1
00:00:00,000 --> 00:00:10,647

2
00:00:10,647 --> 00:00:11,980
DAN ARMENDARIZ: Hello, everyone.

3
00:00:11,980 --> 00:00:16,590
I'm Dan Armendariz, preceptor
in computer science for [? Cs ?]

4
00:00:16,590 --> 00:00:19,890
and today I'm going to be talking
to you about digital photography.

5
00:00:19,890 --> 00:00:24,030
Now, in particular we're going to
do a crash course in just 60 minutes

6
00:00:24,030 --> 00:00:26,701
on a number of topics
in digital photography.

7
00:00:26,701 --> 00:00:28,450
Unfortunately, we have
a packed house here

8
00:00:28,450 --> 00:00:31,070
to sort of like choose
your own adventure,

9
00:00:31,070 --> 00:00:35,290
and we will try to get
through as much as possible.

10
00:00:35,290 --> 00:00:38,600
>> So without further
delay-- unless you happen

11
00:00:38,600 --> 00:00:42,890
to be hiding under a rock--
humanity has for the very first time

12
00:00:42,890 --> 00:00:46,960
put a lander on a comet,
which is a pretty cool thing.

13
00:00:46,960 --> 00:00:50,640
Phi-lay or Phil-y or some
way of actually pronouncing

14
00:00:50,640 --> 00:00:52,890
this-- I've heard it
pronounced a variety of ways,

15
00:00:52,890 --> 00:00:58,320
but of course this lander
and the associated satellite

16
00:00:58,320 --> 00:01:00,470
that actually brought the
lender to the comet each

17
00:01:00,470 --> 00:01:04,069
have some digital cameras
attached and associated with them.

18
00:01:04,069 --> 00:01:10,130
So this is the view of Philae from
Rosetta's OSIRIS narrow angle camera,

19
00:01:10,130 --> 00:01:14,590
so Rosetta is the machine that actually
brought Philae over to the comet.

20
00:01:14,590 --> 00:01:18,250
>> Philae is the lander itself and as
Philae was its way landing on a comet,

21
00:01:18,250 --> 00:01:19,249
it snapped some photos.

22
00:01:19,249 --> 00:01:22,290
And so there's something interesting
about this that I want to point out,

23
00:01:22,290 --> 00:01:25,320
and first of all, this
is just the lander,

24
00:01:25,320 --> 00:01:29,990
of course, but if you notice surrounding
that there seems to be no stars.

25
00:01:29,990 --> 00:01:33,780
So I added a little bit extra black
just sort of design of the slide,

26
00:01:33,780 --> 00:01:36,050
but the very center, the
very corner of this slide

27
00:01:36,050 --> 00:01:41,414
is in fact original, the original image
that came from Rosetta's OSIRIS camera.

28
00:01:41,414 --> 00:01:43,330
So just sort of give
that some consideration--

29
00:01:43,330 --> 00:01:46,250
why, if this is in fact
in deep space, is it

30
00:01:46,250 --> 00:01:50,010
the case that there are no
stars in this photograph.

31
00:01:50,010 --> 00:01:52,920
>> So just a couple of other
things to take a look at-- this

32
00:01:52,920 --> 00:01:58,160
was a photo that came back from
Philae, this was yesterday I think,

33
00:01:58,160 --> 00:01:59,620
after it had actually landed.

34
00:01:59,620 --> 00:02:02,910
And unfortunately, it was the case
where the very first that Philae landed

35
00:02:02,910 --> 00:02:06,020
it bounced a couple times, and so
it's not actually the proper position

36
00:02:06,020 --> 00:02:08,270
that they expected, but
still it has this kind

37
00:02:08,270 --> 00:02:10,919
of neat look of the comet itself.

38
00:02:10,919 --> 00:02:14,010
And one of things that's really neat
about this is that you realize that

39
00:02:14,010 --> 00:02:16,690
Rosetta has been travelling for
about 10 years through space,

40
00:02:16,690 --> 00:02:20,480
so this means that the digital camera
technology that's contained within

41
00:02:20,480 --> 00:02:23,360
Philae and Rosetta is
least 10 years old,

42
00:02:23,360 --> 00:02:26,450
but if you go back through the records
there's actually a scientific paper

43
00:02:26,450 --> 00:02:31,120
that was published back in 1998
that talked about the specifics

44
00:02:31,120 --> 00:02:36,290
of the specifications of the
cameras on each of these satellites.

45
00:02:36,290 --> 00:02:39,360
>> And this is 1988,
that's a long time ago.

46
00:02:39,360 --> 00:02:42,000
Do you have any idea what kind
of digital camera technology

47
00:02:42,000 --> 00:02:43,370
was available back then?

48
00:02:43,370 --> 00:02:48,700
There happen to be a digital
camera called the Canon EOS d2000

49
00:02:48,700 --> 00:02:51,160
and it was really the
first digital camera

50
00:02:51,160 --> 00:02:55,980
that came out that people considered to
be serious and usable digital cameras,

51
00:02:55,980 --> 00:02:58,410
so was it the case
that back in 1998 when

52
00:02:58,410 --> 00:03:01,270
there were creating the
specifications that they simply

53
00:03:01,270 --> 00:03:05,320
duct taped one of these Canon
EOS d2000s to this lander?

54
00:03:05,320 --> 00:03:06,780
Well, of course not.

55
00:03:06,780 --> 00:03:08,720
>> This is meant to be a
scientific instrument

56
00:03:08,720 --> 00:03:11,920
and so there's a lot of detail
that actually went into this,

57
00:03:11,920 --> 00:03:16,560
but just to give you some context,
this top of the line d2000 camera

58
00:03:16,560 --> 00:03:22,280
had two megapixel sensor and could take
photos at about 3.5 frames per second.

59
00:03:22,280 --> 00:03:24,230
So two megapixels is
pretty abysmal, if you

60
00:03:24,230 --> 00:03:29,170
have a modern smartphone such as
an iPhone or Android phone it might

61
00:03:29,170 --> 00:03:31,700
be that the camera on
the front of your device

62
00:03:31,700 --> 00:03:35,230
actually has one or two megapixels,
about the same number of pixels

63
00:03:35,230 --> 00:03:39,960
as the Rosetta camera itself--
that's sort of the high quality one.

64
00:03:39,960 --> 00:03:44,680
The Philae lander
actually has other cameras

65
00:03:44,680 --> 00:03:46,380
that are only one megapixels each.

66
00:03:46,380 --> 00:03:48,580
I think there's an array
of six for panoramas

67
00:03:48,580 --> 00:03:51,580
and then there's another
for some scientific studies

68
00:03:51,580 --> 00:03:54,060
and so basically the photo
that we were just looking at

69
00:03:54,060 --> 00:03:57,570
was taken essentially with
a one megapixel camera.

70
00:03:57,570 --> 00:04:01,090
>> Now of course, this is kind
of not a very fair comparison,

71
00:04:01,090 --> 00:04:04,130
because when we're talking
about the scientific aspect

72
00:04:04,130 --> 00:04:09,662
of digital photography then there's
a lot to of additional work that

73
00:04:09,662 --> 00:04:12,370
has to go into making sure that
it's actually going to be correct

74
00:04:12,370 --> 00:04:16,170
and that they can actually get
some usable data out of this.

75
00:04:16,170 --> 00:04:20,119
And there's some interesting
things about the Rosetta camera

76
00:04:20,119 --> 00:04:23,160
that we can actually learn from the
paper that was published back in '98.

77
00:04:23,160 --> 00:04:26,550
In particular, it had a four megapixel
camera, which was pretty impressive.

78
00:04:26,550 --> 00:04:28,724
It actually had a very
large sensor size--

79
00:04:28,724 --> 00:04:30,140
we'll talk more about sensor size.

80
00:04:30,140 --> 00:04:34,254
That was pretty well equivalent
to a standard 35 millimeter frame.

81
00:04:34,254 --> 00:04:36,670
We'll talk more about that in
just a little bit, hopefully

82
00:04:36,670 --> 00:04:38,770
if we actually get to it.

83
00:04:38,770 --> 00:04:40,880
>> And the maximum shutter
speed, so in other words,

84
00:04:40,880 --> 00:04:45,300
the maximum amount of time that, rather
than the fastest amount of time that

85
00:04:45,300 --> 00:04:49,540
the sensor was actually able to
capture data and to capture the lights

86
00:04:49,540 --> 00:04:51,990
for the exposure was
one 1/100 of a second,

87
00:04:51,990 --> 00:04:56,210
which is frankly pretty abysmal compared
to this digital camera that actually

88
00:04:56,210 --> 00:05:01,820
that came out in 1998, which operated
about 1/4,000 or maybe 1/8,000

89
00:05:01,820 --> 00:05:03,740
of a second.

90
00:05:03,740 --> 00:05:05,850
So let's take a look at
another image from space.

91
00:05:05,850 --> 00:05:09,820
>> This came out of JAXA, which
is Japan's space agency

92
00:05:09,820 --> 00:05:15,075
and this is a picture of they released
a satellite that went around the moon

93
00:05:15,075 --> 00:05:18,630
and took some photographs, and
this was I believe a moon rise that

94
00:05:18,630 --> 00:05:21,250
came over that, and
it's a fantastic image,

95
00:05:21,250 --> 00:05:23,410
but again you have to
wonder what is going on.

96
00:05:23,410 --> 00:05:26,496
Why are there no stars in this scene?

97
00:05:26,496 --> 00:05:29,120
So realize that we we're talking
about digital photography, one

98
00:05:29,120 --> 00:05:33,230
of the most important aspects of
it is to consider the exposure.

99
00:05:33,230 --> 00:05:36,030
And of course, exposure is
not something that we actually

100
00:05:36,030 --> 00:05:38,150
deal with solely in
digital photography, this

101
00:05:38,150 --> 00:05:40,970
applies to film photography
as well and also videography

102
00:05:40,970 --> 00:05:44,650
and a variety of other fields where
we're actually capturing images,

103
00:05:44,650 --> 00:05:48,810
but there's really four major
things that impact the exposure.

104
00:05:48,810 --> 00:05:51,940
>> One of the most important things
is the amount of available light.

105
00:05:51,940 --> 00:05:54,366
Now sometimes you can control
this, if you're in a studio,

106
00:05:54,366 --> 00:05:56,990
for example, or in this room we
can control the amount of light

107
00:05:56,990 --> 00:05:59,200
by turning some lights on,
turning the lights off,

108
00:05:59,200 --> 00:06:02,040
but in the case of the
satellites they really

109
00:06:02,040 --> 00:06:05,460
don't have any control over this.

110
00:06:05,460 --> 00:06:09,520
It is the amount of sunlight
that exist in the sky

111
00:06:09,520 --> 00:06:13,470
or rather in space that reflects
off of each of these objects

112
00:06:13,470 --> 00:06:16,560
and can be collected by this sensor.

113
00:06:16,560 --> 00:06:18,560
So the amount available
light, we may or may not

114
00:06:18,560 --> 00:06:21,230
have control over depending
on the circumstance,

115
00:06:21,230 --> 00:06:24,100
but notice that we also
have three other settings

116
00:06:24,100 --> 00:06:28,870
as well-- shutter speed, ISO, an
aperture through which any camera

117
00:06:28,870 --> 00:06:33,690
actually uses to manipulate to try to
capture the amount of available light

118
00:06:33,690 --> 00:06:35,110
that exist in the environment.

119
00:06:35,110 --> 00:06:37,100
So another way to think
about this is that you

120
00:06:37,100 --> 00:06:40,690
have a sensor in a digital camera, it
can collect a certain amount of light,

121
00:06:40,690 --> 00:06:43,990
there's a range of amount of light
that it can actually collect,

122
00:06:43,990 --> 00:06:47,240
too little light and it won't
register, so it'll look totally dark.

123
00:06:47,240 --> 00:06:50,280
Too much light and it will
actually overwhelm the sensor

124
00:06:50,280 --> 00:06:51,890
and it will look totally white.

125
00:06:51,890 --> 00:06:54,810
So we have these settings
to try to compensate

126
00:06:54,810 --> 00:06:57,560
for the amount available
light that exists in the scene

127
00:06:57,560 --> 00:07:00,860
and fit that amount of light
in the scene to the range

128
00:07:00,860 --> 00:07:04,000
that our sensor can actually capture.

129
00:07:04,000 --> 00:07:07,610
>> So let's take a step back and
talk a little bit about light.

130
00:07:07,610 --> 00:07:10,300
So you might recall from
high school physics,

131
00:07:10,300 --> 00:07:17,780
light is of course is photons that has
properties of both the wave and matter,

132
00:07:17,780 --> 00:07:24,090
and because of its
properties of a wave it

133
00:07:24,090 --> 00:07:27,240
operates in various wavelengths
and we as humans can only

134
00:07:27,240 --> 00:07:30,430
interpret and understand
and receive through our eyes

135
00:07:30,430 --> 00:07:34,420
a small spectrum of the
electromagnetic spectrum, which

136
00:07:34,420 --> 00:07:37,540
represents the color
that we're able to see.

137
00:07:37,540 --> 00:07:41,510
Now, it's interesting to note
of course that our visual system

138
00:07:41,510 --> 00:07:45,460
is a rather complex system that is made
up of a wide variety of parts, not only

139
00:07:45,460 --> 00:07:49,180
just our eyes, but even all of
the sub parts within the eyes,

140
00:07:49,180 --> 00:07:51,566
including the lens,
the iris and the retina

141
00:07:51,566 --> 00:07:53,940
in the very back with all the
cells associated with that,

142
00:07:53,940 --> 00:07:57,350
but also the pathway to the brain
and the visual cortex itself.

143
00:07:57,350 --> 00:08:00,420
>> And this can lead to some very
interesting phenomenon that actually

144
00:08:00,420 --> 00:08:03,610
impact us as photographers,
and perhaps more

145
00:08:03,610 --> 00:08:07,660
correctly impact the design of
cameras and digital cameras.

146
00:08:07,660 --> 00:08:09,692
So this you may or may
not have seen if you've

147
00:08:09,692 --> 00:08:11,900
been trolling around on the
internet for long enough.

148
00:08:11,900 --> 00:08:15,540
It's just an optical
illusion where there

149
00:08:15,540 --> 00:08:20,300
are two tiles that are labeled-- tile A
at the top of this illusion and tile B

150
00:08:20,300 --> 00:08:22,540
at the center, and it
just so happens that they

151
00:08:22,540 --> 00:08:24,638
are in fact exactly the same color.

152
00:08:24,638 --> 00:08:26,513
So even if you know this
fact, you look at it

153
00:08:26,513 --> 00:08:28,096
and it still doesn't look quite right.

154
00:08:28,096 --> 00:08:30,690
This is in fact a very
strong visual perception

155
00:08:30,690 --> 00:08:34,700
that our brain is playing on us.

156
00:08:34,700 --> 00:08:37,789
Just to try to prove
this to you a little bit,

157
00:08:37,789 --> 00:08:40,600
>> I'm going to bring up the
same image in Photoshop

158
00:08:40,600 --> 00:08:46,090
and I'm going to bring up the eyedropper
tool, select the color in the A tile,

159
00:08:46,090 --> 00:08:50,400
and I'm going to draw a little
color bridge between A and B

160
00:08:50,400 --> 00:08:54,170
and hopefully now you can
sort of see what is going on,

161
00:08:54,170 --> 00:08:57,110
or you can at least convince
yourself that this color is

162
00:08:57,110 --> 00:08:59,920
in fact the same in these two tiles.

163
00:08:59,920 --> 00:09:03,470
So let me digress a little bit, because
I really am showing you this just

164
00:09:03,470 --> 00:09:09,990
to make clear the fact that we have a
visual system that complicates matters.

165
00:09:09,990 --> 00:09:14,560
Our eyes do not operate scientifically
like the Philae lander would

166
00:09:14,560 --> 00:09:16,420
and like a digital
camera would, and this

167
00:09:16,420 --> 00:09:20,181
causes some problems that actually
impact us as digital photographers.

168
00:09:20,181 --> 00:09:22,180
So if we take a look at
the structure of the eye

169
00:09:22,180 --> 00:09:24,310
we don't have to really
worry about too much of it,

170
00:09:24,310 --> 00:09:29,070
but there is of course the iris
and the lens that actually focuses

171
00:09:29,070 --> 00:09:32,610
the light into the back of
the eye, which has the retina.

172
00:09:32,610 --> 00:09:36,922
The retina has a variety of cells,
and in the very center of our vision

173
00:09:36,922 --> 00:09:38,880
there exists a structure
called the fovea where

174
00:09:38,880 --> 00:09:41,590
we have a very high concentration
of detail cells that

175
00:09:41,590 --> 00:09:46,020
allow us to see color vision
and a variety of other things.

176
00:09:46,020 --> 00:09:49,425
Now the retina is made up of
a variety of types of cells.

177
00:09:49,425 --> 00:09:51,800
There's two major types that
we're really concerned with.

178
00:09:51,800 --> 00:09:54,430
There's rods and cones,
and each of these

179
00:09:54,430 --> 00:09:56,590
have different properties,
so the rods for example

180
00:09:56,590 --> 00:09:58,500
are primarily associated
with night vision,

181
00:09:58,500 --> 00:10:00,510
whereas cones give us our day vision.

182
00:10:00,510 --> 00:10:03,890
What this means is that rod cells
are more sensitive to light.

183
00:10:03,890 --> 00:10:05,740
They're the ones that
are activated and that

184
00:10:05,740 --> 00:10:08,698
are in use when you're outside in
the middle of the night, for example.

185
00:10:08,698 --> 00:10:11,860
And cones tend to be in use when
you have high detailed vision

186
00:10:11,860 --> 00:10:14,930
or when you're actually in daylight.

187
00:10:14,930 --> 00:10:17,700
So just like we were saying,
rods have more light sensitivity,

188
00:10:17,700 --> 00:10:19,549
cones have less.

189
00:10:19,549 --> 00:10:21,840
In the fovea, which was that
structure that I mentioned

190
00:10:21,840 --> 00:10:26,120
that's in the very middle of the retina
in the center of your field of view

191
00:10:26,120 --> 00:10:30,630
you have a high concentration of
cones and a low concentration of rods.

192
00:10:30,630 --> 00:10:34,690
In fact, the relative presence of
rods overall in your entire retina

193
00:10:34,690 --> 00:10:35,410
is very high.

194
00:10:35,410 --> 00:10:38,870
You have far more rods than you have
cones, which is pretty interesting

195
00:10:38,870 --> 00:10:44,487
and sort of eludes a little bit to the
fact that the greatest amount of detail

196
00:10:44,487 --> 00:10:46,570
that we have and the
greatest amount of day vision

197
00:10:46,570 --> 00:10:49,540
that we have is in the
very center of our vision.

198
00:10:49,540 --> 00:10:54,521
>> When we go outside at night if you've
been to a planetarium for example,

199
00:10:54,521 --> 00:10:56,270
you might have heard
the host actually say

200
00:10:56,270 --> 00:10:58,640
that when you want to look
at something up in the sky

201
00:10:58,640 --> 00:11:01,100
actually look at it in
the corner of your eye.

202
00:11:01,100 --> 00:11:04,020
The reason for that is you have
more rods in your periphery

203
00:11:04,020 --> 00:11:05,950
than you do in the
center, and this means

204
00:11:05,950 --> 00:11:09,210
that you can perhaps see that
detail a little bit better

205
00:11:09,210 --> 00:11:11,400
with that more sensitive cell.

206
00:11:11,400 --> 00:11:13,760
>> Now, the primary stimulus
for cones is trichomatic,

207
00:11:13,760 --> 00:11:16,450
that means that the cones are
really the ones that provide to us

208
00:11:16,450 --> 00:11:20,400
our color vision, so among other
reasons this in combination

209
00:11:20,400 --> 00:11:24,245
is why in broad daylight we can
actually perceive far more colors

210
00:11:24,245 --> 00:11:25,870
than we can in the middle of the night.

211
00:11:25,870 --> 00:11:27,480
You might have noticed if you go
outside in the middle of the night

212
00:11:27,480 --> 00:11:30,050
the colors don't seem to be as bright.

213
00:11:30,050 --> 00:11:32,660
One of the reasons for
that is that the cones

214
00:11:32,660 --> 00:11:35,450
are the ones that provide
to us our color vision,

215
00:11:35,450 --> 00:11:39,960
and the cones are what
become inactive at night.

216
00:11:39,960 --> 00:11:41,974
>> Now similarly, rods
actually detect motion

217
00:11:41,974 --> 00:11:44,640
and this is another reason why
it's very useful in the periphery

218
00:11:44,640 --> 00:11:47,764
and why we can detect motion more in
the periphery than when we're actually

219
00:11:47,764 --> 00:11:50,090
looking directly at something.

220
00:11:50,090 --> 00:11:53,280
Now, the reason that we are able to
actually have trichromatic vision out

221
00:11:53,280 --> 00:11:57,480
of these cones cells is because
we have different types of cones

222
00:11:57,480 --> 00:12:03,120
that respond to different wavelengths
of light, and it's not an exact science.

223
00:12:03,120 --> 00:12:06,500
We don't say that one
specific type of cone cell

224
00:12:06,500 --> 00:12:09,230
responds precisely to some
specific wavelengths of light,

225
00:12:09,230 --> 00:12:11,930
know there's a response curve
that's associated with these.

226
00:12:11,930 --> 00:12:15,160
And that implies that some of them
there's some overlap in this element,

227
00:12:15,160 --> 00:12:20,650
so we might actually have
sort of a non-linear stimulus

228
00:12:20,650 --> 00:12:22,020
to various types of colors.

229
00:12:22,020 --> 00:12:24,936
>> And in fact, this is precisely what
happens, if we take a look at this

230
00:12:24,936 --> 00:12:28,840
we have three different types of
cells-- The s-type cell, which

231
00:12:28,840 --> 00:12:32,120
is for short wavelengths, the
MDL types, which are absolutely

232
00:12:32,120 --> 00:12:34,690
the most prevalent types
of cones within our eye,

233
00:12:34,690 --> 00:12:38,980
and you notice that those are
very high up in this spectrum,

234
00:12:38,980 --> 00:12:41,880
much closer to the green spectrum.

235
00:12:41,880 --> 00:12:43,950
And this actually is
very, very important to us

236
00:12:43,950 --> 00:12:47,230
as digital photographers and in
the construction of digital cameras

237
00:12:47,230 --> 00:12:54,160
because this is one of the primary
reasons why-- well, there's

238
00:12:54,160 --> 00:12:56,640
a lot of things that this
impacts and hopefully we'll

239
00:12:56,640 --> 00:12:57,990
get a chance to get to them.

240
00:12:57,990 --> 00:13:00,980
But the result of this
is that we actually

241
00:13:00,980 --> 00:13:06,250
respond better to green wavelengths
than we do to red or to blue,

242
00:13:06,250 --> 00:13:08,990
and in fact our response curve
is very different for that.

243
00:13:08,990 --> 00:13:11,600
>> And if you sort of close
your eyes for just a minute

244
00:13:11,600 --> 00:13:16,210
and imagine that you have three
similar rooms that are all

245
00:13:16,210 --> 00:13:19,590
totally dark except in the very
center there is a light bulb.

246
00:13:19,590 --> 00:13:22,572
And in one room, you
have a green light bulb,

247
00:13:22,572 --> 00:13:25,780
in one room you have a red light bulb,
in another you have a blue light bulb,

248
00:13:25,780 --> 00:13:28,370
and that's all you have in
this room for illumination.

249
00:13:28,370 --> 00:13:32,470
And if you imagine the relative
brightness of these rooms based

250
00:13:32,470 --> 00:13:37,420
purely on this single light
source, try to imagine

251
00:13:37,420 --> 00:13:41,950
which one might feel brighter,
and the correct answer is green.

252
00:13:41,950 --> 00:13:46,360
Generally what happens is that because
we respond, because our cone cells are

253
00:13:46,360 --> 00:13:50,010
stimulated much more by the green
wavelengths than by any others,

254
00:13:50,010 --> 00:13:55,700
we respond much more to that
light, and so that is actually

255
00:13:55,700 --> 00:13:58,750
very important for our perception
of brightness and luminous,

256
00:13:58,750 --> 00:14:04,130
as opposed to some of
these other colors.

257
00:14:04,130 --> 00:14:08,570
>> Now, if we take a look again at
this, the eye structure that we had,

258
00:14:08,570 --> 00:14:11,810
we had course light that comes in
on the left side of this diagram

259
00:14:11,810 --> 00:14:15,090
through the iris, focused by the lens
and onto this so-called "censor,"

260
00:14:15,090 --> 00:14:19,110
our retina at the very back of
the eye, and this is very similar

261
00:14:19,110 --> 00:14:22,850
to the structure of a digital
camera as well in some ways.

262
00:14:22,850 --> 00:14:26,110
We have a lens, which is actually
used the focus the light.

263
00:14:26,110 --> 00:14:28,320
And that light is then
focused onto the very back

264
00:14:28,320 --> 00:14:31,100
of the camera, which has the sensor.

265
00:14:31,100 --> 00:14:35,546
>> Now this is a diagram of a digital
SLR-- a single lens reflex camera, which

266
00:14:35,546 --> 00:14:37,420
for those of you that
are unfamiliar are sort

267
00:14:37,420 --> 00:14:39,003
of the more professional looking ones.

268
00:14:39,003 --> 00:14:41,720
They're the ones that
allow you to change lenses,

269
00:14:41,720 --> 00:14:45,760
they're the ones that have a hump
on the top of the camera where

270
00:14:45,760 --> 00:14:48,890
the prism and the viewfinder is so
you can actually look through it.

271
00:14:48,890 --> 00:14:51,270
The reason that it works
that way that it does

272
00:14:51,270 --> 00:14:54,390
is that the pentaprism actually
reflects the light that

273
00:14:54,390 --> 00:14:57,350
has come in through the
lens and reflected off

274
00:14:57,350 --> 00:15:00,565
a mirror that operates that
sits at a 45 degree angle.

275
00:15:00,565 --> 00:15:03,440
It goes up through the pentaprism
and then out through the viewfinder

276
00:15:03,440 --> 00:15:06,020
where you are able to see the image.

277
00:15:06,020 --> 00:15:09,930
>> When you actually take the exposure,
the mirror moves up and out of the way,

278
00:15:09,930 --> 00:15:13,930
the shutter is opened, and that allows
the light to pass all the way back

279
00:15:13,930 --> 00:15:18,280
through and directly hit the sensor,
which causes the exposure to happen.

280
00:15:18,280 --> 00:15:24,810
So in the typical configuration you
cannot actually see the image through

281
00:15:24,810 --> 00:15:28,185
the viewfinder in a proper digital
SLR, you cannot actually see the image

282
00:15:28,185 --> 00:15:31,150
through the viewfinder and
also capture the image.

283
00:15:31,150 --> 00:15:32,900
If you happen to have
one of these cameras

284
00:15:32,900 --> 00:15:35,250
you might say well I
have a preview mode,

285
00:15:35,250 --> 00:15:39,620
but what that essentially does it
lifts the mirror out of the way.

286
00:15:39,620 --> 00:15:43,510
It turns off, essentially disables,
the optical viewfinder and it

287
00:15:43,510 --> 00:15:46,866
uses the screen on the back of
the camera based on the light

288
00:15:46,866 --> 00:15:49,592
that the sensor is receiving.

289
00:15:49,592 --> 00:15:54,520
>> Now there's an important aspect of
light to recognize beyond the fact

290
00:15:54,520 --> 00:16:00,360
that it is made up of wavelengths,
that it is made up of colors, that

291
00:16:00,360 --> 00:16:02,360
as a result of the various
wavelengths, and that

292
00:16:02,360 --> 00:16:05,900
is that the individual
photons that make up the light

293
00:16:05,900 --> 00:16:08,580
have a direct correlation
to the relative brightness,

294
00:16:08,580 --> 00:16:10,790
or to the intensity of that light.

295
00:16:10,790 --> 00:16:14,100
So every time that we
double the number of photons

296
00:16:14,100 --> 00:16:16,932
at any particular wavelength
of that light then

297
00:16:16,932 --> 00:16:18,640
we're essentially
doubling the intensity,

298
00:16:18,640 --> 00:16:21,380
we're doubling the
brightness of that light,

299
00:16:21,380 --> 00:16:23,840
and this has a very important
name in photography.

300
00:16:23,840 --> 00:16:25,340
It's called stops.

301
00:16:25,340 --> 00:16:28,680
So when we're talking about exposure,
we talk about stops in this way.

302
00:16:28,680 --> 00:16:35,235
We generally want to try to manipulate
this is quantized notion of photons

303
00:16:35,235 --> 00:16:37,380
that are actually
entering into our camera

304
00:16:37,380 --> 00:16:41,930
by either having or doubling the
amount of light that is allowed in.

305
00:16:41,930 --> 00:16:46,110
So it's very, very
frequent that you'll see

306
00:16:46,110 --> 00:16:48,640
numbers related to this idea of stops.

307
00:16:48,640 --> 00:16:51,576
So for example, the idea
of exposure compensation,

308
00:16:51,576 --> 00:16:53,450
which we'll talk more
about in just a minute,

309
00:16:53,450 --> 00:16:56,920
operates in this notion of
stops where a single stop

310
00:16:56,920 --> 00:16:59,520
is a doubling or halving
depending on the direction

311
00:16:59,520 --> 00:17:03,000
you're going of the amount of
light that's being entered.

312
00:17:03,000 --> 00:17:07,010
>> Now of course, when we're talking about
a number of stops, so for example,

313
00:17:07,010 --> 00:17:11,740
let's say we're talking about a change
of two stops as opposed to one stop.

314
00:17:11,740 --> 00:17:15,530
This means we are not just doubling
it, but we're doubling it again,

315
00:17:15,530 --> 00:17:19,300
so a changing two stops
results in a four times

316
00:17:19,300 --> 00:17:21,740
difference in the
intensity of the light.

317
00:17:21,740 --> 00:17:23,980
Likewise, a three stop
differences is eight,

318
00:17:23,980 --> 00:17:26,230
four stops is 16, so on and so forth.

319
00:17:26,230 --> 00:17:29,760
>> So even a low number
of stops can represent

320
00:17:29,760 --> 00:17:33,980
a wide variety of different
intensities in light.

321
00:17:33,980 --> 00:17:38,350
And in fact, when we're talking
about daylight versus the brightest

322
00:17:38,350 --> 00:17:43,010
day versus the darkest night we're
really talking about 20 stops perhaps

323
00:17:43,010 --> 00:17:44,210
at the absolute most.

324
00:17:44,210 --> 00:17:48,020
That's probably something
closer to 15 stops or so,

325
00:17:48,020 --> 00:17:50,180
but that'll be important
in just a minute as we

326
00:17:50,180 --> 00:17:52,330
keep talking about exposure.

327
00:17:52,330 --> 00:17:55,610
>> So we talked a little bit about
light and so let's talk about some

328
00:17:55,610 --> 00:17:58,320
of these other exposure
settings that actually

329
00:17:58,320 --> 00:18:02,930
allow us to capture the
light that exists in a scene.

330
00:18:02,930 --> 00:18:05,450
There's shutter speed,
there's ISO and aperture,

331
00:18:05,450 --> 00:18:07,870
and we alluded a little bit
to shutter speed before,

332
00:18:07,870 --> 00:18:11,780
but I do have a video that sort
of shows the anatomy of a camera

333
00:18:11,780 --> 00:18:16,530
and also will illuminate this
idea of the shutter itself.

334
00:18:16,530 --> 00:18:19,170
So I have here this
high speed photo which

335
00:18:19,170 --> 00:18:22,170
I happened to find on the
internet, and what you will see

336
00:18:22,170 --> 00:18:26,570
is this action of actually
capturing an exposure

337
00:18:26,570 --> 00:18:29,470
on this particular digital SLR.

338
00:18:29,470 --> 00:18:33,640
>> So as I'm talking I want you to pay
attention to a couple of things.

339
00:18:33,640 --> 00:18:37,640
First, notice that the mirror
moves up out of the way,

340
00:18:37,640 --> 00:18:40,500
recall that we talked about
this in a digital SLR.

341
00:18:40,500 --> 00:18:43,520
Now notice that the thing that
we're seeing it behind that

342
00:18:43,520 --> 00:18:48,280
is not the raw sensor itself, but
it is in fact a piece of plastic

343
00:18:48,280 --> 00:18:53,040
or Kevlar depending on the
quality of the camera that

344
00:18:53,040 --> 00:18:54,060
operates as the shutter.

345
00:18:54,060 --> 00:18:57,040
It's a mechanical shutter
actually move out of the way

346
00:18:57,040 --> 00:18:59,821
and exposes the sensor underneath.

347
00:18:59,821 --> 00:19:01,570
So let's take a look
at this one more time

348
00:19:01,570 --> 00:19:04,640
so you can sort of watch
the action of the shutter.

349
00:19:04,640 --> 00:19:07,330
The mirror moves up by the
way, notice shutter opens

350
00:19:07,330 --> 00:19:11,600
and then very quickly there is
another curtain that closes behind it.

351
00:19:11,600 --> 00:19:16,080
This is a very typical set up for
digital SLRs with mechanical shutters.

352
00:19:16,080 --> 00:19:19,340
We'll have two curtains that
operates either horizontally

353
00:19:19,340 --> 00:19:23,170
or vertically depending
on the particular camera

354
00:19:23,170 --> 00:19:25,240
and it will move across
the entire plane.

355
00:19:25,240 --> 00:19:28,540
First the first curtain will open,
exposing the sensor underneath,

356
00:19:28,540 --> 00:19:33,420
and the second curtain will close
thereby stopping the exposure.

357
00:19:33,420 --> 00:19:36,720
>> Now there are other types of shutters
as well, and really for our purposes

358
00:19:36,720 --> 00:19:40,712
we don't have to worry about them too
much except for the electronic shutter.

359
00:19:40,712 --> 00:19:42,920
So this is a mechanical
shutter, and you'll typically

360
00:19:42,920 --> 00:19:45,875
find this on digital SLR.

361
00:19:45,875 --> 00:19:47,750
And the whole combination
of these movements,

362
00:19:47,750 --> 00:19:49,708
including the mirror
moving up, out of the way,

363
00:19:49,708 --> 00:19:52,800
the shutter opening, and then the
second curtain closing behind it,

364
00:19:52,800 --> 00:19:57,220
results in that characteristic
click that we hear in cameras.

365
00:19:57,220 --> 00:19:59,820
But for cameras that don't
actually make that physical noise,

366
00:19:59,820 --> 00:20:05,010
such as camera phones and
compact cameras and smart phones

367
00:20:05,010 --> 00:20:08,680
and a variety of others is that
they have an electronic shutter.

368
00:20:08,680 --> 00:20:12,130
An electronic shattered doesn't
operate in the same way,

369
00:20:12,130 --> 00:20:15,540
but rather it starts to read data off
the sensor and then immediately stops,

370
00:20:15,540 --> 00:20:21,600
or rather it allows the sensor to
accumulate the data of the changes

371
00:20:21,600 --> 00:20:25,090
in the voltage caused by
photons hitting the sensor

372
00:20:25,090 --> 00:20:29,770
and then it will actually clear it
once the exposure is actually complete.

373
00:20:29,770 --> 00:20:35,140
>> So this is sort of the most rigid
definition of shutter speed,

374
00:20:35,140 --> 00:20:40,900
but what ultimately this means is that
this is defining how much light we

375
00:20:40,900 --> 00:20:45,810
are actually receiving
on the sensor plane,

376
00:20:45,810 --> 00:20:49,060
and ultimately this means
that we can change the shutter

377
00:20:49,060 --> 00:20:51,220
speed in terms of stops.

378
00:20:51,220 --> 00:20:53,930
We might have the shutter
open for a single second,

379
00:20:53,930 --> 00:20:57,290
for example, and so we would say that
our shutter speed is then one second.

380
00:20:57,290 --> 00:21:01,010
And what that means in the mechanical
terms is that the first curtain opens,

381
00:21:01,010 --> 00:21:03,370
the sensor is then exposed
to light for one second,

382
00:21:03,370 --> 00:21:06,060
and then the second
curtain closes behind it.

383
00:21:06,060 --> 00:21:08,030
>> Then of course, we can
change this by a stop

384
00:21:08,030 --> 00:21:11,220
if we go a stop brighter
this means that we then

385
00:21:11,220 --> 00:21:14,010
have to keep the
shutter open for longer,

386
00:21:14,010 --> 00:21:16,240
so that we can collect more photons.

387
00:21:16,240 --> 00:21:20,570
So a stop brighter would result
in two second shutter speed.

388
00:21:20,570 --> 00:21:23,770
Likewise, a stop darker, which would
mean that we have to have the shutter

389
00:21:23,770 --> 00:21:28,149
open for less amount of time so we would
have half a second of a shutter speed.

390
00:21:28,149 --> 00:21:30,690
We can keep going in either
direction, but if you play around

391
00:21:30,690 --> 00:21:32,860
with the settings on
your camera, you probably

392
00:21:32,860 --> 00:21:35,810
will notice that it seems
to approximately double

393
00:21:35,810 --> 00:21:39,130
or halve depending on the
direction of your tuning.

394
00:21:39,130 --> 00:21:43,030
>> Now, the shutter speed because we
can have it open for some arbitrary

395
00:21:43,030 --> 00:21:46,700
amount of time does have
some impact on our image.

396
00:21:46,700 --> 00:21:49,170
In particular, imagine
what happens if you're

397
00:21:49,170 --> 00:21:52,830
capturing all of the photons
in a particular scene

398
00:21:52,830 --> 00:21:54,550
over a couple of seconds.

399
00:21:54,550 --> 00:21:57,740
You might imagine if there's
some movement within this scene,

400
00:21:57,740 --> 00:22:00,610
so for example there's a ball
that moves across the scene,

401
00:22:00,610 --> 00:22:02,370
or in the case of this
photograph there's

402
00:22:02,370 --> 00:22:04,760
a wave that moves across the scene.

403
00:22:04,760 --> 00:22:07,980
>> I'm capturing the photons
from that entire movement,

404
00:22:07,980 --> 00:22:10,380
so this is causing a
motion blur that becomes

405
00:22:10,380 --> 00:22:14,370
very visible within the photograph
and sometimes this is intentional.

406
00:22:14,370 --> 00:22:17,650
Sometimes you actually want to get
some motion blur so you can smooth out

407
00:22:17,650 --> 00:22:20,980
the motion of the waves,
for example, or perhaps you

408
00:22:20,980 --> 00:22:23,900
want to actually capture
movement of a fast moving

409
00:22:23,900 --> 00:22:28,450
car, you want to actually capture the
movement of fireworks, for example.

410
00:22:28,450 --> 00:22:31,990
By the way, many people love to go
outside and take pictures of fireworks

411
00:22:31,990 --> 00:22:35,500
and have very high, fast shutter
speeds, which just looks abysmal,

412
00:22:35,500 --> 00:22:39,241
because it's just the brief moment of
explosion or a couple of seconds after

413
00:22:39,241 --> 00:22:40,490
and then they're all chimping.

414
00:22:40,490 --> 00:22:41,698
>> Do you know what chimping is?

415
00:22:41,698 --> 00:22:45,180
It's like you take a picture, right, and
then you're hunched over your camera,

416
00:22:45,180 --> 00:22:47,471
and you show your friends
and you're like, " oh,oh,oh."

417
00:22:47,471 --> 00:22:48,280
Chimping, right?

418
00:22:48,280 --> 00:22:48,890
OK.

419
00:22:48,890 --> 00:22:52,487
>> So come back, so you have this
idea of fireworks where it's really

420
00:22:52,487 --> 00:22:55,070
the movements of these fireworks
that's really interesting, so

421
00:22:55,070 --> 00:22:57,310
try experimenting with
your shutter speed

422
00:22:57,310 --> 00:23:00,900
and capturing the movement
using a very long shutter speed,

423
00:23:00,900 --> 00:23:02,460
rather than a very short one.

424
00:23:02,460 --> 00:23:05,300
Of course, this means
that you can get motion

425
00:23:05,300 --> 00:23:07,130
blur due to a wide variety of factors.

426
00:23:07,130 --> 00:23:10,680
It might not just be the object in
this scene that's moving quickly,

427
00:23:10,680 --> 00:23:15,200
as is the case in the fireworks here,
or the other car or the environment

428
00:23:15,200 --> 00:23:17,940
in this photo on the
left, but instead imagine

429
00:23:17,940 --> 00:23:22,790
if you're trying to hold of the
phone or your camera for that long.

430
00:23:22,790 --> 00:23:25,110
No matter how much you
actually brace yourself,

431
00:23:25,110 --> 00:23:28,440
you will have a small amount of
movement that translates to some motion

432
00:23:28,440 --> 00:23:30,450
blur within your camera.

433
00:23:30,450 --> 00:23:32,640
>> So if you're try to
counteract that, you either

434
00:23:32,640 --> 00:23:36,630
have to increase the shutter speed so
that it decreases the amount of time

435
00:23:36,630 --> 00:23:39,930
that the shutters actually open
and thereby freezing that motion,

436
00:23:39,930 --> 00:23:42,716
or you need to stabilize
the camera in some way.

437
00:23:42,716 --> 00:23:44,590
In which, case you might
want to use a tripod

438
00:23:44,590 --> 00:23:48,190
or to set the camera down on some stable
table or something along those lines

439
00:23:48,190 --> 00:23:50,785
to actually freeze
that particular motion.

440
00:23:50,785 --> 00:23:52,660
So this is an artistic
question that you have

441
00:23:52,660 --> 00:23:56,080
to ask yourself is in which direction
do I actually want to take this,

442
00:23:56,080 --> 00:24:01,790
do I want to try to capture the motion
by having this intentional motion blur,

443
00:24:01,790 --> 00:24:04,400
or do I want to freeze
the motion, and sometimes

444
00:24:04,400 --> 00:24:07,580
freezing the motion is precisely what
you want, in the example of sports

445
00:24:07,580 --> 00:24:08,610
photography for example.

446
00:24:08,610 --> 00:24:13,260
>> You really want to capture that precise
moment that something is happening,

447
00:24:13,260 --> 00:24:17,610
or perhaps rather than get this smooth
motion of the entirety of some ways

448
00:24:17,610 --> 00:24:20,460
you really want to capture
the sort of instant moment

449
00:24:20,460 --> 00:24:23,070
that a wave crashes or
breaks against the rock

450
00:24:23,070 --> 00:24:24,810
and you want to capture that moment.

451
00:24:24,810 --> 00:24:26,940
You certainly will want to capture this.

452
00:24:26,940 --> 00:24:30,730
By the way, this is what it looks like,
my camera got soaked, I got soaked,

453
00:24:30,730 --> 00:24:31,890
it was totally fine.

454
00:24:31,890 --> 00:24:33,639
Don't worry about it,
a lot of cameras are

455
00:24:33,639 --> 00:24:37,140
a lot stronger than you would imagine.

456
00:24:37,140 --> 00:24:39,950
The buttons on the camera
were a little gritty

457
00:24:39,950 --> 00:24:43,010
from the sand stuff--
ended up being fine.

458
00:24:43,010 --> 00:24:48,290
>> Now sometimes you actually want to mix
both motion and still in one camera.

459
00:24:48,290 --> 00:24:51,040
So imagine what happens if
you have a moving object

460
00:24:51,040 --> 00:24:57,610
and you pan your camera with that object
keeping some part of that object still

461
00:24:57,610 --> 00:25:00,980
totally still relative to
some portion on your sensor,

462
00:25:00,980 --> 00:25:04,680
if you're able to have a long shutter
speed that actually captures movement

463
00:25:04,680 --> 00:25:08,540
of the environment but you keep
that one part of the object

464
00:25:08,540 --> 00:25:12,700
still relative to some portion on
your sensor you can mix both and get

465
00:25:12,700 --> 00:25:18,260
a sort of neat effect where you are
able to get something in sharp focus

466
00:25:18,260 --> 00:25:20,910
and without any motion
blur, but sort of blur

467
00:25:20,910 --> 00:25:24,240
everything else in the environment.

468
00:25:24,240 --> 00:25:26,820
And sometimes this is actually
what you want also for sports,

469
00:25:26,820 --> 00:25:31,230
sometimes you do you want to convey
this motion of the motion itself

470
00:25:31,230 --> 00:25:32,990
or the idea of speed.

471
00:25:32,990 --> 00:25:36,600
So for example, in a
car race you might not

472
00:25:36,600 --> 00:25:39,749
want to totally freeze the
motion of the car and the wheels,

473
00:25:39,749 --> 00:25:42,040
because then it will look
like it's not going anywhere.

474
00:25:42,040 --> 00:25:44,120
It's just standing on
the track, providing

475
00:25:44,120 --> 00:25:51,129
some of that can actually give
some amount of drama to the scene.

476
00:25:51,129 --> 00:25:53,670
So let's take a step back from
the shutter speed a little bit

477
00:25:53,670 --> 00:25:56,410
and talk about some of these
other settings as well.

478
00:25:56,410 --> 00:25:59,340
One of them is ISO, and
you might have heard

479
00:25:59,340 --> 00:26:02,370
of the term in the
context of sensitivity,

480
00:26:02,370 --> 00:26:05,400
but that's not really an accurate
way of thinking about it, at least

481
00:26:05,400 --> 00:26:07,590
in terms of digital cameras.

482
00:26:07,590 --> 00:26:10,211
We're not actually changing
the sensitivity of the camera,

483
00:26:10,211 --> 00:26:12,460
there's actually some other
electronic trickery that's

484
00:26:12,460 --> 00:26:16,240
happening underneath the hood,
but for our purposes for now,

485
00:26:16,240 --> 00:26:19,310
thinking of it as
sensitivity is an OK way

486
00:26:19,310 --> 00:26:22,960
to think about it, especially
In terms of exposure value.

487
00:26:22,960 --> 00:26:26,380
>> So ISO starts generally
at a round value of 100.

488
00:26:26,380 --> 00:26:29,870
It's just sort of an
arbitrary value, and if we

489
00:26:29,870 --> 00:26:33,820
are to think of it in our
simplified terms as sensitivity,

490
00:26:33,820 --> 00:26:37,600
increasing the ISO means that
the sensor become slightly more

491
00:26:37,600 --> 00:26:40,280
sensitive to light,
which would then allow

492
00:26:40,280 --> 00:26:43,950
us to change the shutter
speed to be faster.

493
00:26:43,950 --> 00:26:46,700
So, in other words because we're
trying to get the amount of light

494
00:26:46,700 --> 00:26:51,140
in our scene to match the
specific range of our camera

495
00:26:51,140 --> 00:26:54,630
we have to play with these
settings, so these two settings

496
00:26:54,630 --> 00:26:58,270
that we've mentioned and also aperture
that we'll talk about in just a moment,

497
00:26:58,270 --> 00:27:03,704
in order to really get that precise
range of photons within our sensor.

498
00:27:03,704 --> 00:27:06,620
So one of the ways that we're able
to do this one, and one of the ways

499
00:27:06,620 --> 00:27:08,470
that we're able to
alter our shutter speed

500
00:27:08,470 --> 00:27:12,460
is to also change the
ISO for a given scene.

501
00:27:12,460 --> 00:27:16,420
So by increasing the ISO we
increase the so-called sensitivity,

502
00:27:16,420 --> 00:27:19,820
which allows us to make
the shutter speed faster,

503
00:27:19,820 --> 00:27:23,570
or likewise perhaps we actually want
to make the shutter speed longer.

504
00:27:23,570 --> 00:27:25,950
Perhaps we actually
want to have a lower ISO

505
00:27:25,950 --> 00:27:30,170
and increase the time that the
shutter is open to capture our motion

506
00:27:30,170 --> 00:27:34,330
or to capture that motion blur
for some artistic purpose.

507
00:27:34,330 --> 00:27:36,830
>> Now the downside to ISO of
course, is that we actually

508
00:27:36,830 --> 00:27:39,330
get a fair amount of noise as a result.

509
00:27:39,330 --> 00:27:42,220
And these are some examples
from relatively old cameras,

510
00:27:42,220 --> 00:27:47,570
but generally this shows an
interesting general trend

511
00:27:47,570 --> 00:27:52,500
that larger cameras tend to do slightly
better in combating issues of noise.

512
00:27:52,500 --> 00:27:55,350
And it's not really the case
that larger cameras are doing it,

513
00:27:55,350 --> 00:28:00,000
there's a lot of factors that play
into this-- the age of the sensor

514
00:28:00,000 --> 00:28:03,181
is one important distinction,
but also the size of the pixel,

515
00:28:03,181 --> 00:28:04,930
so it's not really the
size of the camera,

516
00:28:04,930 --> 00:28:08,950
but the size of the pixels itself can
make a huge difference because larger

517
00:28:08,950 --> 00:28:12,150
pixels can capture more light,
there's more area through which you

518
00:28:12,150 --> 00:28:13,850
can actually capture more photons.

519
00:28:13,850 --> 00:28:15,850
And also the electronics
are a little bit bigger

520
00:28:15,850 --> 00:28:21,570
and they cant hold
more voltage, perhaps,

521
00:28:21,570 --> 00:28:24,320
and be able to give us a
better signal to noise ratio.

522
00:28:24,320 --> 00:28:28,720
So there's a variety of reasons why,
but generally speaking, bigger sensors

523
00:28:28,720 --> 00:28:33,245
or bigger pixels more specifically
allow us to get better quality out

524
00:28:33,245 --> 00:28:35,270
of our higher ISO settings.

525
00:28:35,270 --> 00:28:38,750
If you're really struggling with getting
a lot of noise out of your images,

526
00:28:38,750 --> 00:28:41,900
perhaps you are using, for
example, a smartphone that

527
00:28:41,900 --> 00:28:44,710
has a sensor that's really,
really small and because it

528
00:28:44,710 --> 00:28:47,910
has a very high megapixel
count, the pixels also

529
00:28:47,910 --> 00:28:55,190
have to be very small, which results in
a relatively noisy image at high ISOs.

530
00:28:55,190 --> 00:29:00,700
>> So one of things that we've noticed is
that ISO noise improvements has just

531
00:29:00,700 --> 00:29:02,770
been enormous, especially
in recent years.

532
00:29:02,770 --> 00:29:09,020
The sensors essentially a technology
very similar to that of our computers

533
00:29:09,020 --> 00:29:11,390
and over time it's
really, really improved,

534
00:29:11,390 --> 00:29:18,650
and nowadays the noise that we see
in digital cameras really greatly

535
00:29:18,650 --> 00:29:22,020
exceeds the noise capabilities of film.

536
00:29:22,020 --> 00:29:24,560
So in other words, the digital
cameras with digital cameras

537
00:29:24,560 --> 00:29:29,080
we can take images that are far
less grainy, far cleaner than film,

538
00:29:29,080 --> 00:29:31,930
and this is maybe good or bad
depending on how you look at it.

539
00:29:31,930 --> 00:29:34,890
Sometimes you like having that
additional texture for that,

540
00:29:34,890 --> 00:29:39,110
but you can of course add
that later in software.

541
00:29:39,110 --> 00:29:43,770
>> So let's take these two into
combination in these two ideas

542
00:29:43,770 --> 00:29:49,750
and combine them to realize how we
can alter one to impact the other.

543
00:29:49,750 --> 00:29:52,960
So in the context of
ISO and shutter speed,

544
00:29:52,960 --> 00:29:55,720
imagine that I'm taking
this photograph, which

545
00:29:55,720 --> 00:29:58,530
I did many years ago back
in 2007 in New Hampshire.

546
00:29:58,530 --> 00:30:02,730
I was on a dock at the
edge of Lake Winnipesaukee

547
00:30:02,730 --> 00:30:07,000
and there was some cool stars
whose trails I wanted to capture.

548
00:30:07,000 --> 00:30:10,270
So I set my camera
outside, changed the modes

549
00:30:10,270 --> 00:30:13,300
so that I could have several
minutes worth of exposure time,

550
00:30:13,300 --> 00:30:18,060
and just waited outside in the cold
for 15 minutes and got this picture.

551
00:30:18,060 --> 00:30:21,980
>> And so there's a variety of stars
here, it's an OK photograph,

552
00:30:21,980 --> 00:30:25,660
but at the very center I've
highlighted one particular star, which

553
00:30:25,660 --> 00:30:29,511
I think I asked an astronomer friend
and they said it was big at the time.

554
00:30:29,511 --> 00:30:31,260
One of the interesting
things to notice is

555
00:30:31,260 --> 00:30:35,390
that you can of course see the
Earth's rotation in the star trails,

556
00:30:35,390 --> 00:30:38,180
but notice that the
radius of the circle seems

557
00:30:38,180 --> 00:30:41,160
to get smaller as you get
to the upper right portion.

558
00:30:41,160 --> 00:30:44,610
That's because I was pointing
the camera towards the north,

559
00:30:44,610 --> 00:30:49,200
and this appeared just of
the slide just little bit

560
00:30:49,200 --> 00:30:57,900
was the North star through
which the Earth was rotating.

561
00:30:57,900 --> 00:30:58,400
OK.

562
00:30:58,400 --> 00:31:01,280
So anyway, we have this star
that I want to point out.

563
00:31:01,280 --> 00:31:04,170
Vega, it has a specific
length, and realized

564
00:31:04,170 --> 00:31:08,770
that if I wanted to make the
star trail longer the thing

565
00:31:08,770 --> 00:31:11,660
that I would need to do is
to change the shutter speed.

566
00:31:11,660 --> 00:31:15,230
I would have to have the shutter
open for a longer amount of time,

567
00:31:15,230 --> 00:31:17,390
but the amount of light
in this scene is fixed,

568
00:31:17,390 --> 00:31:20,960
I can't actually change the shutter
speed without changing something

569
00:31:20,960 --> 00:31:26,260
else so that the amount of
light that enters into my camera

570
00:31:26,260 --> 00:31:30,840
continues to be correct, and I continue
to get a properly exposed photograph.

571
00:31:30,840 --> 00:31:32,630
>> So I can of course
change the sensitivity,

572
00:31:32,630 --> 00:31:38,490
and if you're able to look at this
relatively small text below each

573
00:31:38,490 --> 00:31:41,400
of these images you'll
see the change that

574
00:31:41,400 --> 00:31:48,955
happened is that I changed the ISO by
one stop, so changing it from ISO 800

575
00:31:48,955 --> 00:31:53,840
to ISO 400, which then allowed
me to increase the shutter

576
00:31:53,840 --> 00:31:57,940
speed approximately by a value of 2.

577
00:31:57,940 --> 00:32:00,030
And that is how we were
able to get precisely

578
00:32:00,030 --> 00:32:04,850
this star trail that was twice as long.

579
00:32:04,850 --> 00:32:09,270
>> All right, so then let's talk
about this third idea of aperture.

580
00:32:09,270 --> 00:32:12,760
Now aperture, unlike
shutter speed and ISO,

581
00:32:12,760 --> 00:32:15,060
doesn't have a very
nice doubling or halving

582
00:32:15,060 --> 00:32:19,100
to represent a single
stop change in exposure.

583
00:32:19,100 --> 00:32:22,070
The reason for that is that
aperture or f-number is really

584
00:32:22,070 --> 00:32:26,630
a ratio of some things
that are related to a lens.

585
00:32:26,630 --> 00:32:30,680
Now this icon is actually from
the now defunct apple Aperture

586
00:32:30,680 --> 00:32:31,940
software, which is too bad.

587
00:32:31,940 --> 00:32:35,840
It was a fantastic software, but one
of the things that this icon has which

588
00:32:35,840 --> 00:32:39,770
is representative of a lot of
lenses that you have on cameras

589
00:32:39,770 --> 00:32:43,271
is the data on the lower
right of this lens.

590
00:32:43,271 --> 00:32:46,520
You notice that it says 50 millimeters,
which is the focal length of the lens,

591
00:32:46,520 --> 00:32:51,060
and it also has this 1:1.4, I know
it's upside down, but you can read it,

592
00:32:51,060 --> 00:32:55,280
it's 1:1.4 and that is
actually this aperture.

593
00:32:55,280 --> 00:33:00,590
This is actually the f-number, the
maximum possible aperture of this lens.

594
00:33:00,590 --> 00:33:02,660
And this is important
because this tells us

595
00:33:02,660 --> 00:33:05,780
quite a few properties about this
particular lens-- the focal length

596
00:33:05,780 --> 00:33:10,690
tells us how zoomed in or zoomed out it
is, 50 millimeters on a typical camera

597
00:33:10,690 --> 00:33:16,100
is a very stand sort of field of
view, it's not too zoomed out,

598
00:33:16,100 --> 00:33:19,380
it's not too zoomed in,
it's perhaps somewhat

599
00:33:19,380 --> 00:33:23,860
equal to how it would look to
our eye, but there's definitely

600
00:33:23,860 --> 00:33:26,170
some changes in the field of view.

601
00:33:26,170 --> 00:33:28,310
>> Let's take a look now at this aperture.

602
00:33:28,310 --> 00:33:34,390
The ratio here is precisely the
ratio of the focal length divided

603
00:33:34,390 --> 00:33:37,800
by the apertures effective diameter,
so what does this actually mean?

604
00:33:37,800 --> 00:33:40,050
So let's keep in mind this
division for just a minute.

605
00:33:40,050 --> 00:33:45,540
The f-number from this previous
slide was in fact this 1.4 value,

606
00:33:45,540 --> 00:33:49,110
the 1 colon just represents
the fact that this is a ratio,

607
00:33:49,110 --> 00:33:52,480
and the focal length
is this 50 millimeters.

608
00:33:52,480 --> 00:33:56,840
So this is important and we'll be
able to find out why in just a second.

609
00:33:56,840 --> 00:34:00,710
>> So here's an oversimplified view of
a lens, it's a side view of the lens.

610
00:34:00,710 --> 00:34:05,260
On the very far right of this image
we have an imaginary sensor plane.

611
00:34:05,260 --> 00:34:08,290
Notice this symbol here, there's
a vertical line with a circle.

612
00:34:08,290 --> 00:34:10,159
That represents a
sensor plane, and if you

613
00:34:10,159 --> 00:34:14,977
happen to have a digital SLR or
some sort of other advanced camera

614
00:34:14,977 --> 00:34:18,060
take a look on the body of that camera,
you might actually find the symbol

615
00:34:18,060 --> 00:34:21,080
and that represents the plane
through which your sensor actually

616
00:34:21,080 --> 00:34:25,480
exist somewhere within
that camera, but anyway we

617
00:34:25,480 --> 00:34:28,431
can measure the focal length from
the nodal point of the lens, which

618
00:34:28,431 --> 00:34:30,139
in this oversimplified
thing just happens

619
00:34:30,139 --> 00:34:34,199
to be in a single lens element, all
the way to the focal plane itself.

620
00:34:34,199 --> 00:34:37,260
And there's an effective
diameter of that lens.

621
00:34:37,260 --> 00:34:40,400
>> The diameter is the maximum
aperture through which

622
00:34:40,400 --> 00:34:45,275
the photons enter and are
focused on to the sensor.

623
00:34:45,275 --> 00:34:48,500
But imagine what might
happen for just a minute

624
00:34:48,500 --> 00:34:52,630
if we had this amount of
light that was actually

625
00:34:52,630 --> 00:34:56,370
able to enter through our lens,
but we actually restricted this,

626
00:34:56,370 --> 00:34:59,870
so we have some sort of device that
actually reduced the amount of light

627
00:34:59,870 --> 00:35:02,600
on the outside from
coming into this lens--

628
00:35:02,600 --> 00:35:04,720
very similar to the iris in our eyes.

629
00:35:04,720 --> 00:35:07,670
When you go outside,
for example, and it's

630
00:35:07,670 --> 00:35:11,050
bright daylight you might actually
notice that your iris constricts

631
00:35:11,050 --> 00:35:14,840
to let in less light, likewise when
you go inside into a very darkroom,

632
00:35:14,840 --> 00:35:16,730
your iris expands to allow more light.

633
00:35:16,730 --> 00:35:21,460
It's precisely am analogous
situation to what we have here.

634
00:35:21,460 --> 00:35:25,930
>> And so what this actually
means is that the f-number has

635
00:35:25,930 --> 00:35:33,170
some indication of precisely how
much light this lens is actually

636
00:35:33,170 --> 00:35:36,910
able to accumulate through this
diameter and the focal length,

637
00:35:36,910 --> 00:35:39,790
because as we actually
increase the focal length,

638
00:35:39,790 --> 00:35:44,970
the diameter would need to increase
to allow the same amount of photons

639
00:35:44,970 --> 00:35:49,200
to enter into the lens
and fall onto the sensor.

640
00:35:49,200 --> 00:35:51,840
So there's some math that we
can do to actually figure out

641
00:35:51,840 --> 00:35:59,780
precisely what a stop difference
is between the various f-numbers.

642
00:35:59,780 --> 00:36:02,760
So I'll hopefully be
able to post a handout

643
00:36:02,760 --> 00:36:05,310
next to the slides that will
actually show you that math.

644
00:36:05,310 --> 00:36:07,610
>> That goes through this and
takes all this into account,

645
00:36:07,610 --> 00:36:10,050
but you can also sort of
figure it out yourself

646
00:36:10,050 --> 00:36:12,500
through this ratio that
we were just talking about

647
00:36:12,500 --> 00:36:16,150
and imagine that the way that
we are able to restrict light

648
00:36:16,150 --> 00:36:19,660
through this mechanism is to
have different amounts of areas

649
00:36:19,660 --> 00:36:21,780
through which the light is able to flow.

650
00:36:21,780 --> 00:36:24,250
So if we have a circular
lens that has an aperture

651
00:36:24,250 --> 00:36:27,530
that's this large that means that
photons are flowing through that area,

652
00:36:27,530 --> 00:36:31,890
but imagine how this might change
if we actually restrict that area.

653
00:36:31,890 --> 00:36:35,050
So because we're actually talking
about a difference in area

654
00:36:35,050 --> 00:36:38,190
rather than some sort of linear
change, such as shutter speed,

655
00:36:38,190 --> 00:36:41,190
this is actually what causes
the very strange numbers

656
00:36:41,190 --> 00:36:43,170
that we see out of f-numbers.

657
00:36:43,170 --> 00:36:45,590
>> So there's an easy way to
remember the differences

658
00:36:45,590 --> 00:36:48,130
in one stop between
all of the f-numbers.

659
00:36:48,130 --> 00:36:54,750
First remember two numbers-- f1 and f1.2
and double each one to get a subsequent

660
00:36:54,750 --> 00:36:55,250
number.

661
00:36:55,250 --> 00:36:58,480
So for example, you would
double f1, we get f2,

662
00:36:58,480 --> 00:37:04,700
so now the string of aperture values
that we have are f1, f1.4, f2.

663
00:37:04,700 --> 00:37:07,400
Now we take that second
number, 1.4 and double that.

664
00:37:07,400 --> 00:37:11,040
So now we have 2 and 2.8, and we
can continue along in this fashion.

665
00:37:11,040 --> 00:37:15,180
4, 5.6, 8 and so on and so forth.

666
00:37:15,180 --> 00:37:19,630
This breaks down after about
the 32 or something like that,

667
00:37:19,630 --> 00:37:23,670
but it's close enough
approximation for our purposes.

668
00:37:23,670 --> 00:37:27,940
>> So just like shutter speed
and ISO, the aperture

669
00:37:27,940 --> 00:37:33,050
does have an impact on our images,
and one of the biggest impacts

670
00:37:33,050 --> 00:37:35,390
that it actually has
beyond the fact that it's

671
00:37:35,390 --> 00:37:38,820
allowing more or less light depending
on whether we have constricted

672
00:37:38,820 --> 00:37:42,570
our aperture or increased it's size,
the biggest change perhaps that it has

673
00:37:42,570 --> 00:37:45,160
is the amount of background
blur that you might actually

674
00:37:45,160 --> 00:37:46,900
have within your image.

675
00:37:46,900 --> 00:37:50,250
The larger the aperture,
the more background blur

676
00:37:50,250 --> 00:37:52,880
you'll actually have in your image.

677
00:37:52,880 --> 00:37:56,710
So you can reduce the size of the
aperture, thereby letting in lets light

678
00:37:56,710 --> 00:38:01,240
and get more of your
scene in focus, or you

679
00:38:01,240 --> 00:38:06,190
can try to increase the size of the
aperture by decreasing the f-number

680
00:38:06,190 --> 00:38:11,032
and you will get less of
the scene in proper focus.

681
00:38:11,032 --> 00:38:12,740
And this can be an
effective tool as well

682
00:38:12,740 --> 00:38:16,550
if you want to isolate your subject from
the background, for example, or perhaps

683
00:38:16,550 --> 00:38:19,770
you actually have a landscape shot
and you want to do the opposite.

684
00:38:19,770 --> 00:38:22,870
You want to try to get as much
of that as possible in focus,

685
00:38:22,870 --> 00:38:26,350
and so what you might actually do is
then decrease the size of the aperture

686
00:38:26,350 --> 00:38:31,460
by increasing your f-number and
altering the other shutter values,

687
00:38:31,460 --> 00:38:35,510
or the other exposure values as
appropriate to actually capture as much

688
00:38:35,510 --> 00:38:39,250
of your scene and focus
as you might like.

689
00:38:39,250 --> 00:38:40,619
>> So this is the big four.

690
00:38:40,619 --> 00:38:43,285
We talked about the amount of
available light, the shutter speed

691
00:38:43,285 --> 00:38:47,280
that's actually there, ISO, and aperture
and how the amount of available light

692
00:38:47,280 --> 00:38:52,330
is we're sort of at the mercy of the
scene that we happen to be capturing,

693
00:38:52,330 --> 00:38:55,500
unless we happen to have an
indoor setup or some other way

694
00:38:55,500 --> 00:38:58,210
that we can impact that
amount of light, and how

695
00:38:58,210 --> 00:39:01,730
we can use the three values--
shutter speed, ISO, and aperture,

696
00:39:01,730 --> 00:39:06,010
to vary the amount of light
that enters to our sensor

697
00:39:06,010 --> 00:39:08,690
and captures our exposure.

698
00:39:08,690 --> 00:39:10,950
And so there's this
discussion of stops and how

699
00:39:10,950 --> 00:39:13,550
I mentioned earlier on how
there's this distinction.

700
00:39:13,550 --> 00:39:16,060
>> There's about 20 stops
difference perhaps

701
00:39:16,060 --> 00:39:20,650
between the brightest bright day and
the darkest dark night without any moon

702
00:39:20,650 --> 00:39:23,480
shining or anything
like that, and cameras

703
00:39:23,480 --> 00:39:26,720
tend to operate in a dynamic
range, so the possible range

704
00:39:26,720 --> 00:39:29,710
of light that they can actually
capture tends to be much lower.

705
00:39:29,710 --> 00:39:34,500
Perhaps along the lines of about 10
stops, or maybe at a maximum 12 stops,

706
00:39:34,500 --> 00:39:37,690
and we're talking about some
really high end cameras here.

707
00:39:37,690 --> 00:39:41,530
You might recall from our discussion
earlier of the Philae lander

708
00:39:41,530 --> 00:39:43,530
that had some phenomenal
technology-- well,

709
00:39:43,530 --> 00:39:48,120
the Rosetta camera had some phenomenal
technology for the time period, 1998,

710
00:39:48,120 --> 00:39:52,000
and that actually has possible
14 stops of dynamic range.

711
00:39:52,000 --> 00:39:54,010
>> But this really implies
something about this

712
00:39:54,010 --> 00:39:57,350
that if we have some object, such
as the moon or a comet that's

713
00:39:57,350 --> 00:40:00,630
illuminated full on by
sunlight with any atmosphere

714
00:40:00,630 --> 00:40:05,700
especially to reflect some of that
light, then anything in the background

715
00:40:05,700 --> 00:40:08,270
is just going to be so
completely dark that we're not

716
00:40:08,270 --> 00:40:10,190
going to be able to see it.

717
00:40:10,190 --> 00:40:16,290
So this is sort of the primary reason
why a lot of these photographs have

718
00:40:16,290 --> 00:40:19,530
such harsh lighting is that there's
no atmosphere to reflect it and sort

719
00:40:19,530 --> 00:40:22,680
of fill in the gaps in the
crevices of the moon, for example,

720
00:40:22,680 --> 00:40:27,430
or the crevices of the comet, but also
because the stars that are actually

721
00:40:27,430 --> 00:40:30,870
within the night sky are so dark
relative to the ground that's being

722
00:40:30,870 --> 00:40:34,980
illuminated by the sun that they fall
away in exposure and we cannot actually

723
00:40:34,980 --> 00:40:37,410
see them whatsoever.

724
00:40:37,410 --> 00:40:40,760
>> So some terminology here,
there's underexposure,

725
00:40:40,760 --> 00:40:43,740
overexposure, sometimes
there's both, underexposure

726
00:40:43,740 --> 00:40:45,591
is when something's a
little bit too dark,

727
00:40:45,591 --> 00:40:47,340
you actually need to
increase the exposure

728
00:40:47,340 --> 00:40:49,280
to actually get all the details.

729
00:40:49,280 --> 00:40:52,690
Underexposure-- the hallmarks of it
is everything just looks way too dark,

730
00:40:52,690 --> 00:40:55,030
the shadow areas have
absolutely no detail.

731
00:40:55,030 --> 00:40:58,070
This one isn't horrendously
underexposed, but it's pretty bad.

732
00:40:58,070 --> 00:40:59,510
>> Overexposure is the opposite.

733
00:40:59,510 --> 00:41:02,020
You have overexposed
portions of your image

734
00:41:02,020 --> 00:41:05,790
and you've lost detail because it's
simply too bright for your sensor.

735
00:41:05,790 --> 00:41:09,800
You might need to change your exposure
values to compensate for that.

736
00:41:09,800 --> 00:41:12,960
And if you have both, we'll
you're just sort of out of luck.

737
00:41:12,960 --> 00:41:16,160
>> So one way to overcome these
issues, because frequently you

738
00:41:16,160 --> 00:41:19,930
will come into a compromise between
the capabilities of your camera

739
00:41:19,930 --> 00:41:24,620
and the amount that you can
actually vary these three exposure

740
00:41:24,620 --> 00:41:28,370
values and the amount of light that
exists in the scene so one of the best

741
00:41:28,370 --> 00:41:31,630
powers that you have, especially if
you're taking photographs outside

742
00:41:31,630 --> 00:41:34,630
is to just wait a little
while for better light.

743
00:41:34,630 --> 00:41:39,990
Generally midday light is really
harsh, it casts very harsh shadows,

744
00:41:39,990 --> 00:41:43,630
there's less atmosphere to actually
reflect and scatter some of the light

745
00:41:43,630 --> 00:41:47,420
and so it just tends to be
not a very good situation.

746
00:41:47,420 --> 00:41:49,650
If you're able to wait
even just a few hours,

747
00:41:49,650 --> 00:41:53,770
wait until dusk or if you're
able to do so, get up at dawn

748
00:41:53,770 --> 00:41:57,220
and you'll be rewarded
with wonderfully soft light

749
00:41:57,220 --> 00:42:01,480
that has a lot of color--
warm colors and tone

750
00:42:01,480 --> 00:42:07,300
that results from the light passing
through more of the atmosphere.

751
00:42:07,300 --> 00:42:11,350
>> Now very quickly, there's
this concept of metering,

752
00:42:11,350 --> 00:42:14,560
which is what the camera
actually does on our behalf

753
00:42:14,560 --> 00:42:19,500
to alter each of these
three exposure values

754
00:42:19,500 --> 00:42:22,270
and try to capture an appropriate image.

755
00:42:22,270 --> 00:42:25,410
And generally what the camera does
is it tries to take the entire scene

756
00:42:25,410 --> 00:42:27,370
and look at it in the
sort of middle gray.

757
00:42:27,370 --> 00:42:30,740
It tries to figure out what is the
middle tones, the middle brightness

758
00:42:30,740 --> 00:42:35,140
of the scene, and it will try to
expose your photograph for it.

759
00:42:35,140 --> 00:42:38,160
>> And typically there's some
additional fantastic goes into this,

760
00:42:38,160 --> 00:42:40,687
it will divide it into
a variety of zones

761
00:42:40,687 --> 00:42:43,520
and it will try to figure out in
which zone you've actually focused,

762
00:42:43,520 --> 00:42:45,710
and say OK that's probably
a very important zone

763
00:42:45,710 --> 00:42:49,780
and so it will apply some extra
weighting or priority to that zone

764
00:42:49,780 --> 00:42:52,520
and all that stuff is
fine, but this will still

765
00:42:52,520 --> 00:42:55,860
have the problem that even though
you might have some images that

766
00:42:55,860 --> 00:43:01,280
are being exposed to this middle
gray, the scene may not actually

767
00:43:01,280 --> 00:43:03,570
be appropriate for that.

768
00:43:03,570 --> 00:43:07,900
And so unless you're using
the absolute most manual mode

769
00:43:07,900 --> 00:43:11,440
available on your camera, you're
probably relying on your cameras meter

770
00:43:11,440 --> 00:43:15,972
to some degree to try to help
you pick these exposure values.

771
00:43:15,972 --> 00:43:17,680
And this means that
occasionally you need

772
00:43:17,680 --> 00:43:20,310
to do something called
exposure compensation to notify

773
00:43:20,310 --> 00:43:23,050
the camera that the scene
is actually a little bit

774
00:43:23,050 --> 00:43:26,180
different than its assumption.

775
00:43:26,180 --> 00:43:30,000
So in particular, if you have a
scene where there's a lot of snow,

776
00:43:30,000 --> 00:43:32,530
or a lot of white sand as
in the case of this image

777
00:43:32,530 --> 00:43:37,580
or it has a lot of dark areas, it's
a very shadowy, very dark alleyway

778
00:43:37,580 --> 00:43:39,830
or something like that, dark
at night and you actually

779
00:43:39,830 --> 00:43:42,750
need to notify the camera
that it needs to not

780
00:43:42,750 --> 00:43:45,630
expose for the very middle
you can apply some exposure

781
00:43:45,630 --> 00:43:48,240
compensation to overcome this issue.

782
00:43:48,240 --> 00:43:51,980
>> So in this example, the original
exposure that the camera wanted

783
00:43:51,980 --> 00:43:52,860
was on the left.

784
00:43:52,860 --> 00:43:57,310
Notice how it looks sort of dull
gray, it's not precisely what you want

785
00:43:57,310 --> 00:44:00,130
and I would suggest that this is
actually one of the best things

786
00:44:00,130 --> 00:44:02,400
that you can do to
improve your photography

787
00:44:02,400 --> 00:44:06,310
is to pay more attention to the exposure
compensation setting on your camera

788
00:44:06,310 --> 00:44:09,700
because most likely if you are taking a
scene in the snow, which is especially

789
00:44:09,700 --> 00:44:11,491
relevant for those of
us here in Cambridge,

790
00:44:11,491 --> 00:44:14,925
very soon it's going to start
to snow, or if you're outside

791
00:44:14,925 --> 00:44:16,800
and it's dark at night
then you actually have

792
00:44:16,800 --> 00:44:18,910
to apply some exposure compensation.

793
00:44:18,910 --> 00:44:22,390
>> So you apply exposure
compensation in stops

794
00:44:22,390 --> 00:44:25,390
and what you do is you tell the
camera to either increase or decrease

795
00:44:25,390 --> 00:44:29,530
the exposure compensation based
on its assumption of middle gray,

796
00:44:29,530 --> 00:44:33,160
in this case, I know that because
the scene was going to be brighter

797
00:44:33,160 --> 00:44:35,470
than the camera was
expecting it I needed

798
00:44:35,470 --> 00:44:39,670
to actually tell it to increase
the exposure compensation,

799
00:44:39,670 --> 00:44:44,430
so by adding a positive 1 stop of
exposure value of exposure compensation

800
00:44:44,430 --> 00:44:47,770
I told the camera that it's actually
brighter than it was anticipating

801
00:44:47,770 --> 00:44:51,910
and would then take a
properly exposed photograph.

802
00:44:51,910 --> 00:44:55,320
Likewise, we might have a
scene that was too dark.

803
00:44:55,320 --> 00:44:58,560
For example, if you are trying
to take an image of someone who's

804
00:44:58,560 --> 00:45:01,690
wearing a dark coat for example then
it might actually confuse the camera

805
00:45:01,690 --> 00:45:03,690
into making everything
little bit too bright,

806
00:45:03,690 --> 00:45:06,650
you might need to dial in some
negative exposure compensation

807
00:45:06,650 --> 00:45:08,930
to overcome this issue.

808
00:45:08,930 --> 00:45:12,200
>> Now many cameras have a wide
variety of metering modes.

809
00:45:12,200 --> 00:45:15,820
In fact, what you will find is
that the simpler the camera,

810
00:45:15,820 --> 00:45:18,200
the cheaper the camera
the more modes it has

811
00:45:18,200 --> 00:45:21,160
and this is just ridiculous
what they've gone through.

812
00:45:21,160 --> 00:45:24,710
I've seen cameras now of course
there is like a self portrait mode,

813
00:45:24,710 --> 00:45:29,230
but they have a party mode, candlelight
mode, a sunset mode, fireworks mode,

814
00:45:29,230 --> 00:45:30,965
beach mode, snow mode.

815
00:45:30,965 --> 00:45:35,600
I saw one camera that had a beach
mode and the beach two mode,

816
00:45:35,600 --> 00:45:38,440
so I have no idea what the
difference between those two was,

817
00:45:38,440 --> 00:45:39,670
but it doesn't matter.

818
00:45:39,670 --> 00:45:41,630
You don't really need
any of those modes,

819
00:45:41,630 --> 00:45:46,680
because the vast majority of the time
they do nothing special to the camera,

820
00:45:46,680 --> 00:45:50,860
to the settings in the camera, other
than altering these three exposure

821
00:45:50,860 --> 00:45:51,474
values.

822
00:45:51,474 --> 00:45:53,890
So if you just sort of think
about what you might want out

823
00:45:53,890 --> 00:45:56,570
of that particular image, you
could overcome those issues

824
00:45:56,570 --> 00:46:00,780
and use one of the simpler, one
of the more raw metering modes

825
00:46:00,780 --> 00:46:05,050
so that you can actually take photos
with a great deal more control.

826
00:46:05,050 --> 00:46:07,060
So for example, in a
portrait you might actually

827
00:46:07,060 --> 00:46:09,930
want to isolate your subject
from the background, which

828
00:46:09,930 --> 00:46:13,270
would mean decreasing the f-number
or having a very large aperture,

829
00:46:13,270 --> 00:46:17,262
so you get very nice background
blur from them or within that shot,

830
00:46:17,262 --> 00:46:18,720
and so that would be your priority.

831
00:46:18,720 --> 00:46:21,580
And that's precisely what the
portrait modes in these cameras do,

832
00:46:21,580 --> 00:46:24,220
is it tries to make the
apertures as large as possible

833
00:46:24,220 --> 00:46:29,280
and alters the other
settings as a result.

834
00:46:29,280 --> 00:46:30,210
>> OK.

835
00:46:30,210 --> 00:46:33,990
So let's go into a completely different
direction and talk a little bit more

836
00:46:33,990 --> 00:46:36,960
about the digital aspect
of digital cameras

837
00:46:36,960 --> 00:46:39,764
and just talk very quickly
about sensors and some

838
00:46:39,764 --> 00:46:41,930
of the different technologies
and some of the things

839
00:46:41,930 --> 00:46:45,060
that actually impact
us as photographers.

840
00:46:45,060 --> 00:46:48,870
I'd alluded to dynamic range
before and we can think of sensors

841
00:46:48,870 --> 00:46:54,760
as being an array of bucket that
capture light in the form of raindrops.

842
00:46:54,760 --> 00:46:57,980
>> So imagine we set out an
array of buckets outside

843
00:46:57,980 --> 00:47:03,080
and they're going to capture rain, and
we can then measure the amount of rain

844
00:47:03,080 --> 00:47:05,080
in each of those buckets
and that is our image,

845
00:47:05,080 --> 00:47:08,870
so-called, and we can take
this analogy quite far

846
00:47:08,870 --> 00:47:11,470
and it's actually a
relatively good analogy

847
00:47:11,470 --> 00:47:15,570
because it alludes to a number of
things within the digital camera.

848
00:47:15,570 --> 00:47:17,040
Imagine a couple of scenarios.

849
00:47:17,040 --> 00:47:21,280
First of all, imagine what might happen
if we allow rain or photons to actually

850
00:47:21,280 --> 00:47:25,150
fall into our bucket and not
a lot to actually falls there.

851
00:47:25,150 --> 00:47:27,750
Now imagine that we have some
sort of way of measuring this,

852
00:47:27,750 --> 00:47:30,650
if we have some measurement
that's not accurate enough

853
00:47:30,650 --> 00:47:34,962
to measure the small amount of water
that we've actually collected then

854
00:47:34,962 --> 00:47:37,170
it's indistinguishable from
noise, we're not actually

855
00:47:37,170 --> 00:47:39,490
going to be able to measure
that as any sort of signal.

856
00:47:39,490 --> 00:47:42,760
>> And so we'll perhaps guess as
to the value that is actually

857
00:47:42,760 --> 00:47:45,760
appropriate for that
small amount of white.

858
00:47:45,760 --> 00:47:49,920
This alludes to this problem of sensors
that do not collect enough photons

859
00:47:49,920 --> 00:47:52,060
and it's just too dark
and so there's noise

860
00:47:52,060 --> 00:47:54,550
in these dark regions of the image.

861
00:47:54,550 --> 00:47:58,380
Likewise, if we allow too much to
collect into this bucket it might fill

862
00:47:58,380 --> 00:48:01,660
up and actually overflow
and so beyond that point

863
00:48:01,660 --> 00:48:05,320
we have no way of measuring or
knowing how much rain has precisely

864
00:48:05,320 --> 00:48:09,610
fallen within this bucket, we just
know that it's beyond the maximum.

865
00:48:09,610 --> 00:48:12,980
That's precisely what happens in these
buckets as well, or in these pixels

866
00:48:12,980 --> 00:48:17,160
as well, is that once we've
gotten to their maximum of voltage

867
00:48:17,160 --> 00:48:20,155
then it's not actually possible
to get any more detail out of that

868
00:48:20,155 --> 00:48:22,560
and we would get an overexposure.

869
00:48:22,560 --> 00:48:25,270
>> We can actually take this
analogy just a little bit further

870
00:48:25,270 --> 00:48:27,420
if you imagine again
this array of buckets

871
00:48:27,420 --> 00:48:29,340
that are sitting next to each other.

872
00:48:29,340 --> 00:48:31,270
One of these buckets
fills up with water.

873
00:48:31,270 --> 00:48:34,850
You can imagine it might spill
over into neighboring buckets,

874
00:48:34,850 --> 00:48:38,630
and this concept is known as
blooming within a digital camera

875
00:48:38,630 --> 00:48:42,640
and we actually see this in a wide
variety of circumstances where

876
00:48:42,640 --> 00:48:48,710
a very, very bright section of the
scene that is extremely overexposed

877
00:48:48,710 --> 00:48:54,380
will actually bleed some of its data
over to the neighboring pixels as well

878
00:48:54,380 --> 00:48:57,570
and cause those to become
overexposed as well, which

879
00:48:57,570 --> 00:48:59,730
is kind of an interesting phenomenon.

880
00:48:59,730 --> 00:49:02,460
>> Now imagine that we're
actually able to take

881
00:49:02,460 --> 00:49:05,300
a division between the
maximum amount of volume

882
00:49:05,300 --> 00:49:07,150
that we're actually
able to measure here,

883
00:49:07,150 --> 00:49:10,160
our full well capacity,
our full bucket capacity,

884
00:49:10,160 --> 00:49:13,600
divided by the smallest possible signal.

885
00:49:13,600 --> 00:49:16,807
This would be our dynamic
range and one of the ways,

886
00:49:16,807 --> 00:49:19,890
there's variety of ways that we can
improve the dynamic range for a camera

887
00:49:19,890 --> 00:49:23,270
and what this essentially says is the
possible range, this range that we were

888
00:49:23,270 --> 00:49:27,500
alluding to before, that allows us to
specify how much or how little light

889
00:49:27,500 --> 00:49:30,414
we can actually capture with our camera.

890
00:49:30,414 --> 00:49:32,830
So there's a variety of ways
to improve this dynamic range

891
00:49:32,830 --> 00:49:33,705
as you might imagine.

892
00:49:33,705 --> 00:49:36,620
One of them is to have a
bigger bucket-- actually

893
00:49:36,620 --> 00:49:39,180
allow us to capture a fuller signal.

894
00:49:39,180 --> 00:49:42,910
Another way to do this is to
minimize the detectable signal,

895
00:49:42,910 --> 00:49:46,250
to actually decrease the
amount of noise that we get out

896
00:49:46,250 --> 00:49:50,910
of the electronics of
this particular sensor,

897
00:49:50,910 --> 00:49:53,110
and some of the
advancements in recent years

898
00:49:53,110 --> 00:49:56,020
have, in fact, been to
decrease the smallest

899
00:49:56,020 --> 00:50:00,650
detectable signal within
the sensor and then

900
00:50:00,650 --> 00:50:03,740
we are able to improve our
dynamic range and get improvements

901
00:50:03,740 --> 00:50:06,960
within our photographs.

902
00:50:06,960 --> 00:50:10,190
>> Now one of the other really important
things to realize with digital cameras

903
00:50:10,190 --> 00:50:12,740
is that they come in a
variety of sensor sizes

904
00:50:12,740 --> 00:50:14,820
and so there's a wide variety of sizes.

905
00:50:14,820 --> 00:50:18,060
One of the great things
of modern digital cameras

906
00:50:18,060 --> 00:50:22,560
is that we are seeing larger and larger
sensors in smaller and smaller cameras,

907
00:50:22,560 --> 00:50:26,070
but there's a wide variety of
things that this actually impacts,

908
00:50:26,070 --> 00:50:30,250
not the least of which is the way
that focal length will actually

909
00:50:30,250 --> 00:50:34,600
change the field of view depending
on the size of the sensor.

910
00:50:34,600 --> 00:50:38,760
So imagine, just for minutes, and sort
of a teaser for what you should look

911
00:50:38,760 --> 00:50:41,350
into after this seminar
is actually over--

912
00:50:41,350 --> 00:50:44,310
imagine that we have a lens that
because it's circular projects

913
00:50:44,310 --> 00:50:47,810
this circular image on to
some location and imagine

914
00:50:47,810 --> 00:50:51,130
we have a sensor that's relatively
large and captures as much

915
00:50:51,130 --> 00:50:55,820
of this area as possible, in
this case our red sensor here.

916
00:50:55,820 --> 00:50:59,190
>> Now imagine we have a smaller
sensor, this blue sensor that

917
00:50:59,190 --> 00:51:01,710
captures the center
portion of this image.

918
00:51:01,710 --> 00:51:04,560
If you blow both up to be
approximately the same size you'll

919
00:51:04,560 --> 00:51:07,230
notice at the blue sensor
seems to be a crop,

920
00:51:07,230 --> 00:51:09,380
it seems to be this
center portion and it

921
00:51:09,380 --> 00:51:12,360
makes it look like you're
using a larger focal length

922
00:51:12,360 --> 00:51:14,340
lens than you actually are.

923
00:51:14,340 --> 00:51:17,600
So for this reason, as we
shrink the size of sensors

924
00:51:17,600 --> 00:51:23,030
we also have to shrink the size
and the focal length of our lenses

925
00:51:23,030 --> 00:51:26,120
in order to compensate for that
change in the field of view.

926
00:51:26,120 --> 00:51:29,070
And as you might recall from
our discussion about aperture

927
00:51:29,070 --> 00:51:31,290
just a few minutes ago,
this means that we also

928
00:51:31,290 --> 00:51:37,070
have to change the diameter of our
aperture to maintain the same f-number.

929
00:51:37,070 --> 00:51:41,795
>> So we can go on and on to a wide variety
of topics in sensor sizes and all

930
00:51:41,795 --> 00:51:44,670
these things, but this is really
just a teaser for some of the things

931
00:51:44,670 --> 00:51:47,047
that you might actually
start looking into.

932
00:51:47,047 --> 00:51:49,130
When we start talking about
this a little bit more

933
00:51:49,130 --> 00:51:51,380
we start talking about 35
millimeter equivalency.

934
00:51:51,380 --> 00:51:58,400
We might have some sort of
reference size of a digital sensor

935
00:51:58,400 --> 00:52:01,440
that we're able to compare
other sensors to in order

936
00:52:01,440 --> 00:52:05,635
to discuss our focal lengths
in a more meaningful way

937
00:52:05,635 --> 00:52:09,530
and so I certainly suggest that you
start doing your research in that area

938
00:52:09,530 --> 00:52:11,830
if you're interested in
doing that, but for now it

939
00:52:11,830 --> 00:52:14,360
seems like I've run out of time
and we'll have to sign off.

940
00:52:14,360 --> 00:52:17,440
>> So I want to thank you
all very much for viewing.

941
00:52:17,440 --> 00:52:19,779
I'll post the slides that
we have here online and also

942
00:52:19,779 --> 00:52:22,070
that handout that allows you
to understand a little bit

943
00:52:22,070 --> 00:52:24,924
more the mathematics
behind the wacky f-numbers,

944
00:52:24,924 --> 00:52:26,840
and I do encourage you
to take a look at that.

945
00:52:26,840 --> 00:52:29,631
And so thank you very much for
watching and I hope to see you soon.

946
00:52:29,631 --> 00:52:32,510

947
00:52:32,510 --> 00:52:33,010
Oh.

948
00:52:33,010 --> 00:52:34,490
Thank you, thank you.

949
00:52:34,490 --> 00:52:37,210
The illustrious audiences enjoys it.

950
00:52:37,210 --> 00:52:38,827