or some other exotic color
space and you try to display it on your
web page the colors would actually just
look totally wrong because it wouldn't
her pit those numbers as some other
values the paint-by-number system would
fail because you you intended it for one
pallet and it was actually displayed in
another and there were a lot of really
interesting examples of this where
people would actually invert the color
space a little bit they would actually
define the color space as red being
green and green being blue and so on and
so you would actually get in these web
browsers that did not have proper color
color management you would see inverted
colors because that paint-by-numbers
system failed and it was totally
ignoring the mapping the color space
which defined what each value should
actually be but nowadays some Safari for
many years has been the winner in in
color color management if you were using
a Mac and using safari even since I
think 2008 or 2007 or something you were
you were safe it was actually able to
display the color information properly
chrome I think followed a few years
later but still it was like maybe 2011
or something like that
and maybe it was even Firefox but now I
think those major ones are doing it and
who knows what internet explorer is
doing I honestly have no idea if
Internet Explorer is properly showing
color information to you but what I'll
do during the break is in a little bit
I'll try to find some webpages that
actually display some color information
or displayed some images with kind of
wacky color values but with the color
space that would fix that information so
that it would actually look appropriate
in a properly color management
environment and we'll see we'll test
some browsers on my computer to see if
that actually works appropriately okay
now there's one thing that I do want to
show with regards to this this gamut
mapping we can understand perhaps a
little bit more about about photoshop's
color information so here let's say I
just want to show you let's see if this
will work now and nope let's see I need
to do
here it is photoshop's color settings
panel so if you've used Photoshop in the
past perhaps one of the scariest things
to look at was the the color settings
panel in Photoshop because it was kind
of if you didn't know much about color
spaces it was really difficult to get
through all of the all of the options
that are here and in fact a lot of these
options still don't really apply to us
but we can take a look at some of these
and try to and try to figure out what
some of these actually are so notice
that we have here this section in the
upper left called working spaces and
this basically defines the color spaces
that we want by default for when we're
working with with photographs in in
Adobe Photoshop and by default you
almost always want to use srgb because
again that is the most widely used color
space and the most likely one to
appropriately display your photos when
you're submitting them online or
submitting them to your friends but you
can of course change that to any one of
the other popular color space as well
like Adobe RGB or others now CMYK is one
you would really worry about only if you
were dealing with print but the standard
for that if you have a printer or a lab
that your that asks you to define a CMYK
color space they probably will tell you
precisely what profile or what color
space do what they would want to use so
the default for that is generally fine
then there was always this other scary
question that I would ask if you would
open an image for example and it would
give you some weird color space error
like you opened an image from the web
and it would say this image doesn't have
a color space what do you want to do is
very common to just kind of freak out
and just click cancel or something and
or click something to actually get the
image to open but now we can sort of
understand just a little bit better what
it was asking you to do so imagine that
you have an image and it doesn't
actually have an embedded color space
information within it well we can sort
of guess that means that it has all of
the numbers available in the in the in
display but that that's that connection
that mapping between what those numbers
mean to the actual color value is
not in fact they're now again most of
the time for images that do not have the
embedded color space we can sort of
assume that srgb is going to be the
correct one so we can apply perhaps an
sRGB color space to that image that
didn't have an embedded color space and
most of the time nine times out of ten
will be correct so you can sort of
assume that that's kind of gonna be your
default position now if for example you
have a file that was created and has an
embedded color space so when I say
embedded I mean this is literally some
information that's embedded within the
metadata so we talked about the exif
metadata and all of these other things
one of the things that can be embedded
within these images is also the color
space information that acts as this this
mapping from the numbers to the colors
but sometimes if there's if you have a
default color space for example srgb and
you're opening an image in some
alternative color space because it has
an Adobe RGB colour space embedded
within for example then Photoshop will
ask you what you want to do there and
generally what you want to do is to
preserve the embedded profile and the
reason for this is that if you were to
convert it to some other profile perhaps
that would then open an sRGB which is
again the working RGB is the the one
that we had set at the very top are
working spaces so it would perhaps
convert it to srgb but generally if I'm
opening a file at least in my case if
I'm if I'm opening a file that I've
downloaded off the internet for example
and it has an embedded color profile I
want to see that image in its intended
way and so I actually want to open it
with its preserved color profile so I'll
try it
so I'll tend to preserve the embedded
profile now the default for older
versions of Photoshop where all of these
things were checked missing profiles it
would ask you when it was opening if
there was profile investment in it was
opening profile mismatches is when you
have working space that was different
than the embedded color space or
similarly when you were pasting it these
would cause the sort of scary prompts
that most people tended to ignore and so
it seems like
Dobies finally caught on and unchecked
by default all of these all of these
things and you can still get that back
if you want it now that you understand
what's going on hopefully understand
better what's going on
you can get that that that feature back
by checking these asked when opening but
but usually it doesn't really seem to be
totally necessary now the rest of these
conversion options for example this in
this case relative colorimetric we were
just talking about relative colorimetric
versus perceptual in this case we might
want to select perceptual as our as our
default but it's pretty rare when I
actually do color space conversions and
the rests we tend to the rest we can
actually somewhat ignore will actually
talk about so there's the second option
here used either we'll talk about what
that actually means actually we can talk
about it now so the color space of
course defines specific colors that the
computer should display to you for a
given value but as you saw especially
with my monitor and with your own
monitor almost certainly your computer
is not actually able to display all of
the possible colors in a typical color
space so what a computer will do is will
and what some software will do like
Photoshop is it will try to trick the
monitor into actually displaying that
that color by using dithering which is
just the technique that a lot that
alters neighboring colors just slightly
so that they match they mix slightly
differently and it actually appears to
be a slightly different color this is
perhaps fine for our case because then
it improves our perception of that out
of gamut
color and it probably isn't going to
necessarily be bad but if you need
absolutely the most accuracy out of your
monitor setup then you probably want to
uncheck use dithering and the rest again
we can pretty much ignore for now any
questions on this
all right so we talked about color
spaces and how we can embed color space
information into images which is almost
always a good idea because if your web
browser if your user is happens to use
some software that it's actually able to
understand that color information that
embedded color space information within
your image then you're sure that it will
reproduce this accurately as possible on
their computer otherwise most likely
what will happen if you don't apply
color spaces that the computer will
assume srgb and try to display it in
srgb which might be fine normally but
perhaps not depending on your particular
circumstance and as a result of having
these different color spaces we might
actually want to map between the two so
we have this concept of gamut mapping
gamut is really just a fancy name for
this color space and we might want to
map from one to the next using a variety
of mapping techniques but for our
purposes perceptual or color metric are
the ones that we would probably use
perceptual or relative colorimetric or
probably the ones that we in fact would
want to that we would want to use so I
talked a bit already about monitor
profiling and color management but I did
want to mention a couple of other things
about this there's different types of
monitors of course so there's there's
the obvious kinds where there's the flat
screens the kind that we have in laptops
and TVs nowadays and then the the
old-school really big fat ones that were
huge on your desk it made your desk
slump or at least make it makes my desk
slump because it's not very good and
there is a wide variety of other ones as
well but even within these flat panels
even within the context of these flat
panels there's different technologies
and these different technologies have
different color reproduction
capabilities so there's the one that you
typically will find as the one that you
want is what's called an IPS display or
an s IPS display and this display is the
one that happens to you have the most
color of reproduction possible and
nowadays we're finding that the wide
variety of displays that we can purchase
are in fact IPs
displays they're all Mac's have had them
for many years a wide variety of pcs now
are also including IPS displays an
increasing frequency but about 10 years
ago or so it actually made a big
difference sometimes you would hear this
this about this display called a TN
display that was just terrible it was
have no contrast had no color and
fortunately we were moving away from all
of that but you generally want to find
if you're working with photographs in
IPS display even so the desktop panels
the panels that tend to be used in the
in the larger monitors that you would
actually have at home can actually be
higher quality and be higher and have
better color reproduction because of
their power requirements they because
they can actually be plugged into the
main the main power line then they can
actually use a little bit more power
they can be a little bit I don't know
less efficient is the correct term but
they can actually spend more power being
a little bit brighter displaying a
little some more colors so generally
what we find is that even so today our
desktop monitors tend to be higher
quality than our laptop monitors and
believe it or not there's actually a
number of monitors that cannot display
full eight bits worth of color
information so all of this color space
in from all this color space data aside
they really can only accept six bits
worth of information which is really
really terrible that's I mean that's
again not 256 that's eight bits that's
not 128 because that's seven bits but
we're talking about 74 distinct values
in the red the green and the blue so
they would do a lot of this dithering to
try to reproduce the color information I
think nowadays though fortunately in
2014 we're seeing fewer and fewer of
these but I bet that a number of laptop
displays remain seven bits or so not
able to actually display all of the
possible colors all right so with regard
to color management there's a couple of
other things that is important to keep
in mind one of them is this concept of
gamma and for our purposes we can really
think of gamma as being some measure of
contrast but it applies to your entire
monitor it's set by the system it's
by your color management system within
your operating system and it actually
defines how much contrast essentially
your computer has and this is an old
term from from back in the day when we
actually wanted to when we had those
tube type TVs and we needed to the way
that the way that they actually
functioned by shooting the electrons
under the phosphorous screen was not
quite accurate linearly so they had to
compress it with gamma there's there's a
bunch of history behind it
that we don't really need to care about
too much but the system still exists to
this day and I've mentioned it only as a
warning for those of you that happens to
be using older Mac's in particular many
older Mac's that have an operating
system of 10 5 or below Leopard or older
had a different value of gamma than the
newer Mac's and all of the versions of
Windows so the standard gamma value that
that I recommend you set your computer
to is 2.2 older Mac's had a much lower
value of 1.8 which means that things
look slightly less contrasty and this
did make a bigger difference a few years
ago when people were much more likely to
be running these old versions of Mac's
but it was extremely frustrating to get
the contrast in the saturation just
right on your image and then look at it
on another computer like a Windows
computer for example and all of a sudden
just looks overly contrasting overly
saturated because of this difference in
gamma so when you're actually color
doing a color profile on your monitor it
will generally ask you your target gamma
set it to gamma of 2.2 and that's
probably the default but just to be safe
said it's 2.2 and you should be good to
go it is worthwhile to note however that
some computers do come with slightly
higher gamma which means that you might
get your your contrast just write down
your machine but then you would look
somewhere else and it actually looks
slightly less contrasting than you would
want so as you're doing your color
profile set its 2.2 and you should be
good to go from there all right so if we
talk about when we're talking about
color it's also important that we talk
about oh yes
so generally um if you're doing
pre-press work
I believe the GAM is actually set
slightly lower at some some value like
1.8 which I think was the reasoning
behind old Macs having a gamma value of
1.8 it I think Galla in so the the
contrast that you are that your printer
will this that your printer will display
probably depends on a wide variety of
factors and may not necessarily have
some specific gamma value that we can
sort of distill easily in class but I
would wager that most printers
especially that you would buy nowadays
would actually come in such a way that
you would be able to print out something
and have it look as identical as
possible to a properly color managed
sequence of production on your on your
computer but I bet if you were to work
with a professional grade lab or some
other professional means of printing
producing some prints then you probably
would have to pay attention to this to
this a little bit more and in fact
that's that's sort of useful too to
think about because it might actually be
the case where when you print it that it
looks a little bit more contrasty in the
print then you might might have wanted
and sometimes setting your your monitors
gamma to a lower value like 1.8 might
try to mitigate some of those some of
those issues okay so the discussion
about color is not really complete
without of course color temperature and
it's it's really the idea behind color
temperature is that it really deals with
the temperature of some object so you've
heard the term glowing white hot glowing
red hot these sorts of things this is
actually a true physical phenomenon when
there is an object and it has a certain
amount of heat it emits some
electromagnetic radiation in the given
spectrum and so it actually does give
off some color things that are hot
enough to actually produce color in
their electromagnetic spectrum and so as
things get hotter and hotter and harder
they go from red all the way up to
and so on and so it's it's there's this
interesting dichotomy because in
culturally speaking we think of the
color blue as being a cool color and we
think of the color red as being a a warm
color but physically speaking that's not
necessarily true as you probably know if
you have a red flame it's not going to
be quite as hot as a blue flame blue
flame is going to actually be a much
hotter temperature and that is actually
represented by this idea of actual color
temperatures when we're talking about
color temperature we use the the the
measure of kelvins which is actually a
measure of temperature to determine the
distribution of color given by some
particular object so our Sun in the
middle of the day for example emits a
certain spectrum of light and our
atmosphere actually allows that certain
spectrum of light to to fall through and
so in the middle of the day we get we
see an apparent color temperature of
about 6500 Kelvin or so so give or take
depending on number of factors including
overcast and where you are on the earth
and altitude and so on and so forth but
about 6500 is pretty reasonable for mid
day but as we go inside for example for
an incandescent lighting or especially
if we're using candle lit lighting where
the color is much is the actual
temperature is much cooler that means
that it gives off redder light instead
and so the color temperature has
actually dropped the the actual
temperature itself has dropped
significantly producing more red tones
again remember this dichotomy that
there's red hots which is a cooler
object and blue hot which is actually
much hotter of an object so as soon the
sooner you're able to sort of let go of
those or distinguish between the the
cultural phenomenon of naming blue is
cool and red as as warm that the sooner
this will hopefully become less
confusing it's just unfortunate that we
have this actual dichotomy between these
these distinctions so realize that it's
really just the the whole like the whole
concept is really just this
this glowing object that's that actually
gives off a certain spectrum of color
and we are perceiving that that that
that color based on the wavelengths of
light that we are receiving and if we
were to actually apply this onto that
same graph we can actually see how color
temperature moves into the sort of blue
realm and so we have very cool colors
are really warm over here on the far
right then as we go left along this line
the temperatures are actually increasing
and so things become increasingly blue
now what this means is that we generally
want to try to capture a photograph in
of the the exact white in the exact
color temperature that might represent
that scene because remember when we're
dealing with color spaces it is going to
display white at this particular value
d65 and so in order to actually
represent that accurately we have to
have this white balance that actually
defines what values actually going to be
white so based on the lighting that's in
the room there's some there's some
spectrum of light that's actually being
output by our whatever our source of
illumination is and we want to try to
capture that using this color
temperature to define what should be
white and then be able to reproduce that
accurately on our computers so you can
generally set this manually on your on
your camera but that is really difficult
in changing circumstances if you're
outside for example with changing light
especially if there's any clouds that
move in or even just some even as the
Sun changes after a few minutes the
color temperature will change because
the atmosphere is is absorbing different
quantities of that spectrum of light
emitted by the Sun using it outdoors is
very difficult but if you're indoors and
you're using lights that have very known
quantities of emissions then you
generally will want to set your white
balance to something very specific that
actually matches the output of your
lamps and set it and forget it basically
just continue using it in a very
consistent
way now what this means though is that
for most of us we tend to take
photographs in in changing situations
unless you are doing a lot of work in a
studio this is it's probably going to be
the is probably going to be the case
that you will in fact need to change
your white balance after the fact and
what this means is fortunately when you
actually use a RAW format you can
actually do this you can actually go and
in post-production change what each of
the values actually represent in terms
of their their white balance information
when we're using a JPEG unfortunately
the camera decides automatically what
the color balance should be what the
white balance should be and we'll set
that in and solidify it in in the values
represented by the by the JPEG and so as
we've talked about before with
posterization and other things this
isn't necessarily the best thing for us
so generally what we want is there's a
variety of techniques to overcome this
in software typically you will have the
opportunity to use a little eyedropper
tool and try to pick a middle gray
generally there is going to be some
pixels in your shot that do represent
some sort of a middle gray or even a
white or like a black and you can use
that to represent that value and it will
be able to change the white balance
based on that selection sometimes though
it's not possible perhaps you're in an
extremely colorful situation and one of
the things that you might want to do is
to first take a picture of just some
white object and reflecting the light
that actually exists within the scene
and then you can later use that find the
color temperature from that and apply
that to the rest of your images as
appropriate okay let's take a quick
five-minute break and when we come back
we will continue talking about artifacts
all right hello everyone welcome back so
before the break we were talking about
color and color spaces and profiling
information and now is making the
argument that color profiling is really
important and it's especially important
if you want to have a consistent look of
color across your own computers across
your own devices you actually want to be
able to accurately
lay color information or your the color
and your photographs to your friends and
to your family and there's a couple of
ways that we can sort of hammer this
point home and one of them is to
actually talk about color in the context
of white balance and right now I'm
assume that for my purposes I'm using an
sRGB workflow and I have my monitor
properly profiled in this case we're
using a projector and it's totally not
profiled in fact it's probably very
inaccurate and but assuming that I have
all of those pieces in place I can now
look at the color that's being presented
to me from my from my software in this
case Lightroom which is hopefully
hopefully taking the color space into
account from my camera and actually
displaying the color information that I
want appropriately so this is a
photograph that I took last week while I
was having fun in Ireland while you guys
were were here watching the the guest
lecture I will admit that while the
guest lecture was happening I was not
actually having any fun at all I was in
middle of a conference and slogging
through it so this is really from the
weekend where rented a car and went out
into the countryside to sort of
appreciate the the the country a little
bit more and this is from the Cliffs of
Moher and unfortunately for my coworker
who was with me and my sister who came
up to visit me it was that time of day
where you know the light was starting to
get good but it wasn't quite there yet
and despite what it looks it was really
cold I mean the the wind was just
blasting across the Atlantic and it was
super windy and it was like really cold
and and my my sister and my coworker
were just they were just huddled and
shivering and I just said hold on just
one more picture one more picture for
probably about half an hour so there's a
few pictures here so there's this one
and it's not meant to be like oh look at
my fantastic photos but instead to show
you these two images and sort of look at
that the difference is between the two
there's some obvious ones in composition
but look at look at the context of these
in terms of the color so the light is
being reflected from the sunlight is
being reflected from from the greenery
here and we get a specific
look in the color of how that how that
image actually looked and I do want to
point out the white balance which we can
change here so this is the one of the
very first things I tend to do when I'm
looking at these photos is to adjust the
white balance make sure that it actually
looks good and in this case I think the
the white balance is actually pretty
appropriate you can kind of look at the
there's a variety of contexts to look at
here one of them is the the clouds in
the sky notice that the clouds do appear
to sort of be fluffy and white and gray
and sort of not have any sort of
unnatural cast to them unnatural blue or
unnatural warm cast to them and so that
the other color is probably correct as a
result but if it wasn't then I would
perhaps want to change my color my color
temperature to either be warmer by
increasing the the relative color
temperature or cooler by decreasing it
and we can sort of compare that to
another photograph which happened a
little bit later and notice that some of
the color is actually changing here but
really the this the color information or
the color that's that's changing is not
necessarily that on this sort of castle
but the the it's starting to become
warmer overall just as the Sun was
setting that the the light was in fact
becoming a bit warmer and the color
temperature did in fact go up slightly
but we can sort of see what happens if I
actually make those those changes
generally what I like especially with
sunset is to just increase the the color
temperature just a little bit to make
those colors pop just a little bit more
but perhaps that's that's a very that's
something that you will want to decide
for yourself and especially you don't
want to change it to such a degree that
aspects of the image that should be
white such as the breaking waves at the
bottom of the clouds at the top are no
longer truly white but that's just one
of the things to actually take a look at
when we are dealing with colors so we
have the color slider most of the time
what I probably use is this eyedropper
and I find some region of the image that
I know should be a proper white or a
proper gray and actually select that and
see how the color or the white balance
for that
image actually changes in this case it
actually increased it just a little bit
and this actually looks pretty good it's
nice amount of warmth in the color being
reflected from the cliffs and and the
color of the the clouds and the and the
breaking waves at the bottom to does
appear to be somewhat grains so there's
probably differences in play here with
between the the projector and the color
information that should be the color
profile for this projector that should
actually properly display this color but
it is what it is
yes yeah right I am when I'm editing
photos I'm constantly looking away and
then looking back especially if I'm for
the for the very same reason as these
illusions actually like you like you
were talking about just a moment ago
changing the contrast just staring at
the image for too long it's probably not
the same response as the illusion itself
the illusion is sort of would be
persistent if there was some aspect of
this image that was actually causing my
color perception to actually change for
that for that image that would be
persistent no matter of how much I can
like look at it and come back to it but
certainly there's an there's an element
of editing fatigue that that comes into
it and looking at an image for too long
I get an idea of how I want to actually
represent that image and after making
the modifications to it to what I think
it should be sometimes leaving and then
coming back actually we'll find that
I've overdone some element of it because
I had grown accustomed in some way and
again not necessarily in the same way as
these as these illusions but I've grown
accustomed to it in some way and needed
to need it to just take a little bit of
a break and come back to it with a fresh
start so this is perhaps more of like a
psychological sort of element than in
actual physiological components
right right oh yeah certainly I do
certainly find find it to be the case
where if I use the eyedropper on a part
of the image that I know should be gray
it will look wrong to me at first
because I had grown in fact accustoms
perhaps in some physiological way to the
the colors that were displayed before
and and there's there's other things
that play as well such as the ambient
light in in the room I think that can
have an even more drastic effect on my
perception of what the color balance
should actually be for this image I have
some lights behind my behind my computer
monitor at work and they're the kind
that can display all sorts of fancy
colors and usually I have something kind
of bright to keep me awake or something
and I noticed that if I edited photo in
that context and then change the color
it will just look so wrong it will just
be so completely off because of my
actual physiological reaction to the
ambient light that that was actually
really impacting my ability to edit the
the photo properly so yes I think that
using the eyedropper that there is there
is some element though that you might
click on an image on a portion of the
image that should be middle gray but you
might actually miss those pixels and
actually select a pixel that for some
reason or another due to camera error or
due to noise or something like that it's
not actually properly gray and might
actually be a red pixel for instance and
then it's actually totally off and then
that that's actually wrong and that can
happen in a more subtle way as well so
generally I will kind of select a
variety of areas just to be doubly sure
just to see if I'm able to consistently
get the same white balance and we can
even sort of see that now now that
you've been sort of staring at this
one image with this one white balance
for a little bit if I go back to the one
before I might actually look and so this
is a very small change we're talking
about a change of only 800 Kelvin or so
or 900 Kelvin or so but even now this
one sort of looks wrong to me now it
just looks too too cool it looks like
I've actually changed the color
temperature for the worse okay all right
so let's come back to the slides here
but let's switch gears a little bit and
we'd been talking about color
up until now which is hopefully you're
getting the sense that it's really
important to have a properly color
managed sequence of processing so that
you can actually properly display your
images oh I forgot during the break to
look up that one website that has the
cool like demonstrations of improperly
managed color browsers and maybe if I'm
if I'm able to find that well tweet it
or something but I did want to switch
gears a little bit to artifacts and
there's a wide array of artifacts that
might actually exist and there's there's
sort of a spoiler here in that we could
talk at length about all the different
types of artifacts and we could spend
some time talking about ways that we
could actually overcome them but I can
tell you right now that a lot of the
ways that we can overcome the artifacts
are is simple is that we'd use the
center part of the image generally the
outside the corners of the image are
going to have the worst representations
of artifacts and also that we stopped
down the aperture using the the lens at
its maximal aperture is actually going
to show the the most amount of artifacts
so those two things are really the way
to solve the problems that we are about
to see and unfortunately there's also
unfortunately that's not always an
option sometimes you actually have to
live with this and this is one of those
things where I'm going to show some of
these to you and you're going to
actually look at some of your photos and
you might it might ruin those those
photos not not ruin the photos but you
might start to realize some of the
issues with your camera but it's kind of
important to realize it because if you
don't notice it maybe someone else will
so the first one are spherical
aberrations and so
if lenses within the if lens elements
within the the lens as a whole are
actually made with a sir if you imagine
like a sphere and you sort of chop off
that portion of that sphere that is
would be a spherical lens element and
that's actually does not properly focus
light on the edges compared to the very
center so the Rays do not actually
converge at the same point at the outer
edges of those compared to this compared
to a spherical lenses which have a shape
that tried to overcome this issue so a
spherical lenses are not spherical and
they have they have more of a tapering
sort of at the edges of the lens to try
to overcome this issue of focusing the
Rays too strongly at the outside at the
outside edges of these so typically what
will results so if you imagine this
concept of the circle of confusion if it
sort of matches this sort of narrow band
here where you can imagine that we will
have some areas of it that actually are
out of focus it'll actually appear as
though there's some some fuzzy areas in
your in your in your photograph now
similarly imagine that we have some
light that comes not straight on but off
axis so it's not actually on the same
axis as the lens itself is is pointing
we can get this interesting results
called a coma which is sort of like the
word for comet because that's what it
actually appears like so the image which
we can see here in the very middle
portion is that we might have a point
source of light that actually sort of
spreads out and it looks like it has a
tail like a wide diffuse tail as a
result of some imperfection in the lens
you can sort of trace the rays of light
as they actually enter this lens and are
from the outside and compared to the
innermost parts of this lens and you can
sort of see why this might actually be
happening and there's a couple of
different there's a couple of different
diagrams here to try to show you why
this might be happening but what this
would actually look like in your
image is this so you have an image this
there's this is this center portion up
along the top here is a single image and
we are cropping three distinct portions
of that image from the very center and
also at the very outside and we can see
how these point sources of light
actually look different in each of these
areas so in this case we have the very
center a point source of light that's
coming on axis it's coming head-on so
the light rays are actually converging
properly and it looks correct it look
actually looks like a single point
source of light but as soon as we get to
these objects that are off to the side
and they're coming at us slightly at an
angle to the to the axis and not head-on
then we start getting some of these
issues and we can see this sort of
smearing this coma that can result as as
a results and so again the outside of
our image is the the problem here and
also the fact that we might be able to
fix this by stopping down a couple of
stops on our aperture by making our
aperture smaller but that's how we would
potentially fix this but again it's not
always necessarily possible now for
those of you that wear glasses or
contacts you may be familiar with the
the concept of astigmatism which also
applies to lenses as well and in this
case if you are a stigmatic like I am
things that should be that should have
some particular shape are in fact
distorted slightly but in such a way
that it's not really like kind of it's
not really like super obvious it's it's
it will stretch the circle for example
to make it a little bit more of an
ellipse it will distort to square to
make it slightly less square and more
rectangular those sorts of things and
this can can cause some interesting
out-of-focus areas as well we might get
some sort of weird diffraction pattern
in our out-of-focus areas which would
impact your bokeh and a variety of other
patterns that we can see here but
usually we would notice this mostly in
the out-of-focus areas and not quite so
much in the actual geometry of the
shapes that are within our scene but
this is just something that that does
happen as well
again this isn't
really something that I would worry too
much about because normally it's not
quite so bad and it's obvious is this
but this is one of the reasons why you
may not have a perfectly smooth bokeh in
your out-of-focus areas now vignettes is
something that has come back into vogue
a little bit in recent years thanks to
things like Instagram where people
intentionally try to add some darkening
to their their corners this is an
extreme example of that where we have
I've been yet that's actually blocking
out the light from the corners of this
image but it's I mentioned this because
there's actually a variety of types have
been yet to watch out for there's
optical there's mechanical there's also
natural vignette and sometimes it's
important to understand the distinction
of these because some of these you can
actually correct and some of them you
cannot so let's first look at at a
optical vignette in which in this case
is due to a lens that has a really wide
aperture and lights that comes in
through the side is not actually going
to be able to is going to be physically
blocked by the actual side of the lens
so imagine we have this lens here you
can sort of see what happens if we are
looking through the lens just off axis
very slightly no longer do we see behind
the lens in a circular way as we
normally would but we see this sort of
cat I like formation and again stopping
down would actually fix this issue but
we actually saw an example of this from
one of my own images a few weeks ago
where out of focus areas in this image
actually have this optical vignette as a
result of this wide aperture that I was
using with this lens and the the light
coming off axis as a result of that so
as you might recall when I sort of
discovered this live that I was sort of
surprised to find it because I'd never
noticed that before and it kind of
annoyed me that's that this actually
happens but it is what it is I guess
there's I would have to get a different
lens to actually be able to overcome
this of course stopping down might be
possibility but in this case it simply
was not I had to use the widest possible
aperture in order to actually get this
image because of how dark it was in the
fact that I was using it that at full
zoom this was at 200 millimetres 35
millimeter equivalent and using it
handheld on a boat that was moving so I
had to make sure that I tried to get the
maximum possible shutter speed now we
might get a natural vignette which is
actually which despite what it looks
like is not really due to light fall-off
but is in fact due to micro lenses so
imagine that we have a lens that has a
very small focal length in which case
the light comes at very strong oblique
angles particularly to the sides of the
sensor what can happen here is that the
micro lenses may not be able to properly
properly focus the light that's coming
from this lens into the pixel and so we
might get some sort of some sort of
natural vignette as a results just in
the corners of this image especially in
wide-angle lenses you notice this
especially in wide-angle lenses with
longer lenses the the light tends to
come in and a lot more parallel angles
and so it sort of reduces the
possibility of this but as you saw
longer lenses might have some other
issues with optical optical vignetting
alright so there's some other ones as
well probably if probably the one that
you most likely have noticed in your own
photographs are barrel or pincushion
distortions so imagine that if you take
your photograph print it out and wrap it
around a barrel that's the kind of
distortion that you would get and barrel
distortion it will make lines sort of
bulge out towards you pincushion
distortion is sort of the opposite
imagine that you put it on a pillow you
put your photograph on a pillow and you
push in at the center and all the lines
sort of converge in words that would be
pincushion distortion and this becomes
very very noticeable typically what you
find is that wide-angle lenses have
barrel distortion telephoto have
pincushion normal lenses can sometimes
get away with this with with these
particular types of distortions you can
correct them easily and software
fortunately especially software that
knows your lot your lens and your
software or your lens and camera
combination but here are some pretty bad
examples on the Left we have some barrel
distortion this this window notice in
particular at the very bottom seems to
be bowing outward like it's wrapped
around a barrel and the the the building
here although at first glance looks
pretty good you will notice that it
actually kind of gets squeezed in the
middle due to some pincushion distortion
this is really found this is really
really prevalent in consumer grade
cameras the ones with fixed lenses
fortunately what we're seeing in
technology these days is that the lenses
and the cameras will actually
automatically correct for distortion
with no and lenses connected to them so
we're actually seeing a reduction in
some of the optical distortions such as
such as this that come out of our our
cameras as a result but still the the
lenses probably still have these these
issues optically even though we may not
see them digitally so there's another
one as well which is when we think about
the light the wave property of light we
can actually get artifacts due to
refraction which it's very similar to
the way that we would have a prism like
this which actually takes white light
and separates the the light because the
light is actually changing the different
wavelengths are changing speeds within
the material after we have a change in
material from light from air to glass
the refractive index changes and so it
sort of splits it all off just a similar
way that that a prism actually works the
light is actually changing speed within
this medium and then once it comes back
out to the other side
it sort of resumed but not without some
error in the angle and this actually
impacts lenses as well and one of the
things we've just really been totally
ignoring this but this can cause some
chromatic aberrations within our
photographs imagine that we have a
simple lens like on the left we will
actually see that the blue light focuses
first then the green light and then the
red light of course it's a spectrum but
what this would actually look like is
some blurring within your image and this
is especially bad on the outsides of the
sensors where blue blue sort of focuses
furthest out due to these due to these
refractive problems and without
correction this can actually cause some
pretty bad problems we can actually see
this in an exaggerated version here
where we have this sort of blurriness
that results but really you can see that
it's like the lines have this spectrum
of color because of the the light that
was split so if you sort of imagine this
in the context of a black-and-white
image this would just look like pure
blur it would look perhaps out-of-focus
or it would look like there was some
issue there and this was one of them the
many reasons why it's really difficult
to get a really high quality
black-and-white image is that any one of
these sort of aberrations can really
just come across as blur in your
black-and-white image whereas in the
color image we can sort of distinguish
between some of these aberrations now
some more aberrations are of course
purple fringing which is really
prevalent on some combinations of
cameras and lenses and this might be the
case that it is actually a poor lens
it's actually not refracting the light
properly but normally the the what we
would see in digital cameras especially
is that the camera would have Micro
lenses that was really only optimized
for green wavelengths and so the red and
the blue wavelengths would be improperly
focused and we would get this
combination we'd get this like purple
fringing that would result because of
that there's of course ways to pick
this stopping down will sometimes help
because then that means that color won't
come in at such an oblique angle but
also the lens itself if it uses some
material like fluorite which really sort
of reduces the the problems of the
diffraction then you might actually have
this or refraction excuse me so there's
other other things as well so we have a
diffraction width so I was using before
the before we were talking about
refraction but diffraction is is
basically when something behaves as a
way this can be water this can be light
when it works its way around an
obstruction we actually gets these
patterns of constructive and destructive
interference where we can actually see
that even though we have waves that are
coming at this angle from left to right
as soon as it hits this instruction here
as the waves work with their way around
that obstruction there are waves that
continue along but we also get these
phases of destructive and constructive
interference where that wave then
propagates incorrectly after that
obstruction and we can actually so one
way that this would actually appear is
that if we have a point source of light
we shine it through a really small hole
we might actually get some diffraction
we wouldn't get a perfect representation
of that but in fact we would get this
sort of Airy disc concentric circle
appearance and this if it exceeds of
course the the circle of confusion would
result in some blurriness in our image
so just sort of show you one issue here
is that as you make your aperture
smaller and smaller and smaller that
it's actually causing a small
obstruction in this way and once you get
close enough to some to some multiple of
wavelength of light you will actually
get to this issue where you'll get
diffraction that causes all sorts of
softness so even though we have an image
that is technically in focus this we
have some diffraction issues that are
softening all of the light that's
actually
coming into the coming into the camera
and so given this very small pixels that
we have in cameras these days you really
will find that even modest increases in
apertures such as f11 or f-16 will cause
diffraction softness to do diffraction
especially on much smaller cameras so
this has been a lot of just sort of a
whirlwind tour of optical aberrations
that can actually happen and impact the
image quality of our camera but there's
also some digital artifacts that can
actually result we've actually gone over
some already we've talked about
posterization what that actually is and
what it means you might recall that that
was we had that photo of the sunset we
increased the contrast to such a degree
that no longer was there a smooth
gradient from blue to black but instead
it was steps from blue to cyan to black
and that was an extreme example of
posterization we've also talked about
noise at length we're increasing ISO
which is not quite the same as actually
changing the sensitivity of a sensor
will change the the noise that comes out
of that that comes out of that sensor
and also some blooming which can happen
when photons actually overwhelm
individual pixels and spill over to
neighboring pixels and we see not just a
small area that's overexposed but a much
larger area than we would actually
expect there are some other digital
artifacts as well we've talked about
some of these as well like aliasing and
anti-aliasing where we can actually
change the we would actually get
smoother edges but we actually do get
some loss of resolution in the context
for this of course was the anti-alias
filter the low-pass filter that exists
on the top of many bayer color filter
array cameras the sensors of those
cameras and there was myra as well
there's just a couple more examples of
that where we take this and this you
might actually find this outside of the
context of just simply taking a
photograph but if you were to have an
image such as this which has a somewhat
high frequency repeating pattern in this
case the alternating red and white
between the brick and the mortar and you
change this
size of this this image you actually
resize it to be smaller it's actually
very difficult for the computer to
represent that same pattern if some of
those pixels drop below one pixel if it
should actually become smaller than one
pixel and if it can't actually replicate
that that that same pattern in the
smaller size very well you'll get this
sort of strange repeating artifact this
Moyra artifact because some pixels will
be essentially rounded up and some
pixels will essentially be rounded down
now the reason that you would see this
is if repetitive detail sort of exceeds
the resolution of the sensor or of the
image that we're actually creating which
in this case the repetitive detail does
once we reach ange the size of this
image to be smaller then we also talked
about moira and maize patterns where we
can get wire up but also in combination
with some strange some strange maize
patterns as a result of this the D
mosaicing algorithms that occur when we
have a color filter away and were
reconstructing that image one of the
things that we really didn't get a
chance to provide many examples of
however was we're sharpening halos when
you take an image and you're trying to
combat this issue where you have a sum
once so sure it might the focus might be
ideal and you might have a somewhat
sharp image you want to reduce the
softness even more you want to make that
image even sharper to try to counteract
your anti-aliasing filter that probably
exists in your camera you might actually
be tempted to use a sharpening tool that
Lightroom or Photoshop or some other
software actually has but if you the way
that these tend to work is that it finds
areas of contrasts such as the interface
between this circle and the background
and will try to increase the contrast
increase the contrast between the two of
them by taking the darker edge and
making it slightly darker and taking the
brighter edge and making it slightly
brighter and if you go too far you'll
actually get this halo appearance
because you the inside is now too dark
on the outside is now
right so there is there is definitely a
good middle ground between when we have
too soft
and when we apply some sharpening so
that we actually do get some nice some
nice amount of definition back and when
we overdo it and in fact increase the
sharpening to such a degree that we get
some halos so if you are in fact going
to apply some sharpening to your image
take a look around at a variety of
places not just one place but zoom in to
a hundred percent and make sure that you
haven't overdone it and gotten some
strange halos such as this you can
really help out your your a case by
using Brahe of course that will help to
sort of fix these also what you will
find is that if you use JPEG by default
in the camera the camera will
automatically apply some sharpening and
some cameras will go so aggressive that
they will in fact give you for free some
halos and that's probably not what you
want so be mindful of that to use raw
when possible and then you can define
for yourself exactly how much sharpening
you want out of your image and when
we're talking about compression we
talked about we at the very beginning of
the semester we were talking about the
lossy compression of JPEG versus
lossless compression of other file
formats such as ping or PSD and really
we can get some some compression
artifacts out of that if we of course
use quality that is a quality measure
that is just far too low so the
compression will discard some color
information it will discard fine detail
based on some measure of our perception
of how we would actually perceive that
so when you're working on images what I
would what I would recommend is try to
have a non-destructive workflow trying
to use try to take photos in RAW use
some software like like Lightroom or
some other RAW processing software that
allows you to apply changes but doesn't
actually finalize those changes until
you export into a lossy format like JPEG
that way no matter how many times you
open that file and and change it and
then we save your your changes to that
file you can then receive
the highest possible quality out of that
image if you were to take an image a
JPEG image for example out of your
camera make some modifications to it
save it go back make some more
modifications to it save it again the
the lossy aspect of JPEG will in fact
change this this image over time now
there is sort of a little asterisk here
if you just open a JPEG file and save it
and save it and say just keep opening it
without actually doing any other
modifications most JPEGs saving
algorithms won't actually change the
image for the worst so you should
actually be able to do this over and
over again but some will so just be
mindful of that and again you can get
around all of these issues by just using
wrong using a non-destructive editing
format so with all of these hopefully
you'll be able to be aware of some of
the artifacts and some of the problems
that you can have hopefully you're a
little bit more knowledgeable about
color information why it's actually
important to profile your your monitor
and hopefully that will be a good
starting place as you work on your final
projects which of course are due next
week and until next week well thank you
all very much for coming and enjoy your
the rest of the week and we'll see you
then