1
00:00:00,000 --> 00:00:03,820
these ruins our tourist attractions are
chances are there really is a picture
2
00:00:03,820 --> 00:00:06,960
postcard somewhere that portrays them
hopefully is printed with a little
3
00:00:06,960 --> 00:00:11,420
better filing this regional currently is
that's ok we're going to fix it here in
4
00:00:11,420 --> 00:00:16,059
the first session are Simon always is to
seek out colors that can't possibly be
5
00:00:16,059 --> 00:00:21,710
right colors like for example blue sea
and under Hot Desert Sun this step
6
00:00:21,710 --> 00:00:25,269
deciding whether we're going to do that
deciding whether there are any such
7
00:00:25,269 --> 00:00:29,990
colors has to be done quickly we're
going to go away over time limit to get
8
00:00:29,990 --> 00:00:35,260
the image done I want to be able to
analyze this image and the weather curve
9
00:00:35,260 --> 00:00:39,710
has to be applied to it within 10
seconds in order to do that I think
10
00:00:39,710 --> 00:00:44,070
we've got to be thinking in leb I don't
think we can evaluate the image that
11
00:00:44,070 --> 00:00:49,850
rapidly in RGB for reasons that I'm bout
to explain how do we do this we look for
12
00:00:49,850 --> 00:00:53,520
colors that we know certain callers were
not going to be able to know like what
13
00:00:53,520 --> 00:00:58,469
color is this woman's jacket I have no
clue could be anything but the clouds
14
00:00:58,469 --> 00:01:02,940
now the clouds in the sky they're not
very well defined what you think they're
15
00:01:02,940 --> 00:01:06,270
probably pretty close to being white and
here you see on the right side of the
16
00:01:06,270 --> 00:01:11,100
Info palette do not require a blue also
little bit green the sky itself as a
17
00:01:11,100 --> 00:01:16,700
little bit more improving the clouds are
silent think this problem wrong part me
18
00:01:16,700 --> 00:01:18,710
i didnt explain mellanby died
19
00:01:18,710 --> 00:01:25,750
well let's get right to that let me
minimize the image years
20
00:01:25,750 --> 00:01:32,940
a schematic of how leb works good thing
about evaluating callers in leb is that
21
00:01:32,940 --> 00:01:37,390
you can ignore the L channel because the
other channel is detail on me so we only
22
00:01:37,390 --> 00:01:40,870
have two numbers to look at the aim to
be there are the ones that defined color
23
00:01:40,870 --> 00:01:45,440
I've got a couple videos in the book on
this subject so he need more definite
24
00:01:45,440 --> 00:01:50,020
there's plenty of sources for it but for
the time being
25
00:01:50,020 --> 00:01:56,340
for the purposes of evaluating whether
the current colors are right I think
26
00:01:56,340 --> 00:02:00,090
it's pretty easy to understand once you
get used to it there are two channels
27
00:02:00,090 --> 00:02:06,850
the a and the B that we're looking at
each one if the value was negative it's
28
00:02:06,850 --> 00:02:12,600
a cold color that's green and the a
channel ruin the beach channel if it's a
29
00:02:12,600 --> 00:02:17,590
positive number that's a warm color
magenta in the AGM yeah when the beach
30
00:02:17,590 --> 00:02:22,970
and if it's a value of 0 that would be
neither one either magellan or green and
31
00:02:22,970 --> 00:02:29,450
the a channel either yellow or blue in
the EB channel so 430 we have a values
32
00:02:29,450 --> 00:02:34,670
Uruguay 0 be that's a neutral could be a
white could be a black doesn't matter
33
00:02:34,670 --> 00:02:38,470
because we don't take this into
consideration in the end the beach or
34
00:02:38,470 --> 00:02:43,959
could be great so if the clouds that we
are interested in we're truly white
35
00:02:43,959 --> 00:02:50,380
truly unique role than you would expect
them to be 080 be now if they weren't
36
00:02:50,380 --> 00:02:55,870
exactly 080 be you'd think they would
probably be negative not warm little bit
37
00:02:55,870 --> 00:03:01,340
cold like the rest the sky is ok how do
we use this information while there
38
00:03:01,340 --> 00:03:06,430
certainly areas that come up over and
over again certain types of telltale
39
00:03:06,430 --> 00:03:10,750
regions of the image that we can look at
and determine whether the A&B values are
40
00:03:10,750 --> 00:03:13,150
correct they are as follows
41
00:03:13,150 --> 00:03:19,110
neutral says we stay in beer both 0 if
they're not exactly zero then negative
42
00:03:19,110 --> 00:03:23,709
numbers are called the positive numbers
are warm so these clouds if they weren't
43
00:03:23,709 --> 00:03:29,530
0 you expect them to be negative words
if we go back to the desert image you
44
00:03:29,530 --> 00:03:30,300
would think that
45
00:03:30,300 --> 00:03:34,770
that although the same was nearly
neutral it probably would show positive
46
00:03:34,770 --> 00:03:40,710
numbers in both cases being the beach
animals are both positive because of
47
00:03:40,710 --> 00:03:45,670
faces more imaginative is green and
faces more yellow than his blue and with
48
00:03:45,670 --> 00:03:51,040
the exception of small children every
pic ski skins and light skinned
49
00:03:51,040 --> 00:03:54,530
caucasians chiefly blondes you would
expect that the B would be higher than
50
00:03:54,530 --> 00:03:59,500
the a natural greenery which will see a
lot of the picture-postcard workflow the
51
00:03:59,500 --> 00:04:03,920
A's negative more green than with Jana
kabhi is positive more yellow than blue
52
00:04:03,920 --> 00:04:08,140
and usually the B is further away from
zero then the ages and sometimes I
53
00:04:08,140 --> 00:04:13,440
double what the value of the A's and
skies skies are always blue sotheby's
54
00:04:13,440 --> 00:04:18,660
gonna be negative big-time the aid is
probably pretty close to zero could be
55
00:04:18,660 --> 00:04:24,380
off a little bit because in the aids
your would be neither magenta more green
56
00:04:24,380 --> 00:04:28,600
now typically in disguise the A's little
bit negative little bit more green that
57
00:04:28,600 --> 00:04:31,050
is sometimes the other way around
58
00:04:31,050 --> 00:04:40,169
ok so let's go back to our image and see
if we can analyze it in this way here we
59
00:04:40,169 --> 00:04:44,880
are back again what is it that we know
about this image what colors can we
60
00:04:44,880 --> 00:04:49,900
stayed for a fact what they should be
well this guy should be some kind and it
61
00:04:49,900 --> 00:04:54,780
is the clouds are we really saw midas
4-12 that's a little bit green blue and
62
00:04:54,780 --> 00:05:00,570
pink and comparison what I would have in
mind now greens we said the target is
63
00:05:00,570 --> 00:05:06,330
going to be a negative be positive and
be a little bit higher than the ASO down
64
00:05:06,330 --> 00:05:15,100
here minus 13 + 29 works over here in
the shadow same deal minus nine give me
65
00:05:15,100 --> 00:05:20,740
a plus 12 in the B and the beach
channels for the way from 08 is that
66
00:05:20,740 --> 00:05:21,740
looks good too
67
00:05:21,740 --> 00:05:26,080
about the shadow region you think that
would probably be fairly neutral
68
00:05:26,080 --> 00:05:32,090
underneath the arches and neutral would
be 00 and that's pretty close pretty
69
00:05:32,090 --> 00:05:36,419
close minus one plus one that's close
enough and how about the church itself
70
00:05:36,419 --> 00:05:38,470
that's got to be brown
71
00:05:38,470 --> 00:05:44,520
what we what we know about that if it's
truly brown brown as a species of red
72
00:05:44,520 --> 00:05:53,300
and red would be positive a positive be
what if it's not exactly a red Real
73
00:05:53,300 --> 00:06:00,480
Redwood be equal a will be if this
church is not that possibly quickly but
74
00:06:00,480 --> 00:06:04,750
if it's not then would you expect that
it's on the DL side of red would you
75
00:06:04,750 --> 00:06:09,310
expect this on purpose I'm afraid I
would have to think it should be on the
76
00:06:09,310 --> 00:06:15,600
young side of ready for anything so the
numbers I'll be looking for either equal
77
00:06:15,600 --> 00:06:21,300
a will be a really doubt that but I sort
of expected both the A&B with positive
78
00:06:21,300 --> 00:06:24,980
and the people behind us let's see
whether that's true or not K plus two
79
00:06:24,980 --> 00:06:29,560
plus seven thats good plus five plus
nine that's what I expect those five +
80
00:06:29,560 --> 00:06:37,730
10 + 50 + 14 + three-plus 5 that looks
good too so as far as I can see the only
81
00:06:37,730 --> 00:06:41,700
issue here is that the clouds are not
really the right color time for
82
00:06:41,700 --> 00:06:54,020
adjustment layer how we gonna him down
the red
83
00:06:54,020 --> 00:07:08,050
now it seems to me that the rest of the
sky's getting to green minus 8-22 that
100:00:00,000--> 00:00:03,820我们的旅游景点在这些废墟很有可能真的有一张图片200:00:03,820--> 00:00:06,960明信片在某个地方,描绘了他们希望被印有一点点300:00:06,960--> 00:00:11,420更好目前归档这区域是没关系,我们打算去修理它在400:00:11,420--> 00:00:16,059第一届会议是 Simon 总是对找出不可能的颜色500:00:16,059--> 00:00:21,710正确的颜色,例如喜欢蓝色的大海根据热沙漠太阳这一步600:00:21,710--> 00:00:25,269决定我们是否要这样做决定是否有任何此类700:00:25,269--> 00:00:29,990颜色有快速做我们要超过时间限制,要去走800:00:29,990--> 00:00:35,260图像做想要能够分析这幅图像和天气曲线900:00:35,260--> 00:00:39,710具有要应用于它在 10 内为此,我认为秒1000:00:39,710--> 00:00:44,070我们得在 leb 我不思考相信我们可以衡量图像,1100:00:44,070--> 00:00:49,850迅速在 RGB 的原因,我是布特解释如何做到这,我们寻找1200:00:49,850--> 00:00:53,520我们知道某些调用方的颜色了不会能够知道喜欢什么1300:00:53,520--> 00:00:58,469颜色是我没有这个女人的夹克线索能云呢?1400:00:58,469 --> 00:01:02,940now the clouds in the sky they're notvery well defined what you think they're1500:01:02,940 --> 00:01:06,270probably pretty close to being white andhere you see on the right side of the1600:01:06,270 --> 00:01:11,100Info palette do not require a blue alsolittle bit green the sky itself as a1700:01:11,100 --> 00:01:16,700little bit more improving the clouds aresilent think this problem wrong part me1800:01:16,700 --> 00:01:18,710i didnt explain mellanby died1900:01:18,710 --> 00:01:25,750well let's get right to that let meminimize the image years2000:01:25,750 --> 00:01:32,940a schematic of how leb works good thingabout evaluating callers in leb is that2100:01:32,940 --> 00:01:37,390you can ignore the L channel because theother channel is detail on me so we only2200:01:37,390 --> 00:01:40,870have two numbers to look at the aim tobe there are the ones that defined color2300:01:40,870 --> 00:01:45,440I've got a couple videos in the book onthis subject so he need more definite2400:01:45,440 --> 00:01:50,020there's plenty of sources for it but forthe time being2500:01:50,020 --> 00:01:56,340for the purposes of evaluating whetherthe current colors are right I think2600:01:56,340 --> 00:02:00,090it's pretty easy to understand once youget used to it there are two channels2700:02:00,090 --> 00:02:06,850the a and the B that we're looking ateach one if the value was negative it's2800:02:06,850 --> 00:02:12,600a cold color that's green and the achannel ruin the beach channel if it's a2900:02:12,600 --> 00:02:17,590positive number that's a warm colormagenta in the AGM yeah when the beach3000:02:17,590 --> 00:02:22,970and if it's a value of 0 that would beneither one either magellan or green and3100:02:22,970 --> 00:02:29,450the a channel either yellow or blue inthe EB channel so 430 we have a values3200:02:29,450 --> 00:02:34,670Uruguay 0 be that's a neutral could be awhite could be a black doesn't matter3300:02:34,670 --> 00:02:38,470because we don't take this intoconsideration in the end the beach or3400:02:38,470 --> 00:02:43,959could be great so if the clouds that weare interested in we're truly white3500:02:43,959 --> 00:02:50,380truly unique role than you would expectthem to be 080 be now if they weren't3600:02:50,380 --> 00:02:55,870exactly 080 be you'd think they wouldprobably be negative not warm little bit3700:02:55,870 --> 00:03:01,340cold like the rest the sky is ok how dowe use this information while there3800:03:01,340 --> 00:03:06,430certainly areas that come up over andover again certain types of telltale3900:03:06,430 --> 00:03:10,750regions of the image that we can look atand determine whether the A&B values are4000:03:10,750 --> 00:03:13,150correct they are as follows4100:03:13,150 --> 00:03:19,110neutral says we stay in beer both 0 ifthey're not exactly zero then negative4200:03:19,110 --> 00:03:23,709numbers are called the positive numbersare warm so these clouds if they weren't4300:03:23,709 --> 00:03:29,5300 you expect them to be negative wordsif we go back to the desert image you4400:03:29,530 --> 00:03:30,300would think that4500:03:30,300 --> 00:03:34,770that although the same was nearlyneutral it probably would show positive4600:03:34,770 --> 00:03:40,710numbers in both cases being the beachanimals are both positive because of4700:03:40,710 --> 00:03:45,670faces more imaginative is green andfaces more yellow than his blue and with4800:03:45,670 --> 00:03:51,040the exception of small children everypic ski skins and light skinned4900:03:51,040 --> 00:03:54,530caucasians chiefly blondes you wouldexpect that the B would be higher than5000:03:54,530 --> 00:03:59,500the a natural greenery which will see alot of the picture-postcard workflow the5100:03:59,500 --> 00:04:03,920A's negative more green than with Janakabhi is positive more yellow than blue5200:04:03,920 --> 00:04:08,140and usually the B is further away fromzero then the ages and sometimes I5300:04:08,140 --> 00:04:13,440double what the value of the A's andskies skies are always blue sotheby's5400:04:13,440 --> 00:04:18,660gonna be negative big-time the aid isprobably pretty close to zero could be5500:04:18,660 --> 00:04:24,380off a little bit because in the aidsyour would be neither magenta more green5600:04:24,380 --> 00:04:28,600now typically in disguise the A's littlebit negative little bit more green that5700:04:28,600 --> 00:04:31,050is sometimes the other way around5800:04:31,050 --> 00:04:40,169ok so let's go back to our image and seeif we can analyze it in this way here we5900:04:40,169 --> 00:04:44,880are back again what is it that we knowabout this image what colors can we6000:04:44,880 --> 00:04:49,900stayed for a fact what they should bewell this guy should be some kind and it6100:04:49,900 --> 00:04:54,780is the clouds are we really saw midas4-12 that's a little bit green blue and6200:04:54,780 --> 00:05:00,570pink and comparison what I would have inmind now greens we said the target is6300:05:00,570 --> 00:05:06,330going to be a negative be positive andbe a little bit higher than the ASO down6400:05:06,330 --> 00:05:15,100here minus 13 + 29 works over here inthe shadow same deal minus nine give me6500:05:15,100 --> 00:05:20,740a plus 12 in the B and the beachchannels for the way from 08 is that6600:05:20,740 --> 00:05:21,740looks good too6700:05:21,740 --> 00:05:26,080about the shadow region you think thatwould probably be fairly neutral6800:05:26,080 --> 00:05:32,090underneath the arches and neutral wouldbe 00 and that's pretty close pretty6900:05:32,090 --> 00:05:36,419close minus one plus one that's closeenough and how about the church itself7000:05:36,419 --> 00:05:38,470that's got to be brown7100:05:38,470 --> 00:05:44,520what we what we know about that if it'struly brown brown as a species of red7200:05:44,520 --> 00:05:53,300and red would be positive a positive bewhat if it's not exactly a red Real7300:05:53,300 --> 00:06:00,480Redwood be equal a will be if thischurch is not that possibly quickly but7400:06:00,480 --> 00:06:04,750if it's not then would you expect thatit's on the DL side of red would you7500:06:04,750 --> 00:06:09,310expect this on purpose I'm afraid Iwould have to think it should be on the7600:06:09,310 --> 00:06:15,600young side of ready for anything so thenumbers I'll be looking for either equal7700:06:15,600 --> 00:06:21,300a will be a really doubt that but I sortof expected both the A&B with positive7800:06:21,300 --> 00:06:24,980and the people behind us let's seewhether that's true or not K plus two7900:06:24,980 --> 00:06:29,560plus seven thats good plus five plusnine that's what I expect those five +8000:06:29,560 --> 00:06:37,73010 + 50 + 14 + three-plus 5 that looksgood too so as far as I can see the only8100:06:37,730 --> 00:06:41,700issue here is that the clouds are notreally the right color time for8200:06:41,700 --> 00:06:54,020adjustment layer how we gonna him downthe red8300:06:54,020 --> 00:07:08,050now it seems to me that the rest of thesky's getting to green minus 8-22 that
正在翻譯中..