Would you like to inspect the original subtitles? These are the user uploaded subtitles that are being translated:
1
00:00:07,166 --> 00:00:08,766
welcome back to Nuke 2:04
2
00:00:08,766 --> 00:00:10,200
this is class number two
3
00:00:10,200 --> 00:00:11,333
and this is the shot we'll be
4
00:00:11,333 --> 00:00:13,966
using this week to learn the lens distortion
5
00:00:13,966 --> 00:00:16,400
and camera tracking tools in Nuke X
6
00:00:16,733 --> 00:00:18,466
um pretty shaky camera
7
00:00:18,466 --> 00:00:20,133
but not untrackable of course
8
00:00:20,133 --> 00:00:21,200
and a good example
9
00:00:21,200 --> 00:00:22,333
and it's a nice long shot too
10
00:00:22,333 --> 00:00:23,800
so once you have a camera track for this time
11
00:00:23,800 --> 00:00:25,400
there's lots of possibilities
12
00:00:25,800 --> 00:00:28,300
of um adding things to the shot
13
00:00:28,300 --> 00:00:29,700
I think one good thing to do for those
14
00:00:29,700 --> 00:00:31,166
you building a real would be
15
00:00:31,200 --> 00:00:32,700
extending the height of the building
16
00:00:32,700 --> 00:00:33,733
at the end of the shot
17
00:00:33,733 --> 00:00:34,966
so for example
18
00:00:36,400 --> 00:00:37,866
once you have the camera track information
19
00:00:37,866 --> 00:00:40,200
you could make this building go all the way off frame
20
00:00:40,200 --> 00:00:41,633
rather than having it
21
00:00:42,466 --> 00:00:44,200
be sure that would be a kind of a cool
22
00:00:44,200 --> 00:00:46,333
cool challenge and definitely
23
00:00:46,333 --> 00:00:47,300
be good enough for your real
24
00:00:47,300 --> 00:00:50,533
so I think that's what I put back one more time
25
00:00:50,533 --> 00:00:51,766
and then we'll go ahead and
26
00:00:52,066 --> 00:00:54,600
and please notice these two guys walking in the frame
27
00:00:54,600 --> 00:00:56,266
because we'll have to deal with them in a minute
28
00:00:56,700 --> 00:00:58,666
otherwise everything else is totally static
29
00:00:59,500 --> 00:01:01,800
except for the camera of course alright
30
00:01:01,800 --> 00:01:04,500
so let's go ahead and get into nuke and start using
31
00:01:04,500 --> 00:01:05,866
with the lens to start the
32
00:01:05,866 --> 00:01:08,233
with the lens distortion tool so that we can make our
33
00:01:08,266 --> 00:01:09,466
track that much easier
34
00:01:11,266 --> 00:01:12,700
I've already gone ahead and
35
00:01:13,066 --> 00:01:14,466
wrote it off our two
36
00:01:15,466 --> 00:01:16,866
people walking through the scene here
37
00:01:17,366 --> 00:01:18,766
now but this guy here
38
00:01:18,766 --> 00:01:20,366
he's in the center of frame which
39
00:01:20,500 --> 00:01:22,500
usually has less distortion than everything else anyway
40
00:01:22,500 --> 00:01:23,533
unless there's a very weird lens
41
00:01:23,533 --> 00:01:24,666
that almost always will
42
00:01:24,900 --> 00:01:25,800
and he's very tiny
43
00:01:25,800 --> 00:01:29,266
so I'm being a little anal by roto hang him off um
44
00:01:29,266 --> 00:01:30,600
and spending the time to do that
45
00:01:31,066 --> 00:01:33,266
but it's one of the situations where it looks
46
00:01:33,266 --> 00:01:34,500
only takes me a couple minutes to roto
47
00:01:34,500 --> 00:01:37,000
and if he does end up impacting the
48
00:01:37,133 --> 00:01:38,400
calculation of one's distortion
49
00:01:38,400 --> 00:01:39,600
and I'm troubleshooting it later
50
00:01:39,600 --> 00:01:40,366
it could take me hours
51
00:01:40,366 --> 00:01:44,266
so it's worth rotoing anything that um
52
00:01:44,266 --> 00:01:45,466
moves in any way
53
00:01:45,700 --> 00:01:46,466
and with this guy here
54
00:01:46,466 --> 00:01:48,433
he does get big enough and he actually gets to the
55
00:01:48,466 --> 00:01:49,200
corner of frame
56
00:01:49,200 --> 00:01:50,900
which is where most of the distortion usually happens
57
00:01:50,900 --> 00:01:52,400
that he could absolutely affect
58
00:01:52,400 --> 00:01:54,400
the ones that torsion map that nuke generates
59
00:01:54,400 --> 00:01:56,666
so I'm gonna go ahead and mask both those guys off
60
00:01:56,666 --> 00:01:57,566
and that's already done
61
00:01:57,566 --> 00:01:58,766
cause we covered Rhodell last week
62
00:01:58,766 --> 00:02:01,200
there was no reason for me to obviously do that again
63
00:02:01,300 --> 00:02:02,300
and show you guys how to do that
64
00:02:02,300 --> 00:02:04,866
so our alpha channel is just those two rectangles
65
00:02:07,166 --> 00:02:08,500
now let's go ahead with that being said
66
00:02:08,500 --> 00:02:09,733
bring in our one's distortion tool
67
00:02:09,733 --> 00:02:11,233
which is under the transform menu
68
00:02:13,900 --> 00:02:17,333
and you'll see that we have two inputs
69
00:02:17,333 --> 00:02:18,633
source and mask
70
00:02:18,700 --> 00:02:19,800
I'm gonna plug
71
00:02:20,400 --> 00:02:22,266
actually I can leave just the mask unplugged
72
00:02:22,266 --> 00:02:23,466
but when I do use a mask
73
00:02:23,466 --> 00:02:24,600
I like to plug it in
74
00:02:24,866 --> 00:02:26,600
to something so that it's not just hanging out
75
00:02:26,600 --> 00:02:27,000
here that way
76
00:02:27,000 --> 00:02:27,600
anyone who looks
77
00:02:27,600 --> 00:02:29,466
at my script will know that if there's no mask
78
00:02:29,466 --> 00:02:31,066
that it is being masked by something
79
00:02:31,066 --> 00:02:32,766
so let's plug that in
80
00:02:33,700 --> 00:02:34,533
and I'll show you why
81
00:02:34,533 --> 00:02:35,700
it doesn't matter if I plug it in or not
82
00:02:35,700 --> 00:02:37,766
because when you do the Image analysis tab
83
00:02:37,766 --> 00:02:39,300
you have the choice between
84
00:02:39,600 --> 00:02:40,166
the mask channel
85
00:02:40,166 --> 00:02:41,300
it could be the source alpha
86
00:02:41,300 --> 00:02:42,966
so it doesn't have to have a mask
87
00:02:43,100 --> 00:02:43,666
in this case
88
00:02:43,666 --> 00:02:45,766
I'm just gonna call the mask Alpha since
89
00:02:45,766 --> 00:02:47,333
it's the same thing um
90
00:02:47,333 --> 00:02:49,366
it just makes things a little bit clearer
91
00:02:49,366 --> 00:02:50,600
for anyone who might pick up the script
92
00:02:50,600 --> 00:02:51,933
or maybe for myself if
93
00:02:51,933 --> 00:02:55,000
I'm still working on the shot weeks later um
94
00:02:55,366 --> 00:02:57,066
okay so going back to actually
95
00:02:57,066 --> 00:02:58,733
let's go ahead and start with the Image analysis
96
00:02:58,733 --> 00:03:00,566
tab before we go back to the 1 of the distortion tab
97
00:03:00,566 --> 00:03:03,466
so what you have three options when you do
98
00:03:03,466 --> 00:03:04,600
one of the distortion in nuke
99
00:03:05,800 --> 00:03:08,000
image analysis is nuke literally
100
00:03:08,333 --> 00:03:09,300
place to the sequence
101
00:03:09,300 --> 00:03:11,600
whatever clip range you tell it to and
102
00:03:11,666 --> 00:03:13,866
guesses what the one distortion would be based on how
103
00:03:13,866 --> 00:03:15,366
the pixels move in the image
104
00:03:15,533 --> 00:03:16,800
um this is gonna be the way
105
00:03:16,800 --> 00:03:18,100
you're gonna wanna use it most of the time
106
00:03:18,100 --> 00:03:20,966
mainly because you're not gonna have options for this
107
00:03:20,966 --> 00:03:22,100
the absolute best way to do it
108
00:03:22,100 --> 00:03:23,166
which is the grid analysis
109
00:03:23,166 --> 00:03:24,333
and we'll talk about that in a little while
110
00:03:24,333 --> 00:03:26,766
but that means that you were actually onset that day
111
00:03:26,866 --> 00:03:28,266
or your supervisor was
112
00:03:28,266 --> 00:03:30,300
and they shot a grid
113
00:03:30,533 --> 00:03:32,200
of every lens that was shot
114
00:03:32,200 --> 00:03:33,566
and then you can take that
115
00:03:34,366 --> 00:03:36,133
scanned or it was red camera
116
00:03:36,133 --> 00:03:36,566
for example
117
00:03:36,566 --> 00:03:37,966
a digital file of that grid that
118
00:03:37,966 --> 00:03:38,900
was shot with that camera
119
00:03:38,900 --> 00:03:39,600
with that lens
120
00:03:39,600 --> 00:03:40,966
and plug it right in here
121
00:03:41,000 --> 00:03:42,766
and it'll analyze the grid um
122
00:03:42,766 --> 00:03:43,566
this is much faster
123
00:03:43,566 --> 00:03:45,866
cause it helped me has to analyze one frame and
124
00:03:45,900 --> 00:03:48,100
it also is extremely accurate um
125
00:03:48,100 --> 00:03:50,166
if you're doing a blue screen shoot and the
126
00:03:50,166 --> 00:03:51,333
there's two people in the foreground
127
00:03:51,333 --> 00:03:52,866
and nothing blue screen behind them
128
00:03:52,900 --> 00:03:55,300
the image analysis is not gonna do anything for you
129
00:03:55,466 --> 00:03:57,266
you you'd have to mask up the people cause they move
130
00:03:57,266 --> 00:03:59,633
and then the blue screen has nothing to
131
00:03:59,966 --> 00:04:01,066
analyze with
132
00:04:01,066 --> 00:04:02,266
if you had some tracking markers
133
00:04:02,266 --> 00:04:02,966
if you had enough of them
134
00:04:02,966 --> 00:04:05,266
then image analysis could probably still work
135
00:04:05,466 --> 00:04:06,400
but I wouldn't bet on it
136
00:04:06,400 --> 00:04:08,033
you're gonna have to have a great analysis
137
00:04:08,733 --> 00:04:11,466
or draw very clear tracking markers that
138
00:04:11,466 --> 00:04:12,400
you know that they're straight
139
00:04:12,400 --> 00:04:13,733
because then you can use line analysis
140
00:04:13,733 --> 00:04:15,966
which we'll talk about in a second as well alright
141
00:04:15,966 --> 00:04:17,066
so going back to image analysis
142
00:04:17,066 --> 00:04:18,266
because in this case
143
00:04:18,266 --> 00:04:20,000
this is gonna work like a champ for us
144
00:04:20,166 --> 00:04:21,000
um so I mean
145
00:04:21,000 --> 00:04:23,033
I'm gonna mask out those things that are moving
146
00:04:23,200 --> 00:04:24,733
and then the only thing left to do
147
00:04:24,733 --> 00:04:27,200
is change my camera motion from free camera
148
00:04:27,733 --> 00:04:29,000
to rotation only
149
00:04:29,000 --> 00:04:30,466
because by looking at the shot
150
00:04:30,466 --> 00:04:31,366
I've determined that
151
00:04:31,366 --> 00:04:33,266
is just a camera on a tripod
152
00:04:33,266 --> 00:04:35,466
or someone's holding it um
153
00:04:35,566 --> 00:04:36,600
and just moving it around
154
00:04:36,600 --> 00:04:38,333
so it's not something that's dollying
155
00:04:38,333 --> 00:04:39,166
it's not being
156
00:04:39,166 --> 00:04:40,100
pulled or pushed
157
00:04:40,100 --> 00:04:42,600
and because it's just rotation um
158
00:04:42,866 --> 00:04:43,666
Nuke's gonna be
159
00:04:43,666 --> 00:04:45,733
a little faster at picking up lens distortion
160
00:04:45,733 --> 00:04:47,366
and a little bit more accurate as well
161
00:04:47,366 --> 00:04:49,066
so I'm gonna go ahead and
162
00:04:49,066 --> 00:04:49,800
click rotation only
163
00:04:49,800 --> 00:04:51,566
I'm gonna say analyze image
164
00:04:53,966 --> 00:04:56,066
and the nuke will plug away and should a progress bar
165
00:04:56,066 --> 00:04:57,866
should pop up force and links how long it will take
166
00:05:03,133 --> 00:05:04,700
it does take a little while it'll
167
00:05:04,700 --> 00:05:07,000
it'll obviously be a lot faster depending on your CPU
168
00:05:07,500 --> 00:05:08,900
and your Io to an extent
169
00:05:08,900 --> 00:05:09,700
if this was
170
00:05:09,866 --> 00:05:11,200
these are small JPEG frames
171
00:05:11,200 --> 00:05:12,766
but if these were 2K
172
00:05:13,100 --> 00:05:16,300
or 4 k Vista frames or red frames or 30
173
00:05:16,300 --> 00:05:17,300
500 or scan film
174
00:05:17,300 --> 00:05:18,933
it could certainly take quite a while
175
00:05:18,933 --> 00:05:21,166
and your Io would certainly impact the speed of it
176
00:05:21,200 --> 00:05:22,100
but most of the point
177
00:05:22,100 --> 00:05:23,500
most of the time it's gonna be your CPU
178
00:05:23,500 --> 00:05:25,266
so as you using a laptop for this
179
00:05:25,266 --> 00:05:26,333
prepared to be patient
180
00:05:26,333 --> 00:05:28,300
cause it is a very intensive process
181
00:05:28,300 --> 00:05:30,300
that nuke does take a little while to accomplish
182
00:05:31,933 --> 00:05:33,466
but there's no way around it unfortunately
183
00:05:33,466 --> 00:05:34,500
this point so
184
00:05:38,066 --> 00:05:38,766
and that being said
185
00:05:38,766 --> 00:05:39,900
actually there will be
186
00:05:40,166 --> 00:05:41,766
gives you more reason to use the grid analysis
187
00:05:41,766 --> 00:05:43,400
if you happen to be able to get a grid
188
00:05:43,700 --> 00:05:45,533
or the line analysis if you have like a shot
189
00:05:45,533 --> 00:05:47,200
like this is a perfect one for line analysis
190
00:05:47,200 --> 00:05:49,400
because we have lots of lines in the buildings
191
00:05:49,400 --> 00:05:51,166
with which to analyze in it
192
00:05:51,166 --> 00:05:52,566
although both of those are much quicker
193
00:05:52,566 --> 00:05:54,100
process than image analysis
194
00:05:58,400 --> 00:06:02,566
and actually will go forward and back um
195
00:06:02,566 --> 00:06:05,500
before it's done to make sure it's got everything right
196
00:06:09,200 --> 00:06:10,900
and I will since it's got another four minutes
197
00:06:10,900 --> 00:06:12,366
I'm gonna go ahead and pause
198
00:06:13,600 --> 00:06:14,333
my recording
199
00:06:14,333 --> 00:06:16,300
and I'll be right back as soon as it's done
200
00:06:17,200 --> 00:06:18,633
so distortion is done
201
00:06:19,966 --> 00:06:22,300
and you'll see the features actually will still be on
202
00:06:22,333 --> 00:06:23,533
the footage because I'm showing it
203
00:06:23,533 --> 00:06:26,133
it's having it to show me the features if I can
204
00:06:26,133 --> 00:06:26,800
features and tracks
205
00:06:26,800 --> 00:06:28,433
I can see every track it drew
206
00:06:28,700 --> 00:06:30,366
to determine what the distortion is
207
00:06:31,266 --> 00:06:32,333
um so let's just go to none
208
00:06:32,333 --> 00:06:33,800
so we don't need to see those anymore
209
00:06:33,933 --> 00:06:35,200
and take a look at our image
210
00:06:35,200 --> 00:06:36,433
see what we've come up with here
211
00:06:36,933 --> 00:06:40,100
so here's a good frame because lots of lines to see so
212
00:06:40,800 --> 00:06:42,033
make this a little bigger
213
00:06:43,066 --> 00:06:44,366
so if I look at
214
00:06:45,066 --> 00:06:47,000
those distortion of an actual distortion
215
00:06:47,700 --> 00:06:50,233
you'll see that we are losing our lens distortion
216
00:06:54,333 --> 00:06:56,000
and we'll look at the registerion tab
217
00:06:56,000 --> 00:06:58,266
we'll see that these values are now filled out
218
00:06:58,266 --> 00:06:59,933
including the card parameters down below
219
00:06:59,933 --> 00:07:02,166
so we could just plug these parameters into a card
220
00:07:02,466 --> 00:07:05,066
and instead of having to do any other math in here
221
00:07:05,066 --> 00:07:06,566
the card itself would then
222
00:07:06,933 --> 00:07:08,266
undistort the image for us
223
00:07:08,533 --> 00:07:10,200
and now this undistort button is is key
224
00:07:10,200 --> 00:07:11,133
because by default
225
00:07:11,133 --> 00:07:11,933
it's gonna undistort
226
00:07:11,933 --> 00:07:13,333
but if I wanted it to distort
227
00:07:13,333 --> 00:07:13,866
for example
228
00:07:13,866 --> 00:07:15,833
if we had rendered CG
229
00:07:16,000 --> 00:07:17,866
that doesn't have distortion in it
230
00:07:17,866 --> 00:07:18,900
I would wanna unclick this
231
00:07:18,900 --> 00:07:21,866
so that I'd be distorting the image um
232
00:07:22,000 --> 00:07:22,800
and then plug that through
233
00:07:22,800 --> 00:07:24,366
now you can use that simply by
234
00:07:24,366 --> 00:07:27,366
having multiple under distortion to tools in
235
00:07:27,400 --> 00:07:30,133
but what I prefer to do is output the displacement map
236
00:07:30,133 --> 00:07:32,133
which can be done up here under output type
237
00:07:32,133 --> 00:07:34,700
you can change to displacement instead of image
238
00:07:34,933 --> 00:07:36,100
and when you do that
239
00:07:36,466 --> 00:07:36,933
first of all
240
00:07:36,933 --> 00:07:37,733
the distortion
241
00:07:37,800 --> 00:07:38,900
nothing happens to the image
242
00:07:38,900 --> 00:07:41,866
but what happens is new channels are created
243
00:07:42,000 --> 00:07:44,466
it creates the Motion dot U and Motion dot V pass
244
00:07:44,466 --> 00:07:46,200
so if we channel shuffle that
245
00:07:51,066 --> 00:07:52,466
you'll see this is what we get
246
00:07:52,900 --> 00:07:54,000
and this is our lens map
247
00:07:54,000 --> 00:07:56,366
this is now gonna be the actual mapping of the lens
248
00:07:56,366 --> 00:07:59,533
as far as nukes as tools have calculated
249
00:07:59,533 --> 00:08:01,833
so now we can render this out as a single frame
250
00:08:02,600 --> 00:08:04,300
and I recommend doing that most of the time
251
00:08:04,300 --> 00:08:06,800
so I'm gonna go ahead and render that into our file
252
00:08:06,800 --> 00:08:07,533
so you guys have that
253
00:08:07,533 --> 00:08:07,900
so for these
254
00:08:07,900 --> 00:08:08,533
you on laptops
255
00:08:08,533 --> 00:08:11,433
you don't need necessarily render everything
256
00:08:20,733 --> 00:08:23,566
I actually make sure I call it analysis too
257
00:08:28,466 --> 00:08:29,633
since we know
258
00:08:32,066 --> 00:08:33,166
that this was done with an image
259
00:08:33,166 --> 00:08:34,200
analysis will be doing it with a
260
00:08:34,200 --> 00:08:35,233
couple different things
261
00:08:35,466 --> 00:08:37,133
and you have to make sure you write it as an EXR
262
00:08:37,133 --> 00:08:37,700
or any high
263
00:08:37,700 --> 00:08:38,666
dynamic range image
264
00:08:38,666 --> 00:08:39,366
but EXR is really
265
00:08:39,366 --> 00:08:40,900
the only one that's gonna be high enough for you
266
00:08:40,900 --> 00:08:42,566
I think DPX would probably work too
267
00:08:42,566 --> 00:08:44,566
but um EXR is a safe bet
268
00:08:44,566 --> 00:08:47,100
cause there's data in here that is well above 1
269
00:08:47,100 --> 00:08:48,866
and well below 0 in some cases
270
00:08:48,966 --> 00:08:50,666
and whether there is not
271
00:08:50,666 --> 00:08:51,966
you have to make sure that you're covering for that
272
00:08:51,966 --> 00:08:53,700
so I'll go ahead and render that
273
00:08:53,700 --> 00:08:54,933
and it doesn't matter what frame
274
00:08:54,933 --> 00:08:56,500
I'll just pick my current frame
275
00:08:57,533 --> 00:08:59,066
so now I have that written out
276
00:08:59,066 --> 00:09:01,200
and I'm gonna go ahead and bring it back into my scene
277
00:09:04,266 --> 00:09:05,066
as expected
278
00:09:05,066 --> 00:09:06,966
it looks exactly like the frame we were creating
279
00:09:06,966 --> 00:09:09,633
so let's bring it back at our node graph here
280
00:09:10,533 --> 00:09:12,766
and I'm gonna bring in an s t map tool
281
00:09:12,766 --> 00:09:16,300
which is under transform s t map
282
00:09:18,533 --> 00:09:20,766
and this tool is extraordinarily easy to use
283
00:09:20,766 --> 00:09:22,733
so it's gonna move my Rotel pane
284
00:09:22,733 --> 00:09:23,866
I was a source to the side
285
00:09:23,866 --> 00:09:25,300
and go ahead and look at my plate again
286
00:09:25,400 --> 00:09:28,400
so I'm gonna plug in the SRC
287
00:09:28,600 --> 00:09:30,833
the source arrow into my plate
288
00:09:31,166 --> 00:09:34,633
and the ST map into the UVs we just created
289
00:09:35,300 --> 00:09:37,366
and with the change a couple settings on the s t map
290
00:09:37,366 --> 00:09:40,900
so mainly channels are all which is fine UV channels
291
00:09:40,900 --> 00:09:44,133
this is gonna be the data that we're plugging into it
292
00:09:44,133 --> 00:09:44,766
so in this case
293
00:09:44,766 --> 00:09:47,166
since we shuffled our UV channels into our RGB
294
00:09:47,166 --> 00:09:49,666
we need to change the UV channels to RGB
295
00:09:50,566 --> 00:09:53,366
and if we look at our image before and after
296
00:09:56,466 --> 00:09:58,433
you'll see we're getting the proper distortion
297
00:10:00,533 --> 00:10:02,200
and this shot does have tons of distortion
298
00:10:02,200 --> 00:10:04,500
but you can see that we are in fact changing the image
299
00:10:05,333 --> 00:10:07,500
now for those of you that want to do an experiment
300
00:10:07,500 --> 00:10:09,066
if you run the camera tracker
301
00:10:09,066 --> 00:10:11,333
which will go over of course in a few minutes here
302
00:10:11,333 --> 00:10:12,566
if you run the camera tracker on
303
00:10:12,566 --> 00:10:14,433
the undistorted image versus the
304
00:10:14,966 --> 00:10:16,333
raw plate you will see that
305
00:10:16,333 --> 00:10:18,700
the result is much better with the undistorted camera
306
00:10:18,900 --> 00:10:21,866
so it's um good experiment
307
00:10:21,866 --> 00:10:23,033
for those you that don't
308
00:10:23,100 --> 00:10:24,800
understand why I do lungs distortion
309
00:10:25,133 --> 00:10:26,000
but this is also the case
310
00:10:26,000 --> 00:10:28,100
and especially when you're putting anything CG into a
311
00:10:28,100 --> 00:10:29,566
in a 2 shot like this
312
00:10:30,800 --> 00:10:33,966
taking out the lens distortion is vital because
313
00:10:34,300 --> 00:10:35,966
the CG itself doesn't have one distortion
314
00:10:35,966 --> 00:10:37,400
and there's no way to render it in
315
00:10:37,400 --> 00:10:40,566
so you need to be able to do that later by
316
00:10:41,200 --> 00:10:42,466
running it through something like this
317
00:10:42,466 --> 00:10:43,366
with the G with an ST map
318
00:10:43,366 --> 00:10:44,800
we want to redistor it which was mean
319
00:10:44,800 --> 00:10:46,866
we'd go through the lens distortion tool again
320
00:10:47,700 --> 00:10:49,866
and instead of clicking understort we click distort
321
00:10:49,866 --> 00:10:51,600
which would give us another map
322
00:10:53,733 --> 00:10:55,200
so although this is already distorts
323
00:10:55,200 --> 00:10:55,866
let's go ahead and look at
324
00:10:55,866 --> 00:10:57,400
plug this through to our steam map
325
00:10:57,400 --> 00:10:58,800
which is un distorting the image
326
00:11:01,533 --> 00:11:02,666
and look at our original
327
00:11:02,800 --> 00:11:04,600
and we should see no change so there we have it
328
00:11:04,600 --> 00:11:06,900
so this is distorting the image
329
00:11:08,366 --> 00:11:09,366
this is the original
330
00:11:09,366 --> 00:11:10,766
and then if we look at our lens distortion
331
00:11:10,766 --> 00:11:11,766
which is now
332
00:11:12,500 --> 00:11:14,633
distorting instead of un distorting the image
333
00:11:15,200 --> 00:11:16,566
and look at our original plate
334
00:11:16,700 --> 00:11:18,200
you'll see that everything is exactly the same
335
00:11:18,200 --> 00:11:20,266
except that we've softened the image slightly
336
00:11:20,500 --> 00:11:21,766
just very slightly
337
00:11:22,066 --> 00:11:24,100
due to the fact that it belongs distortion
338
00:11:24,333 --> 00:11:25,733
which is another tip actually
339
00:11:25,733 --> 00:11:26,300
while thinking of it
340
00:11:26,300 --> 00:11:27,800
you never wanted to distort the plate
341
00:11:27,800 --> 00:11:30,300
so I'm undistorting the plate here
342
00:11:30,933 --> 00:11:32,300
for tracking purposes
343
00:11:32,766 --> 00:11:35,566
but unless there's a special case where I wanna
344
00:11:35,566 --> 00:11:37,000
make sure the plate never
345
00:11:37,100 --> 00:11:38,566
gets distorted and it's just seen that way
346
00:11:38,566 --> 00:11:40,266
then I can just add a sharpening filter to
347
00:11:40,266 --> 00:11:41,900
to bring back some of that detail
348
00:11:42,100 --> 00:11:43,066
but ideally
349
00:11:43,566 --> 00:11:45,800
I only use this for tracking purposes
350
00:11:45,800 --> 00:11:47,033
and I never
351
00:11:47,166 --> 00:11:49,900
never distort the plate for an actual final comp
352
00:11:49,900 --> 00:11:51,366
so now that I'm gonna sort this
353
00:11:51,366 --> 00:11:51,966
I would track it
354
00:11:51,966 --> 00:11:53,000
put in my CG
355
00:11:53,100 --> 00:11:54,200
and instead of
356
00:11:54,400 --> 00:11:55,333
und distorting the plate
357
00:11:55,333 --> 00:11:57,266
I would redestort the the CG
358
00:11:57,266 --> 00:11:58,700
so then the CG would be distorted
359
00:11:58,700 --> 00:12:00,833
and then put on top of the distorted plate
360
00:12:00,933 --> 00:12:03,800
that is always the best way to work
361
00:12:03,800 --> 00:12:04,600
because then you're using
362
00:12:04,600 --> 00:12:06,700
the original plate with no softening going on
363
00:12:06,700 --> 00:12:08,066
it's exactly what were shot
364
00:12:08,066 --> 00:12:09,566
you're not degrading the image at all
365
00:12:09,966 --> 00:12:11,766
of course there are occasions where you need to
366
00:12:11,766 --> 00:12:13,366
but rare um
367
00:12:13,366 --> 00:12:14,566
so just on a quick work flow note
368
00:12:14,566 --> 00:12:16,466
that's what you wanna try to maintain
369
00:12:16,500 --> 00:12:17,900
throughout your entire pipeline
370
00:12:18,400 --> 00:12:19,000
okay so now
371
00:12:19,000 --> 00:12:21,400
let's look at another way to use the
372
00:12:21,533 --> 00:12:22,633
lens distortion tool
373
00:12:22,800 --> 00:12:24,133
so there's two more tabs here
374
00:12:24,133 --> 00:12:25,666
grid analysis and line analysis
375
00:12:25,666 --> 00:12:27,666
let's go ahead and talk about line analysis first
376
00:12:28,000 --> 00:12:29,066
so in a shot like this
377
00:12:29,066 --> 00:12:31,166
line analysis will actually work really well
378
00:12:31,200 --> 00:12:32,066
and will be much
379
00:12:32,066 --> 00:12:34,033
much faster than image analysis
380
00:12:34,100 --> 00:12:35,700
so let's go ahead and just see
381
00:12:35,700 --> 00:12:36,533
you pick a good frame with it
382
00:12:36,533 --> 00:12:38,000
lots of lines on each
383
00:12:38,000 --> 00:12:39,000
edge of frame
384
00:12:39,166 --> 00:12:40,466
let's think at the beginning actually
385
00:12:40,466 --> 00:12:40,966
better place
386
00:12:40,966 --> 00:12:42,466
so this is gonna be a good place to
387
00:12:42,766 --> 00:12:43,333
to judge it
388
00:12:43,333 --> 00:12:45,700
so let's try to make sure we have lines off frame okay
389
00:12:45,700 --> 00:12:46,700
so line analysis
390
00:12:46,700 --> 00:12:48,766
what it does is you draw lines on the picture
391
00:12:48,766 --> 00:12:49,900
and then nuke will
392
00:12:49,900 --> 00:12:51,733
figure out based on just a few lines
393
00:12:51,733 --> 00:12:52,833
exactly what
394
00:12:52,933 --> 00:12:54,266
the lens distortion is
395
00:12:54,366 --> 00:12:56,533
uh it's not quite as accurate as image analysis
396
00:12:56,533 --> 00:12:57,000
and it does
397
00:12:57,000 --> 00:13:00,900
depend heavily on straight lines to be in your shot
398
00:13:00,900 --> 00:13:02,066
so it's not always gonna work
399
00:13:02,066 --> 00:13:03,633
but when you have lines like this
400
00:13:03,900 --> 00:13:04,933
it works really
401
00:13:04,933 --> 00:13:05,533
really close
402
00:13:05,533 --> 00:13:07,000
and it's much faster image analysis
403
00:13:07,000 --> 00:13:08,233
so let's take a quick look
404
00:13:08,300 --> 00:13:09,333
and just start drawing our lines
405
00:13:09,333 --> 00:13:10,866
so the first thing you need to do
406
00:13:11,000 --> 00:13:12,700
is turn drawing mode on
407
00:13:13,266 --> 00:13:14,966
everything else is pretty much gonna be the same
408
00:13:14,966 --> 00:13:16,066
so you just click
409
00:13:16,066 --> 00:13:17,833
one point for the start of a line
410
00:13:18,500 --> 00:13:20,233
and you do wanna be pretty accurate here
411
00:13:20,333 --> 00:13:21,900
go to the middle point of the line
412
00:13:21,900 --> 00:13:23,400
and we'll go to the end of the line
413
00:13:24,300 --> 00:13:25,800
and once you have all three points
414
00:13:25,800 --> 00:13:26,600
you right click
415
00:13:26,600 --> 00:13:27,866
and a line will be drawn
416
00:13:28,366 --> 00:13:29,900
and let's go pick another line down here
417
00:13:29,900 --> 00:13:31,833
I'm gonna game up a little bit so we can see better
418
00:13:32,966 --> 00:13:34,866
and getting one be pretty accurate so
419
00:13:35,100 --> 00:13:36,166
right there
420
00:13:37,466 --> 00:13:40,033
right there and then all the way over here
421
00:13:40,300 --> 00:13:41,500
and I'm gonna right click
422
00:13:42,500 --> 00:13:45,033
and let's try to pick a line over here
423
00:13:45,300 --> 00:13:46,066
see if we can see one
424
00:13:46,066 --> 00:13:47,733
it's kind of hard to see the ones on the side
425
00:13:47,733 --> 00:13:49,166
and see if we game it down again
426
00:13:49,866 --> 00:13:52,866
alright so this one actually might work so 1
427
00:13:54,200 --> 00:13:57,700
2 3 and then again you right click
428
00:13:59,933 --> 00:14:01,333
and let's see if we can pick another one
429
00:14:01,333 --> 00:14:02,400
these are a little harder to see
430
00:14:02,400 --> 00:14:03,700
but I think it's it's worth it
431
00:14:03,700 --> 00:14:05,800
because it's gonna be a lot of good distortion
432
00:14:06,200 --> 00:14:07,666
you can really be able to understand the distortion
433
00:14:07,666 --> 00:14:09,566
when there's things closer to the edge of frame
434
00:14:09,966 --> 00:14:12,666
so you wanna give it as much detail as you can so one
435
00:14:14,900 --> 00:14:18,166
2 3 right click
436
00:14:19,666 --> 00:14:21,366
and let's see this line
437
00:14:21,366 --> 00:14:22,700
that might be a good one actually
438
00:14:24,366 --> 00:14:26,866
1 2
439
00:14:28,200 --> 00:14:29,500
3 right click
440
00:14:30,400 --> 00:14:32,100
and let's see
441
00:14:32,300 --> 00:14:33,733
see if we can pick any more horizontal ones
442
00:14:33,733 --> 00:14:36,033
let's go ahead and pick one of these right here
443
00:14:37,766 --> 00:14:39,200
that one actually is a little too high
444
00:14:39,200 --> 00:14:40,800
there seems to be a bump in the line
445
00:14:40,800 --> 00:14:42,833
so we'll have to go below that a little bit
446
00:14:42,933 --> 00:14:44,633
so I'm just gonna delete last point
447
00:14:50,566 --> 00:14:52,300
and you may need to look through your sequence quite a
448
00:14:52,300 --> 00:14:54,300
bit to find a line that's gonna work best for you
449
00:14:54,666 --> 00:14:57,200
and you do there's really no
450
00:14:57,200 --> 00:14:58,933
perfect number of lines um
451
00:14:58,933 --> 00:15:00,600
you definitely need at least three
452
00:15:00,600 --> 00:15:01,966
to to really make it a good line
453
00:15:01,966 --> 00:15:03,133
and it really depends how long they are
454
00:15:03,133 --> 00:15:04,400
the ones that go all the way across frame
455
00:15:04,400 --> 00:15:05,900
and are nice and clear
456
00:15:05,900 --> 00:15:07,566
you might need less but um
457
00:15:07,566 --> 00:15:09,500
I mean you're not gonna get a result without lip
458
00:15:09,666 --> 00:15:10,866
without at least three
459
00:15:10,866 --> 00:15:12,666
and nuke will yell at you if you try to
460
00:15:12,666 --> 00:15:13,900
pick two few lines
461
00:15:18,266 --> 00:15:21,300
and right click and
462
00:15:22,566 --> 00:15:23,700
let's try to get one more
463
00:15:23,700 --> 00:15:25,600
we'll go ahead and just pick this one right here
464
00:15:30,566 --> 00:15:32,400
so they don't have to be on the straight plane
465
00:15:32,400 --> 00:15:33,366
they can be
466
00:15:33,766 --> 00:15:35,166
in other places so
467
00:15:35,900 --> 00:15:36,866
go ahead and pick this one too
468
00:15:36,866 --> 00:15:38,466
I guess actually a sidewalks of bound to pick
469
00:15:38,466 --> 00:15:39,666
because those are rarely perfect
470
00:15:39,666 --> 00:15:41,100
so I'll delete last point
471
00:15:43,666 --> 00:15:46,133
buildings of course are great ones to pick
472
00:15:46,133 --> 00:15:49,033
and let's go ahead and just pick this line right here
473
00:15:54,366 --> 00:15:56,300
alright so let's see what nuke does with this value
474
00:15:57,166 --> 00:15:58,466
so once you have all your lines drawn
475
00:15:58,466 --> 00:15:59,833
you just say analyze lines
476
00:16:00,533 --> 00:16:02,733
I had this plugged into our undistorted plate
477
00:16:02,733 --> 00:16:03,900
so it's actually not gonna have any
478
00:16:03,900 --> 00:16:06,666
distortion to remove so
479
00:16:07,733 --> 00:16:09,600
let's go ahead and kill this one's distortion tool
480
00:16:09,600 --> 00:16:10,866
and I'm gonna bring in a new one
481
00:16:12,333 --> 00:16:13,766
and do the same thing again
482
00:16:18,200 --> 00:16:19,566
and I'll do it very quick this time
483
00:16:19,566 --> 00:16:20,466
I shall just
484
00:16:20,500 --> 00:16:22,300
I'll pause and draw them and come right back
485
00:16:23,200 --> 00:16:24,466
alright so I do the some lines again
486
00:16:24,466 --> 00:16:27,466
I didn't draw them in the exact same place but you know
487
00:16:27,666 --> 00:16:28,500
doesn't really matter
488
00:16:28,500 --> 00:16:29,733
as long as you find some straight lines to draw
489
00:16:29,733 --> 00:16:30,300
so you can see that
490
00:16:30,300 --> 00:16:33,466
the distortion is now happening again very accurately
491
00:16:34,400 --> 00:16:35,500
or the undistortion rather
492
00:16:35,500 --> 00:16:36,533
let me make sure I'm on undistort
493
00:16:36,533 --> 00:16:37,800
I am on undistort
494
00:16:38,000 --> 00:16:39,233
so there's our
495
00:16:40,800 --> 00:16:41,733
one analysis
496
00:16:41,733 --> 00:16:42,700
oh and I kept
497
00:16:44,666 --> 00:16:45,133
adding more lines
498
00:16:45,133 --> 00:16:47,033
so I'll turn off drawing mode now that I'm done
499
00:16:47,466 --> 00:16:48,566
so there are my lines
500
00:16:48,600 --> 00:16:50,166
and now something good to know
501
00:16:50,166 --> 00:16:51,266
that I think I mentioned before
502
00:16:51,266 --> 00:16:53,966
so the lens distortion tool can only be
503
00:16:54,266 --> 00:16:55,966
analyzed in Nuke X
504
00:16:55,966 --> 00:16:57,600
it's a Nuke X only tool
505
00:16:57,666 --> 00:16:59,766
but once that analyzation is done
506
00:16:59,766 --> 00:17:03,300
I can take this node and bring it into nuke 6
507
00:17:03,400 --> 00:17:04,700
and or even
508
00:17:04,866 --> 00:17:06,300
no I guess it nuke 6 only
509
00:17:06,300 --> 00:17:08,166
I could bring this into older versions of Nuke
510
00:17:08,166 --> 00:17:09,366
Nuke 5 and such
511
00:17:09,366 --> 00:17:10,900
but this node will work
512
00:17:10,900 --> 00:17:14,200
and distort and understort images fine in nuke 6
513
00:17:14,900 --> 00:17:17,600
you only need to Nuke 6 x
514
00:17:17,900 --> 00:17:19,533
when you're actually doing the analyzation
515
00:17:19,533 --> 00:17:20,200
whether it's image
516
00:17:20,200 --> 00:17:21,133
greater line
517
00:17:21,133 --> 00:17:22,300
so that's just good to know
518
00:17:22,300 --> 00:17:24,666
so that for those facilities or people that are
519
00:17:24,766 --> 00:17:26,500
hoping to own multiple copies of nuke
520
00:17:26,533 --> 00:17:29,433
you don't have to buy Nuke X just to do one distortion
521
00:17:30,566 --> 00:17:32,000
I only one copy of it rather
522
00:17:32,000 --> 00:17:34,166
and then the rest of your nuke stations will be able to
523
00:17:34,166 --> 00:17:35,133
distort the image fine
524
00:17:35,133 --> 00:17:37,300
I just can't analyze the lens distortion alright
525
00:17:37,300 --> 00:17:38,900
so that's how one distortion works
526
00:17:38,900 --> 00:17:41,900
if still populates this the values in the
527
00:17:41,900 --> 00:17:44,033
in the settings in the car parameters as well
528
00:17:44,400 --> 00:17:46,166
just like the image analysis does
529
00:17:46,466 --> 00:17:47,533
now let's take a look at grid analysis
530
00:17:47,533 --> 00:17:49,700
so for grid analysis because I don't have grid
531
00:17:49,700 --> 00:17:51,466
lens grids that were shot for this
532
00:17:51,666 --> 00:17:54,366
um we have to go and generate our own as an experiment
533
00:17:54,900 --> 00:17:56,666
so I'm gonna bring in a checkerboard
534
00:17:57,266 --> 00:17:58,666
and I'm gonna go ahead and make it
535
00:17:59,166 --> 00:18:01,566
the HD size
536
00:18:01,733 --> 00:18:03,000
and now what a lens grid is
537
00:18:03,000 --> 00:18:05,900
is you literally shoot a grid so you'll print
538
00:18:05,900 --> 00:18:08,566
you would make a large grid to bring to set
539
00:18:08,900 --> 00:18:10,366
and usually it's black and white
540
00:18:12,866 --> 00:18:14,233
let's make it black and white
541
00:18:17,133 --> 00:18:18,466
with no centerline
542
00:18:20,700 --> 00:18:22,000
and we'll make these a little bit bigger
543
00:18:22,000 --> 00:18:23,433
not that it matters too much
544
00:18:23,500 --> 00:18:25,900
you can make this squares any size you want really
545
00:18:26,866 --> 00:18:28,200
just make enough to fill frame
546
00:18:28,200 --> 00:18:30,066
so you would literally just shoot
547
00:18:30,333 --> 00:18:31,666
the lens good with the camera
548
00:18:31,900 --> 00:18:33,766
and the important thing is you want to have to
549
00:18:33,766 --> 00:18:35,133
you want to fill frame so you get
550
00:18:35,133 --> 00:18:37,500
the curvature of the lens all the way on all sides
551
00:18:37,666 --> 00:18:40,500
and um that will distort the grid obviously
552
00:18:40,500 --> 00:18:42,100
because you're shooting through the lens so
553
00:18:42,166 --> 00:18:43,966
we'll have to create one since we don't have one
554
00:18:43,966 --> 00:18:44,400
in this case
555
00:18:44,400 --> 00:18:45,466
so I'm just gonna
556
00:18:45,533 --> 00:18:47,966
use the lens distortion tool we have
557
00:18:48,333 --> 00:18:50,533
and I'm gonna look at the grid checker word through it
558
00:18:50,533 --> 00:18:52,666
and I'm gonna change undistort to distort
559
00:18:53,133 --> 00:18:54,300
so now our grid
560
00:18:54,300 --> 00:18:55,466
is distorted
561
00:18:56,400 --> 00:18:57,366
to match the plates enough
562
00:18:57,366 --> 00:18:58,700
we take this
563
00:18:59,000 --> 00:19:00,866
and add another lens distortion tool
564
00:19:01,733 --> 00:19:02,866
and again all we're doing with this
565
00:19:02,866 --> 00:19:04,800
lens distortion is just storing the checkerboard
566
00:19:04,966 --> 00:19:07,200
as if as if we had shot it with the camera
567
00:19:07,200 --> 00:19:08,333
and the lens that shot this
568
00:19:08,333 --> 00:19:10,300
and that's ideally how this would work
569
00:19:10,333 --> 00:19:11,466
you'd never actually
570
00:19:11,466 --> 00:19:13,600
take a checkerboard in distort it
571
00:19:13,666 --> 00:19:14,900
and then run through grid analysis
572
00:19:14,900 --> 00:19:15,900
but for our experiments
573
00:19:15,900 --> 00:19:16,733
I don't have a lens grid
574
00:19:16,733 --> 00:19:17,766
this is how we do it
575
00:19:17,766 --> 00:19:20,366
so lens distortion gonna go to grid analysis
576
00:19:20,500 --> 00:19:23,400
and um all you really have to do is click analyze grid
577
00:19:24,533 --> 00:19:25,700
you know take a look at that
578
00:19:25,700 --> 00:19:27,400
and anything you can do is a grid overlay
579
00:19:27,400 --> 00:19:28,966
so you can actually look at the original grid
580
00:19:28,966 --> 00:19:30,900
so it'll show you an overlay of the grid
581
00:19:30,900 --> 00:19:32,633
or you can look at the understorted grid
582
00:19:32,733 --> 00:19:34,533
so you can see what it's actually doing
583
00:19:34,533 --> 00:19:36,500
you can see this is just map the grid perfectly and
584
00:19:36,500 --> 00:19:37,100
original grid
585
00:19:37,100 --> 00:19:39,100
you can see that it is curved that way
586
00:19:39,166 --> 00:19:40,966
so now that we've done this
587
00:19:41,666 --> 00:19:43,200
I wanna take a look at
588
00:19:43,300 --> 00:19:45,400
the result of this to see how accurate we were
589
00:19:45,400 --> 00:19:49,066
so if we look at our source
590
00:19:49,700 --> 00:19:51,166
which let me make sure this is on undistored
591
00:19:51,166 --> 00:19:53,166
so we're gonna undistort our image based on this
592
00:19:53,700 --> 00:19:55,400
and now we're going to take
593
00:19:56,400 --> 00:19:57,700
our ones distortion
594
00:19:57,800 --> 00:19:59,200
that we did with our grid
595
00:19:59,700 --> 00:20:01,966
turn off the overlay
596
00:20:02,466 --> 00:20:04,100
and I'm going to tell it to
597
00:20:04,366 --> 00:20:06,133
undistort and they're both should be set to undistort
598
00:20:06,133 --> 00:20:07,500
so now if we look at both of these
599
00:20:09,266 --> 00:20:10,766
we should seen exactly the same image
600
00:20:10,766 --> 00:20:13,000
and as you can see pixel for pixel perfect
601
00:20:13,366 --> 00:20:14,966
now it should because we're in a digital environment
602
00:20:14,966 --> 00:20:16,600
but still you can see that
603
00:20:16,700 --> 00:20:18,100
by creating
604
00:20:18,566 --> 00:20:20,300
the distortion data
605
00:20:20,800 --> 00:20:22,166
distorting the checkerboard through it
606
00:20:22,166 --> 00:20:23,966
and then analyzing that distortion
607
00:20:23,966 --> 00:20:26,100
has got us the exact same result
608
00:20:26,666 --> 00:20:27,466
now in the perfect world
609
00:20:27,466 --> 00:20:28,466
since digital is perfect
610
00:20:28,466 --> 00:20:29,533
it should work like this
611
00:20:29,533 --> 00:20:31,700
but it's it's cool to see the proofs in the pudding
612
00:20:31,700 --> 00:20:33,266
and it actually is working perfectly
613
00:20:33,966 --> 00:20:34,733
and that is a lens grin
614
00:20:34,733 --> 00:20:36,666
this is gonna always be your ideal way
615
00:20:36,933 --> 00:20:38,266
image analysis
616
00:20:39,066 --> 00:20:39,800
look at a shot like this
617
00:20:39,800 --> 00:20:41,600
image analysis is always gonna is gonna nail it
618
00:20:41,600 --> 00:20:43,366
you're gonna get great results but
619
00:20:43,700 --> 00:20:45,533
if there was more people walking through this frame
620
00:20:45,533 --> 00:20:46,200
if there was something
621
00:20:46,200 --> 00:20:47,566
bigger walking through this frame
622
00:20:47,566 --> 00:20:50,100
or when you're shooting blue screen and you have
623
00:20:50,133 --> 00:20:51,366
nothing to analyze
624
00:20:51,800 --> 00:20:53,900
you know a couple tracking markers might not be enough
625
00:20:53,900 --> 00:20:55,633
for image analysis to solve your problem
626
00:20:55,733 --> 00:20:57,900
you want the one sure way to always get
627
00:20:58,600 --> 00:20:59,666
the lens data you need
628
00:20:59,666 --> 00:21:01,000
shoot a lens grid
629
00:21:01,266 --> 00:21:02,866
as long as you want to make sure that the
630
00:21:02,866 --> 00:21:04,300
squares are all exactly the same size
631
00:21:04,300 --> 00:21:05,700
and all exactly the same point apart
632
00:21:05,700 --> 00:21:06,166
other than that
633
00:21:06,166 --> 00:21:07,200
doesn't matter how big they are
634
00:21:07,200 --> 00:21:07,733
how small they are
635
00:21:07,733 --> 00:21:09,566
as long as they're all exactly the same size
636
00:21:09,733 --> 00:21:11,566
just fill the frame with the lens
637
00:21:11,900 --> 00:21:13,766
shoot it and bring it into nuke and run
638
00:21:13,766 --> 00:21:14,500
run grid analysis
639
00:21:14,500 --> 00:21:16,333
this is by far gonna be your best bet
640
00:21:16,333 --> 00:21:17,766
cause it will work every time
641
00:21:17,766 --> 00:21:19,333
line analysis will work sometimes
642
00:21:19,333 --> 00:21:20,800
and image analysis will work sometimes
643
00:21:20,800 --> 00:21:22,666
a grid analysis will always
644
00:21:22,666 --> 00:21:24,800
always work um
645
00:21:24,800 --> 00:21:25,666
and that goes
646
00:21:25,666 --> 00:21:26,666
uh for I mean
647
00:21:26,666 --> 00:21:27,566
I just can't
648
00:21:27,566 --> 00:21:28,900
preach this enough because
649
00:21:28,900 --> 00:21:30,300
so many shows have gotten much more
650
00:21:30,300 --> 00:21:32,200
difficult because no one shot lens grids
651
00:21:32,333 --> 00:21:33,966
or we couldn't find them after they were shot
652
00:21:33,966 --> 00:21:35,433
or things like that so
653
00:21:35,700 --> 00:21:37,966
uh No. 1 priority for any of those
654
00:21:37,966 --> 00:21:39,300
any of you who supervise the set
655
00:21:39,300 --> 00:21:41,066
get those lens grids okay
656
00:21:41,066 --> 00:21:43,800
so um that's a little detortion tool
657
00:21:43,800 --> 00:21:44,933
now we have a nice clean
658
00:21:44,933 --> 00:21:45,600
undesorted plate
659
00:21:45,600 --> 00:21:47,233
let's go ahead and start checking out
660
00:21:47,400 --> 00:21:48,400
our camera tracker
661
00:21:48,400 --> 00:21:49,800
so let me delete what we don't need
662
00:21:51,500 --> 00:21:52,400
and we still have our rotto paint
663
00:21:52,400 --> 00:21:54,866
which is good because we can use that as well for our
664
00:21:57,000 --> 00:21:58,700
our arc camera tracker
665
00:21:58,700 --> 00:22:00,833
so I'm going to plug that back into our source
666
00:22:02,733 --> 00:22:03,800
and you'll see that we're still have our
667
00:22:03,800 --> 00:22:04,866
undistorted plate
668
00:22:07,300 --> 00:22:08,766
so this is now undistorted
669
00:22:10,100 --> 00:22:11,366
and I'll delete this one
670
00:22:12,500 --> 00:22:14,033
and let's bring in our camera tracker
671
00:22:16,500 --> 00:22:18,266
under 3D camera tracker
672
00:22:18,900 --> 00:22:19,900
and there's lots of settings
673
00:22:19,900 --> 00:22:21,066
and lots of ins and outs of this
674
00:22:21,066 --> 00:22:22,700
but let's just start by getting ourselves
675
00:22:22,700 --> 00:22:24,666
ourselves some data so let me close this
676
00:22:25,500 --> 00:22:28,066
and I'm just gonna click track features
677
00:22:28,966 --> 00:22:30,700
and let it let it start working
678
00:22:35,366 --> 00:22:36,333
and we consider and watch it
679
00:22:36,333 --> 00:22:37,700
you'll see the track of the features track
680
00:22:37,700 --> 00:22:39,266
just like they did one distortion
681
00:22:39,366 --> 00:22:41,000
and because it is gonna take a few minutes
682
00:22:41,000 --> 00:22:43,166
it does go forwards and then all the way back
683
00:22:43,166 --> 00:22:45,466
so I'm gonna let it do that and I'm just gonna uh
684
00:22:46,600 --> 00:22:48,400
pause and come back while when it's done
685
00:22:48,800 --> 00:22:50,400
so my tracks complete
686
00:22:51,000 --> 00:22:52,900
you can see the features I'll track well
687
00:22:52,900 --> 00:22:53,966
if we play through it all
688
00:22:53,966 --> 00:22:56,100
you'll see how many tracking markers were chosen and
689
00:22:56,400 --> 00:22:57,866
they all track relatively well
690
00:23:02,366 --> 00:23:03,466
and there's lots of information
691
00:23:03,466 --> 00:23:04,600
and all these individually will tell you
692
00:23:04,600 --> 00:23:05,800
if you click on them
693
00:23:06,700 --> 00:23:08,066
you'll see the track links being told
694
00:23:08,066 --> 00:23:09,633
how many frames it was tracked for
695
00:23:10,300 --> 00:23:12,300
you can delete ones that aren't very good
696
00:23:17,200 --> 00:23:19,433
you can also choose to display track quality
697
00:23:19,700 --> 00:23:20,800
which is a really useful
698
00:23:20,800 --> 00:23:22,400
you'll see what was good and what isn't
699
00:23:22,400 --> 00:23:24,666
nuke usually discards the ones that aren't for you
700
00:23:24,866 --> 00:23:26,300
but the green is a great track
701
00:23:26,300 --> 00:23:27,666
it means it has a really good track
702
00:23:27,666 --> 00:23:29,100
and it as it gets yellow and yellower
703
00:23:29,100 --> 00:23:31,033
it's getting less and less consistent
704
00:23:31,133 --> 00:23:32,600
and if it goes red it's a
705
00:23:32,933 --> 00:23:34,300
downright bad track
706
00:23:34,300 --> 00:23:36,300
and nuke usually will ignore those already
707
00:23:37,366 --> 00:23:40,900
and this one actually really tracked pretty good and um
708
00:23:42,133 --> 00:23:42,466
and the route
709
00:23:42,466 --> 00:23:43,966
I guess nuke doesn't ignore the red ones
710
00:23:43,966 --> 00:23:45,066
that's that's the wrong thing to say
711
00:23:45,066 --> 00:23:46,166
but because we have our
712
00:23:46,300 --> 00:23:48,600
number of features at a hundred and fifty um
713
00:23:48,600 --> 00:23:50,200
nuke track the hundred and fifty best features
714
00:23:50,200 --> 00:23:51,533
you could find um
715
00:23:51,533 --> 00:23:53,466
so if we added this to 1,000
716
00:23:53,466 --> 00:23:55,100
I'm sure it would take hours to
717
00:23:55,100 --> 00:23:55,933
to find them all
718
00:23:55,933 --> 00:23:57,100
but it would also be lots of red ones
719
00:23:57,100 --> 00:23:58,433
most likely because
720
00:23:58,466 --> 00:24:00,666
it's not gonna find 1,000 great points in this
721
00:24:00,666 --> 00:24:01,866
in this shot to track
722
00:24:02,133 --> 00:24:03,266
now one mistake I made was
723
00:24:03,266 --> 00:24:04,533
even though I have an alpha channel
724
00:24:04,533 --> 00:24:06,300
I didn't tell it to mask anything
725
00:24:06,300 --> 00:24:07,966
I'm not gonna go back and reanalyze
726
00:24:07,966 --> 00:24:08,966
but for future reference
727
00:24:08,966 --> 00:24:10,866
when you have masked out something like moving
728
00:24:10,866 --> 00:24:11,700
moving guys
729
00:24:12,000 --> 00:24:13,933
you wanna make sure you actually have it
730
00:24:13,933 --> 00:24:15,133
tell it to ignore those guys
731
00:24:15,133 --> 00:24:16,833
by telling it what alpha channel to choose
732
00:24:17,000 --> 00:24:17,566
in this case
733
00:24:17,566 --> 00:24:18,766
doesn't think we have any
734
00:24:19,200 --> 00:24:21,700
markers that were tracked by our guy walking around
735
00:24:21,700 --> 00:24:23,033
so I think we might be okay
736
00:24:23,166 --> 00:24:23,866
maybe this one
737
00:24:23,866 --> 00:24:25,666
so I'm gonna go ahead and delete this track
738
00:24:26,400 --> 00:24:28,100
so we definitely can't affect them
739
00:24:28,666 --> 00:24:29,933
and I'll go ahead and take a look at our
740
00:24:29,933 --> 00:24:32,666
gentleman down here and make sure he was never touched
741
00:24:33,366 --> 00:24:34,433
by anything
742
00:24:37,600 --> 00:24:38,966
and this is fine to do after as well
743
00:24:38,966 --> 00:24:40,400
if you don't need to write one you know
744
00:24:40,400 --> 00:24:41,800
there's only a few things you'd have to mask out
745
00:24:41,800 --> 00:24:43,600
there's no reason not to mask out one of these
746
00:24:43,600 --> 00:24:44,733
so I'm gonna go ahead and delete this one
747
00:24:44,733 --> 00:24:46,066
just to make sure it didn't touch him
748
00:24:47,900 --> 00:24:48,866
and this one did touch him
749
00:24:48,866 --> 00:24:50,366
so I'm gonna go ahead and delete that
750
00:24:52,300 --> 00:24:53,733
cause you will sometimes get occasionally
751
00:24:53,733 --> 00:24:55,066
like he's very trackable
752
00:24:55,066 --> 00:24:56,900
you might get a good track Mark on him
753
00:24:57,100 --> 00:24:58,366
but nuke doesn't understand that he's
754
00:24:58,366 --> 00:25:00,133
moving in the frame and not a fixed point
755
00:25:00,133 --> 00:25:01,400
so you definitely wanna make sure that any
756
00:25:01,400 --> 00:25:03,266
tracking marks that touch him are gone
757
00:25:04,400 --> 00:25:05,366
and here's another example
758
00:25:05,366 --> 00:25:06,466
so this is a good thing to do
759
00:25:06,466 --> 00:25:07,666
to go through here and delete these
760
00:25:07,666 --> 00:25:08,600
it's actually kind of a good thing
761
00:25:08,600 --> 00:25:09,900
I didn't mask him out
762
00:25:10,266 --> 00:25:11,133
so just play through
763
00:25:11,133 --> 00:25:11,800
make sure oh
764
00:25:11,800 --> 00:25:12,833
there's another one
765
00:25:16,266 --> 00:25:17,666
make sure we're seeing the whole image
766
00:25:21,666 --> 00:25:22,900
this one wasn't a good track
767
00:25:22,900 --> 00:25:24,100
but we'll get rid of it anyway
768
00:25:27,933 --> 00:25:29,200
make sure none of those popped up
769
00:25:29,200 --> 00:25:30,833
okay he's off frame now
770
00:25:36,800 --> 00:25:38,500
just verify a few more
771
00:25:40,400 --> 00:25:42,366
looks like we avoid him most of the time
772
00:25:42,366 --> 00:25:44,400
anyway alright
773
00:25:44,400 --> 00:25:45,533
so that's because that way of the feature is
774
00:25:45,533 --> 00:25:46,700
you can solve the camera
775
00:25:47,200 --> 00:25:48,200
now before you solve the camera
776
00:25:48,200 --> 00:25:49,266
you wanna fill out
777
00:25:49,300 --> 00:25:51,300
any information you have under solver
778
00:25:51,366 --> 00:25:54,700
so this can come from a camera report
779
00:25:54,866 --> 00:25:56,866
um which the script supervisor will manage for you
780
00:25:56,866 --> 00:25:58,000
and you'll have lens information
781
00:25:58,000 --> 00:25:59,700
you'll have hopefully lens information
782
00:25:59,733 --> 00:26:01,600
potentially angles to start with and everything
783
00:26:01,600 --> 00:26:04,033
but all we really wanna know is um
784
00:26:04,100 --> 00:26:05,366
if we can get the focal length gray
785
00:26:05,366 --> 00:26:07,600
but mostly we need to know is this a constant lens
786
00:26:07,600 --> 00:26:09,700
is it a fixed lens or is it a zoom lens
787
00:26:09,700 --> 00:26:11,200
this shot from everything I'm seeing
788
00:26:11,200 --> 00:26:12,200
it looks like a fixed lens
789
00:26:12,200 --> 00:26:14,200
it doesn't look like we zoom in or zoom out at all
790
00:26:14,200 --> 00:26:17,100
so I'm gonna say it's an unknown constant um
791
00:26:17,100 --> 00:26:19,266
I can also say approximate constant
792
00:26:19,300 --> 00:26:21,733
and then I can tell it what lens I think it is
793
00:26:21,733 --> 00:26:23,700
but it will it won't take that for the bible
794
00:26:23,900 --> 00:26:25,533
if I know for sure with the lenses
795
00:26:25,533 --> 00:26:26,466
I can say it's known
796
00:26:26,466 --> 00:26:28,766
and it's gonna say this is a 50 millimeter lens
797
00:26:29,100 --> 00:26:30,100
nothing else to do
798
00:26:30,100 --> 00:26:31,500
this is definitely a lens
799
00:26:31,500 --> 00:26:32,500
I don't know that's the case
800
00:26:32,500 --> 00:26:34,166
so I'm gonna go with unknown constant
801
00:26:34,300 --> 00:26:35,166
if it was a zoom lens
802
00:26:35,166 --> 00:26:37,366
I would definitely wanna know unknown varying
803
00:26:37,366 --> 00:26:38,966
because I want nuke to do
804
00:26:39,000 --> 00:26:40,266
perform zooms in the camera
805
00:26:40,266 --> 00:26:42,433
so that my CG will track in properly
806
00:26:42,733 --> 00:26:43,766
now the filmback size
807
00:26:43,766 --> 00:26:45,466
although I don't have a camera report
808
00:26:45,466 --> 00:26:47,566
I'm guessing that since this was shot by Epic
809
00:26:47,566 --> 00:26:49,466
PhD was shot with the red camera
810
00:26:49,466 --> 00:26:52,166
so I'm gonna change this to 24.4
811
00:26:53,466 --> 00:26:55,433
and 13.7 which is the
812
00:26:55,533 --> 00:26:57,200
the red cameras film back size
813
00:26:57,400 --> 00:26:58,966
and you can get that information on their website
814
00:26:58,966 --> 00:27:00,066
and most cameras
815
00:27:00,400 --> 00:27:02,366
if you know what camera shot with especially digital
816
00:27:02,366 --> 00:27:03,733
you can look it up online
817
00:27:03,733 --> 00:27:05,300
and find out what the film back size is
818
00:27:05,300 --> 00:27:06,800
and make sure you always convert it to millimeters
819
00:27:06,800 --> 00:27:08,166
or change this to images
820
00:27:08,533 --> 00:27:08,966
for example
821
00:27:08,966 --> 00:27:10,400
the Genesis um
822
00:27:10,400 --> 00:27:11,000
is point nine
823
00:27:11,000 --> 00:27:11,500
three inches
824
00:27:11,500 --> 00:27:12,266
you could change it to inches
825
00:27:12,266 --> 00:27:14,600
instead of doing millimeter conversion okay
826
00:27:14,600 --> 00:27:16,166
now Camera Motion is free camera
827
00:27:16,766 --> 00:27:19,666
this camera might be on a tripod or close enough to it
828
00:27:19,666 --> 00:27:20,733
I mean not a tripod
829
00:27:20,733 --> 00:27:21,466
but someone's holding it
830
00:27:21,466 --> 00:27:23,000
and just kind of shaking it around
831
00:27:23,000 --> 00:27:25,866
I don't believe that there's gonna be much um
832
00:27:25,866 --> 00:27:27,766
dollying or anything involved with this move
833
00:27:27,766 --> 00:27:28,600
but I don't know for sure
834
00:27:28,600 --> 00:27:30,100
so I'm gonna leave free camera on
835
00:27:30,366 --> 00:27:32,000
if I knew for sure this was on a tripod
836
00:27:32,000 --> 00:27:33,133
and someone just moving it around
837
00:27:33,133 --> 00:27:34,566
then I'm gonna get a better result
838
00:27:34,566 --> 00:27:35,833
if I say rotation only
839
00:27:36,000 --> 00:27:36,800
but with this solve
840
00:27:36,800 --> 00:27:39,100
I'm guessing we're gonna get a nodal move only anyway
841
00:27:39,100 --> 00:27:40,166
or rotation only move
842
00:27:40,166 --> 00:27:41,833
so I'm just gonna leave it on free camera
843
00:27:41,933 --> 00:27:43,066
and hope for the best
844
00:27:43,266 --> 00:27:44,966
now keyframe separations are interesting
845
00:27:44,966 --> 00:27:46,000
these are gonna
846
00:27:46,166 --> 00:27:47,200
for most of the case
847
00:27:47,200 --> 00:27:48,033
you're gonna get
848
00:27:48,066 --> 00:27:50,800
the default centers are gonna be fine um
849
00:27:50,933 --> 00:27:51,933
but there definitely
850
00:27:51,933 --> 00:27:53,200
are some powerful features in here to use
851
00:27:53,200 --> 00:27:54,833
when you have a really hard solve
852
00:27:54,900 --> 00:27:56,866
and you're not getting a proper solve
853
00:27:57,566 --> 00:27:59,733
you can change how often nuke looks at the keyframe
854
00:27:59,733 --> 00:28:01,400
you can even set a reference frame to know okay
855
00:28:01,400 --> 00:28:03,666
this is a frame that tracks properly
856
00:28:03,800 --> 00:28:06,133
use this as your reference um
857
00:28:06,133 --> 00:28:07,900
and you can also use 3 frame initialization
858
00:28:07,900 --> 00:28:10,500
which means Nuke's gonna find 3 frames overall
859
00:28:10,600 --> 00:28:11,466
instead of just two
860
00:28:11,466 --> 00:28:12,500
which is default
861
00:28:12,533 --> 00:28:13,566
to find out that okay
862
00:28:13,566 --> 00:28:14,700
these these frames are good
863
00:28:14,700 --> 00:28:15,700
and we're locked on these frames
864
00:28:15,700 --> 00:28:19,666
so use these as a as as frames to to match to um
865
00:28:20,000 --> 00:28:21,700
again most of the time you're not gonna deal with this
866
00:28:21,700 --> 00:28:23,366
but the more tracking you do in anything
867
00:28:23,366 --> 00:28:25,233
you're gonna need to find to some of this stuff
868
00:28:25,333 --> 00:28:25,933
but at this point
869
00:28:25,933 --> 00:28:27,166
I'm happy with the solve
870
00:28:27,333 --> 00:28:28,800
and I'm just gonna go to Camera Tracker
871
00:28:28,800 --> 00:28:30,266
and say solve Camera
872
00:28:31,666 --> 00:28:32,866
and it'll take just a second
873
00:28:32,866 --> 00:28:34,733
not nearly as long as tracking the features did
874
00:28:34,733 --> 00:28:36,566
it's gonna take about 20 seconds
875
00:28:43,166 --> 00:28:44,200
and no one actually has a
876
00:28:44,200 --> 00:28:46,366
reference frame is equals 1:45 oh
877
00:28:46,366 --> 00:28:46,700
it's solved
878
00:28:46,700 --> 00:28:47,300
I had an error
879
00:28:47,300 --> 00:28:48,500
it solved okay
880
00:28:48,500 --> 00:28:50,333
so it does give us our total reprojection error of
881
00:28:50,333 --> 00:28:51,566
1.56 speckles
882
00:28:51,566 --> 00:28:54,000
which is doable for what we're using it for now
883
00:28:54,000 --> 00:28:55,033
I'm fine with that
884
00:28:56,266 --> 00:28:57,533
and so nothing changes
885
00:28:57,533 --> 00:28:59,133
cause we're still just looking at the camera tracker
886
00:28:59,133 --> 00:29:00,633
so we actually haven't done anything yet
887
00:29:00,666 --> 00:29:02,600
so the next thing you do is create a scene
888
00:29:04,566 --> 00:29:05,500
so as soon as I do that
889
00:29:05,500 --> 00:29:07,733
we have three nodes that appear C node
890
00:29:07,733 --> 00:29:09,666
a camera node and a tracker point cloud
891
00:29:09,666 --> 00:29:12,366
so we're just gonna click on seen and press number one
892
00:29:12,400 --> 00:29:14,633
and now we're in our 3D environment
893
00:29:14,933 --> 00:29:16,966
and our point cloud is being generated
894
00:29:16,966 --> 00:29:19,400
and you can see that we are now moving around our point
895
00:29:19,400 --> 00:29:21,233
cloud so if we lock to our camera
896
00:29:21,366 --> 00:29:23,300
I'm just gonna change this camera name to cam
897
00:29:23,300 --> 00:29:24,200
underscore Anna
898
00:29:25,666 --> 00:29:27,333
which I like to do um
899
00:29:27,333 --> 00:29:29,300
anytime I'm gonna use this camera to render anything
900
00:29:29,300 --> 00:29:31,000
any camera any camera nuke is just a habit
901
00:29:31,000 --> 00:29:32,600
I've done this way since I was
902
00:29:32,600 --> 00:29:33,966
I was taught this way 10 years ago
903
00:29:33,966 --> 00:29:35,100
and I just like to keep it this way
904
00:29:35,100 --> 00:29:36,533
so that's good to know
905
00:29:36,533 --> 00:29:37,766
so I'm gonna change this to came in
906
00:29:37,766 --> 00:29:38,766
I'm gonna lock to my view
907
00:29:38,766 --> 00:29:40,833
and just kind of go through and see what we got here
908
00:29:41,300 --> 00:29:43,400
you can actually see looking at through the camera
909
00:29:43,400 --> 00:29:44,700
although just like a bunch of points
910
00:29:44,700 --> 00:29:46,066
you can see that there are some shapes
911
00:29:46,066 --> 00:29:46,666
taking place here
912
00:29:46,666 --> 00:29:48,366
so the top of our buildings here
913
00:29:50,166 --> 00:29:51,633
our little um
914
00:29:52,766 --> 00:29:53,866
bench or whatever that
915
00:29:53,866 --> 00:29:55,566
structure was on the sidewalks right there
916
00:29:55,566 --> 00:29:57,433
and our rounded buildings over here
917
00:29:58,500 --> 00:30:00,200
and our sidewalk and cars over here
918
00:30:00,200 --> 00:30:02,500
so it's we got a good scene right now
919
00:30:03,200 --> 00:30:04,766
and I think if you when you
920
00:30:04,766 --> 00:30:06,500
this is the one place you'll really notice a difference
921
00:30:06,500 --> 00:30:08,400
when you run this through a distorted plate
922
00:30:08,400 --> 00:30:11,100
you'll see that there's some more phantom points
923
00:30:11,366 --> 00:30:12,733
in places that they really shouldn't be
924
00:30:12,733 --> 00:30:14,900
there's really nothing there so okay
925
00:30:14,900 --> 00:30:16,600
so we have our saw
926
00:30:16,600 --> 00:30:17,833
we have our point cloud
927
00:30:18,533 --> 00:30:19,566
and we have our camera
928
00:30:19,966 --> 00:30:21,833
let's go back to 2D view
929
00:30:24,933 --> 00:30:27,000
and let's change our default settings to HD
930
00:30:27,000 --> 00:30:28,566
to make sure we're getting the right data
931
00:30:29,466 --> 00:30:31,800
okay so one thing that I like to do
932
00:30:31,800 --> 00:30:33,366
although this camera actually does a pretty good
933
00:30:33,366 --> 00:30:34,800
job of defining the ground plane
934
00:30:34,800 --> 00:30:36,233
we're actually already like it
935
00:30:36,466 --> 00:30:37,633
pretty much where you wanna be
936
00:30:38,066 --> 00:30:39,466
uh there it is important
937
00:30:39,466 --> 00:30:41,066
you can define a ground plane um
938
00:30:41,066 --> 00:30:42,333
it's not so much important when
939
00:30:42,333 --> 00:30:44,066
you're working by yourself because you know the shot
940
00:30:44,066 --> 00:30:45,300
but when you're working with others
941
00:30:45,300 --> 00:30:47,233
you wanna be able to find a plane of the
942
00:30:47,566 --> 00:30:48,066
of the ground
943
00:30:48,066 --> 00:30:49,300
so we know where to put things
944
00:30:49,300 --> 00:30:50,533
we know where to place things
945
00:30:50,533 --> 00:30:52,100
we know what aren't the camera
946
00:30:52,100 --> 00:30:54,600
so what we wanna do is see if we can find points
947
00:30:54,600 --> 00:30:56,133
there's enough points that are the ground
948
00:30:56,133 --> 00:30:57,366
so let's click this
949
00:30:57,500 --> 00:30:58,633
this is on the ground
950
00:30:58,800 --> 00:31:00,300
and I'll shift click this one
951
00:31:00,666 --> 00:31:02,366
and these points are on the ground
952
00:31:02,366 --> 00:31:03,666
so let's click these
953
00:31:05,400 --> 00:31:07,900
I can move frames and select more points
954
00:31:10,866 --> 00:31:12,266
let's see I think those are the only points that
955
00:31:12,266 --> 00:31:13,500
are actually on this piece of ground
956
00:31:13,500 --> 00:31:15,100
so let's just right click on this
957
00:31:15,666 --> 00:31:17,833
and say ground plane set to selected
958
00:31:19,966 --> 00:31:21,300
so there's our ground plane
959
00:31:21,366 --> 00:31:24,366
and this will reorient the scene based on
960
00:31:24,366 --> 00:31:25,933
the camera tracker data which goes through here
961
00:31:25,933 --> 00:31:27,000
so now we look through a scene again
962
00:31:27,000 --> 00:31:28,433
although it won't change much
963
00:31:31,100 --> 00:31:32,833
we now have our ground plane to find
964
00:31:33,966 --> 00:31:34,566
in the scene
965
00:31:34,566 --> 00:31:36,466
so if we unlock from the camera
966
00:31:36,466 --> 00:31:37,466
and that's what this button does
967
00:31:37,466 --> 00:31:39,000
locking to and from the camera
968
00:31:39,000 --> 00:31:40,333
so when this is locked and I move around
969
00:31:40,333 --> 00:31:42,600
you'll see we don't really get to move around too much
970
00:31:42,866 --> 00:31:44,666
and if I unlock and hold down control
971
00:31:44,666 --> 00:31:46,000
I can rotate around
972
00:31:46,100 --> 00:31:47,200
held that all same thing
973
00:31:47,200 --> 00:31:48,900
so that's the difference
974
00:31:48,900 --> 00:31:50,766
so now you can see that because we chose those points
975
00:31:50,766 --> 00:31:51,900
every single one of those points
976
00:31:51,900 --> 00:31:54,166
was orientated so that it's on our ground
977
00:31:54,166 --> 00:31:55,900
our zero or on the ground plane
978
00:31:56,266 --> 00:31:56,866
and I should have shown
979
00:31:56,866 --> 00:31:58,200
you what it looked like before with those points
980
00:31:58,200 --> 00:31:59,200
but that's okay
981
00:31:59,200 --> 00:32:00,100
these are their ground plan
982
00:32:00,100 --> 00:32:02,166
they're all exactly intercepting the grid
983
00:32:02,566 --> 00:32:03,933
and they were probably pretty close before
984
00:32:03,933 --> 00:32:05,400
but now they're definitely intersecting the grid
985
00:32:05,400 --> 00:32:07,433
exactly so anything we place on here
986
00:32:07,466 --> 00:32:08,566
will be on the ground now
987
00:32:08,566 --> 00:32:10,400
so if I throw in a piece of geometry
988
00:32:11,766 --> 00:32:12,566
just make something
989
00:32:12,566 --> 00:32:13,500
let's make a
990
00:32:14,100 --> 00:32:15,200
make a cylinder
991
00:32:16,066 --> 00:32:19,033
and let's bring it way down in size
992
00:32:19,866 --> 00:32:21,166
and you'll see that it's
993
00:32:21,266 --> 00:32:22,933
spliced through our ground plant
994
00:32:22,933 --> 00:32:23,866
so just bring it right up
995
00:32:23,866 --> 00:32:25,400
so that it's right on the grid
996
00:32:28,466 --> 00:32:30,866
and let's bring it back in Z space
997
00:32:33,666 --> 00:32:34,800
and because we defined our grid
998
00:32:34,800 --> 00:32:36,200
we didn't have to do anything else
999
00:32:36,300 --> 00:32:38,300
I'm just gonna make it a little bit smaller still
1000
00:32:40,133 --> 00:32:41,233
and make sure it's
1001
00:32:43,900 --> 00:32:46,500
directly on the ground plane
1002
00:32:52,700 --> 00:32:54,066
which does appear to be
1003
00:32:56,166 --> 00:32:59,000
okay it's in camera view
1004
00:33:01,366 --> 00:33:02,333
okay there it is
1005
00:33:02,333 --> 00:33:04,666
and let's go ahead and throw color bars on it
1006
00:33:04,666 --> 00:33:06,766
just so we can see what it looks like when we render it
1007
00:33:07,866 --> 00:33:08,966
add it to the scene
1008
00:33:14,666 --> 00:33:16,000
and hit I'm hitting tab
1009
00:33:16,000 --> 00:33:18,966
I'm sorry hitting tab to go back between 2D and 3D
1010
00:33:18,966 --> 00:33:21,100
and let's add a scale and render so we can see this
1011
00:33:22,000 --> 00:33:23,266
so when you add a scale and render
1012
00:33:23,266 --> 00:33:24,400
connected to the object
1013
00:33:24,400 --> 00:33:27,000
scene arrow goes to the object scene
1014
00:33:27,333 --> 00:33:28,466
object or the scene
1015
00:33:28,600 --> 00:33:29,800
and then our camera
1016
00:33:30,333 --> 00:33:33,900
and as long as you um have your settings set properly
1017
00:33:33,900 --> 00:33:36,100
which is what we did just a second ago with HD
1018
00:33:36,166 --> 00:33:38,066
you don't need to connect anything to the background
1019
00:33:38,333 --> 00:33:40,266
um of course if you wanna see what it looks like
1020
00:33:40,266 --> 00:33:41,966
comp on top of the background you certainly can
1021
00:33:41,966 --> 00:33:44,400
so let's make sure we connect to the lens distortion 1
1022
00:33:44,400 --> 00:33:46,466
otherwise it won't be tracking properly
1023
00:33:48,666 --> 00:33:50,600
and although I'm on the ground playing with my cylinder
1024
00:33:50,600 --> 00:33:52,366
I put it too far to the right
1025
00:33:52,366 --> 00:33:55,400
so it's actually not gonna be on the ground because
1026
00:33:55,400 --> 00:33:56,366
or not appear to be on the ground
1027
00:33:56,366 --> 00:33:57,566
because I moved it too far to the right
1028
00:33:57,566 --> 00:33:59,166
so let's go ahead and move that back in
1029
00:34:07,166 --> 00:34:08,633
and let's move it a little bit
1030
00:34:13,400 --> 00:34:15,500
short hits of Z is gonna be facing us
1031
00:34:21,666 --> 00:34:23,300
and bring a closer to us in camera
1032
00:34:30,933 --> 00:34:32,400
and now if we play through that
1033
00:34:34,266 --> 00:34:35,966
it should lock to the ground pretty well
1034
00:34:44,733 --> 00:34:45,533
and it is
1035
00:34:48,700 --> 00:34:49,500
okay so then that's
1036
00:34:49,500 --> 00:34:50,966
if you're gonna put this thing on the ground
1037
00:34:51,366 --> 00:34:53,000
and you don't necessarily have points
1038
00:34:53,000 --> 00:34:54,000
way back there on the ground
1039
00:34:54,000 --> 00:34:55,366
so we just by defying the ground plane
1040
00:34:55,366 --> 00:34:56,600
now we can put anything on the ground we want
1041
00:34:56,600 --> 00:34:58,066
we know exactly where the ground is
1042
00:34:58,600 --> 00:35:00,733
but we don't know necessarily where
1043
00:35:00,733 --> 00:35:02,066
say the top of this building is
1044
00:35:02,866 --> 00:35:03,900
we go down further
1045
00:35:04,466 --> 00:35:07,200
so say we want to put an item on top of the building
1046
00:35:08,000 --> 00:35:08,966
it's not as easy as that
1047
00:35:08,966 --> 00:35:10,366
cause we've only to found the ground plane
1048
00:35:10,366 --> 00:35:12,200
and although we know where this point cloud is
1049
00:35:12,200 --> 00:35:12,866
we don't actually
1050
00:35:12,866 --> 00:35:14,300
we don't wanna have to hand
1051
00:35:14,366 --> 00:35:16,966
you know go find out which exact point this is and then
1052
00:35:17,300 --> 00:35:18,533
copy its math somehow
1053
00:35:18,533 --> 00:35:19,600
it would just be too annoying
1054
00:35:19,600 --> 00:35:20,466
so we need to find
1055
00:35:20,466 --> 00:35:22,533
out where that point is as quickly as possible
1056
00:35:22,533 --> 00:35:24,566
so I'm gonna open up the camera tracker again
1057
00:35:26,700 --> 00:35:28,800
and you'll see that this feature is actually
1058
00:35:28,800 --> 00:35:29,900
one of our tracking marks
1059
00:35:29,900 --> 00:35:30,966
I'm gonna right click
1060
00:35:31,300 --> 00:35:34,200
and say create and I'm gonna have a create an axis
1061
00:35:35,133 --> 00:35:38,266
and this axis is now created on that point
1062
00:35:40,400 --> 00:35:42,366
as you can see there so if we move through
1063
00:35:44,166 --> 00:35:45,566
take a second to buffer
1064
00:35:46,566 --> 00:35:47,800
you'll see that axis is stuck there
1065
00:35:47,800 --> 00:35:49,633
so now we can put any piece of geometry
1066
00:35:50,266 --> 00:35:52,566
attach this axis and it's gonna be mapped right there
1067
00:35:52,800 --> 00:35:53,966
now if you want something
1068
00:35:54,066 --> 00:35:55,400
already created instead of an axis
1069
00:35:55,400 --> 00:35:56,566
let's go ahead and delete that
1070
00:35:56,566 --> 00:35:57,966
let's click that point again
1071
00:35:58,966 --> 00:36:00,133
and you can say right click create
1072
00:36:00,133 --> 00:36:01,366
and let's create a cube
1073
00:36:02,166 --> 00:36:05,200
and collect that cube to our scene
1074
00:36:07,933 --> 00:36:09,400
connect that to Color Bar so we can see it
1075
00:36:09,400 --> 00:36:11,566
and it's way too big so let me just scale way down
1076
00:36:18,366 --> 00:36:19,833
okay close that
1077
00:36:20,066 --> 00:36:21,600
and let me close our camera tracker
1078
00:36:21,600 --> 00:36:22,800
so we can just see the plate
1079
00:36:22,800 --> 00:36:24,466
and the scale and the point cloud
1080
00:36:25,466 --> 00:36:27,600
and let's play through that see how it tracks
1081
00:36:30,700 --> 00:36:31,966
so there we have it just a right click
1082
00:36:31,966 --> 00:36:33,700
we created our cube and there it is
1083
00:36:36,800 --> 00:36:38,500
I'll just play a few a few frames there
1084
00:36:38,500 --> 00:36:40,666
and I'll go back and make sure it does track well
1085
00:36:42,966 --> 00:36:43,866
and there it is
1086
00:36:50,133 --> 00:36:51,566
this is a very brief overview of
1087
00:36:51,566 --> 00:36:52,900
1 distortion of camera tracking
1088
00:36:52,900 --> 00:36:53,666
and honestly
1089
00:36:53,666 --> 00:36:55,166
we can spend an entire semester
1090
00:36:55,166 --> 00:36:56,533
just going over these two notes
1091
00:36:56,533 --> 00:36:57,933
especially the camera tracker
1092
00:36:57,933 --> 00:37:00,500
um and since there are camera tracking classes in
1093
00:37:00,500 --> 00:37:02,266
FXPG that are cover whole term
1094
00:37:02,266 --> 00:37:04,600
that proves the point that there's always so many ins
1095
00:37:04,600 --> 00:37:06,666
and outs with any kind of camera track
1096
00:37:06,900 --> 00:37:08,166
and it comes down to really
1097
00:37:08,166 --> 00:37:09,700
the more information you have from set
1098
00:37:09,700 --> 00:37:11,700
the better and easier it will be to track
1099
00:37:11,800 --> 00:37:13,566
and also the more tracking markers you have
1100
00:37:13,566 --> 00:37:14,333
and you know
1101
00:37:14,333 --> 00:37:16,166
Fusion Blue screen is their tracking markers
1102
00:37:16,166 --> 00:37:17,366
not only on the screen
1103
00:37:17,466 --> 00:37:19,833
but in front of the screen so that we can see parallax
1104
00:37:19,900 --> 00:37:20,466
things like that
1105
00:37:20,466 --> 00:37:21,933
there's lots of ins and out to camera tracking
1106
00:37:21,933 --> 00:37:22,566
and there's nothing
1107
00:37:22,566 --> 00:37:24,433
there's no one gonna be able to cover all that
1108
00:37:24,600 --> 00:37:26,100
in this wholesome in this whole semester
1109
00:37:26,100 --> 00:37:27,200
let alone in one class
1110
00:37:27,200 --> 00:37:28,200
so I'm just walking through a
1111
00:37:28,200 --> 00:37:29,866
couple of the other gimmes that
1112
00:37:29,866 --> 00:37:30,566
are gotchas
1113
00:37:30,566 --> 00:37:32,000
rather than with the camera tracker
1114
00:37:32,133 --> 00:37:33,166
so most of the time
1115
00:37:33,166 --> 00:37:34,066
when you have a decent shot
1116
00:37:34,066 --> 00:37:35,100
you're gonna be able to follow
1117
00:37:35,100 --> 00:37:36,333
the steps we did in this class
1118
00:37:36,333 --> 00:37:37,633
and you're gonna get a track
1119
00:37:37,733 --> 00:37:39,766
especially if you can get the lens distortion done
1120
00:37:39,766 --> 00:37:41,366
if you can't get that done
1121
00:37:41,466 --> 00:37:42,900
um it's gonna make a little more difficult
1122
00:37:42,900 --> 00:37:44,100
but not impossible
1123
00:37:44,166 --> 00:37:44,933
so there's a couple of
1124
00:37:44,933 --> 00:37:46,700
the settings in the camera tracker
1125
00:37:46,700 --> 00:37:48,200
so obviously
1126
00:37:48,200 --> 00:37:49,000
by doing lens distortion
1127
00:37:49,000 --> 00:37:49,600
when we did it
1128
00:37:49,600 --> 00:37:50,666
we had to run
1129
00:37:50,666 --> 00:37:52,166
the feature track twice
1130
00:37:52,200 --> 00:37:53,333
once for lens distortion
1131
00:37:53,333 --> 00:37:55,800
and once for camera tracking when we did image analysis
1132
00:37:55,866 --> 00:37:57,500
so to avoid that
1133
00:37:57,500 --> 00:37:59,100
you can actually undistort the
1134
00:37:59,100 --> 00:38:01,500
the image in the camera tracker
1135
00:38:01,500 --> 00:38:02,566
so you can just say
1136
00:38:02,566 --> 00:38:03,466
unknown lens
1137
00:38:03,533 --> 00:38:04,466
lens type is spherical
1138
00:38:04,466 --> 00:38:05,166
that's correct
1139
00:38:05,166 --> 00:38:06,266
instead of anamorphic
1140
00:38:06,300 --> 00:38:07,700
and click undistort
1141
00:38:07,700 --> 00:38:08,366
and it'll actually
1142
00:38:08,366 --> 00:38:09,300
undistort the image
1143
00:38:09,300 --> 00:38:12,133
and use the features to um
1144
00:38:12,133 --> 00:38:12,766
track the camera
1145
00:38:12,766 --> 00:38:15,433
so it'll actually undistort for you um
1146
00:38:16,700 --> 00:38:18,766
the reason to do this inside the tracker is
1147
00:38:18,766 --> 00:38:20,566
really feeling gonna work alone because then you know
1148
00:38:20,566 --> 00:38:21,866
you don't need to worry about the distortion
1149
00:38:21,866 --> 00:38:23,900
and if you're also not gonna be putting a CG in
1150
00:38:24,066 --> 00:38:26,366
you're just gonna be doing stuff in 2D and that's it
1151
00:38:26,366 --> 00:38:28,966
then there's no reason to really not do it this way uh
1152
00:38:28,966 --> 00:38:31,200
however it's just when you're working with a team or
1153
00:38:31,200 --> 00:38:32,166
you're working on a bigger project
1154
00:38:32,166 --> 00:38:34,400
it's always good to have its own lens distortion tool
1155
00:38:34,566 --> 00:38:36,266
you can see the results easier it's
1156
00:38:36,700 --> 00:38:37,733
you know for sure it's working
1157
00:38:37,733 --> 00:38:39,766
cause you can check it um
1158
00:38:40,066 --> 00:38:41,200
and it's just a little bit easier to do
1159
00:38:41,200 --> 00:38:43,166
plus you get a lot more data out of the distortion node
1160
00:38:43,166 --> 00:38:44,800
you can actually apply to a 3D card
1161
00:38:44,800 --> 00:38:47,133
you can do a lot of different things with it so um
1162
00:38:47,133 --> 00:38:48,866
and we might talk about that later in the term
1163
00:38:48,866 --> 00:38:50,566
but just so you know
1164
00:38:50,566 --> 00:38:51,866
when you're in a rush you can do
1165
00:38:51,866 --> 00:38:54,500
have the camera tracker do the lens distortion for you
1166
00:38:55,500 --> 00:38:56,866
now there's another note that are tab
1167
00:38:56,866 --> 00:38:57,600
the scene tab
1168
00:38:57,600 --> 00:38:58,866
this allows you to change
1169
00:38:59,000 --> 00:39:00,166
the translation order
1170
00:39:00,166 --> 00:39:01,566
rotation order of the scene
1171
00:39:01,666 --> 00:39:03,200
you can actually change the whole scene
1172
00:39:03,200 --> 00:39:04,700
so if we wanted to make this thing bigger
1173
00:39:04,700 --> 00:39:05,566
for some reason
1174
00:39:07,566 --> 00:39:10,133
now the scene's monstrous so if we had done this now
1175
00:39:10,133 --> 00:39:11,366
let's go ahead and take a look at our scan right here
1176
00:39:11,366 --> 00:39:13,333
so our cube went away um
1177
00:39:13,333 --> 00:39:14,900
let's make this 111
1178
00:39:16,266 --> 00:39:17,500
wow this is nice and big
1179
00:39:17,500 --> 00:39:20,100
so now if we right click on one of these points
1180
00:39:20,100 --> 00:39:21,600
let's pick a different point now
1181
00:39:22,133 --> 00:39:23,566
and I say create cube
1182
00:39:24,733 --> 00:39:27,300
that cube is gonna be extraordinarily tiny
1183
00:39:27,900 --> 00:39:29,133
once it's plugged into the scene
1184
00:39:29,133 --> 00:39:31,066
instead of being too big like it was before
1185
00:39:31,733 --> 00:39:32,200
as you can see
1186
00:39:32,200 --> 00:39:32,500
there it is
1187
00:39:32,500 --> 00:39:33,500
we can barely see it
1188
00:39:33,566 --> 00:39:35,500
even though the uniform scale still 1
1189
00:39:35,500 --> 00:39:39,800
so this can be useful when you want your scene to um
1190
00:39:40,566 --> 00:39:42,066
you know be able to bring in geometry that's
1191
00:39:42,066 --> 00:39:43,766
closer to the right size out and do anything
1192
00:39:43,766 --> 00:39:45,566
so the bigger I make this
1193
00:39:45,566 --> 00:39:47,533
obviously the smaller the cube will get
1194
00:39:47,533 --> 00:39:49,600
so if we just go to a uniform scale of 10
1195
00:39:49,600 --> 00:39:50,600
and let's go ahead and delete the cube
1196
00:39:50,600 --> 00:39:52,200
I just create and do one more
1197
00:39:52,533 --> 00:39:54,200
on this will do the same point
1198
00:39:57,266 --> 00:39:58,433
and connect to our scene
1199
00:39:59,400 --> 00:40:01,500
so there's a little bit more of what we'd expect
1200
00:40:01,500 --> 00:40:02,766
so the uniform scale is 10
1201
00:40:02,766 --> 00:40:04,200
and now our cubes are coming in
1202
00:40:04,333 --> 00:40:06,700
closer to a normal size and not filling the frame
1203
00:40:06,900 --> 00:40:08,133
so those are just useful things
1204
00:40:08,133 --> 00:40:09,100
you can rotate the scene
1205
00:40:09,100 --> 00:40:11,466
do things like that just to kind of orientate yourself
1206
00:40:11,466 --> 00:40:12,166
so for example
1207
00:40:12,166 --> 00:40:13,900
if I wanted this built the face of this building
1208
00:40:13,900 --> 00:40:15,566
this is gonna be the heart of my shot
1209
00:40:15,566 --> 00:40:17,366
this is gonna be where I'm gonna put a lot of stuff
1210
00:40:17,366 --> 00:40:19,700
I may wanna rotate the scene until this
1211
00:40:19,700 --> 00:40:21,800
is the ground plane or just select points on this
1212
00:40:21,800 --> 00:40:23,066
and call it the ground plane
1213
00:40:23,666 --> 00:40:24,766
and actually when we do that real quick
1214
00:40:24,766 --> 00:40:27,000
that'll be kind of a fun little experiment
1215
00:40:28,533 --> 00:40:30,333
so just click on a few of these points
1216
00:40:30,333 --> 00:40:32,400
let's go to a frame where we see more of that building
1217
00:40:40,066 --> 00:40:41,000
and the good thing about this
1218
00:40:41,000 --> 00:40:42,600
you can you can constantly
1219
00:40:43,500 --> 00:40:45,500
reset the ground point so let me just
1220
00:40:45,866 --> 00:40:47,200
click on a few of these shift
1221
00:40:47,700 --> 00:40:49,400
shift to try to pick green one
1222
00:40:49,400 --> 00:40:51,033
so I know that it's good tracking
1223
00:40:54,933 --> 00:40:57,066
anywhere on the face of this building is gonna be good
1224
00:41:02,266 --> 00:41:04,300
I can also drag and select
1225
00:41:04,300 --> 00:41:05,900
so if I have a couple of those I know are good
1226
00:41:05,900 --> 00:41:06,966
or a couple tracks I know are good
1227
00:41:06,966 --> 00:41:08,033
I can do it that way
1228
00:41:09,600 --> 00:41:11,200
again try to stay to green ones
1229
00:41:11,200 --> 00:41:12,866
just can you can get a better result
1230
00:41:16,600 --> 00:41:17,466
it's like that
1231
00:41:17,466 --> 00:41:19,800
so now I right click and say ground plane
1232
00:41:19,800 --> 00:41:20,966
set to selected
1233
00:41:20,966 --> 00:41:22,833
I've now reorientated my camera
1234
00:41:22,933 --> 00:41:24,233
so if I hit tab
1235
00:41:24,333 --> 00:41:26,500
you'll see my scene is now completely changing
1236
00:41:26,500 --> 00:41:28,633
the angles kind of confusing in a lot of ways
1237
00:41:28,900 --> 00:41:30,433
and if I take this away
1238
00:41:33,400 --> 00:41:35,266
you'll see that nuke did the best it could
1239
00:41:35,600 --> 00:41:37,366
to make that the ground plane
1240
00:41:37,533 --> 00:41:39,400
not nearly as clean as it's not the ground plane
1241
00:41:39,400 --> 00:41:41,433
I probably picked a couple bad markers in there
1242
00:41:41,666 --> 00:41:42,600
but it did do its best
1243
00:41:42,600 --> 00:41:43,100
and now I want to
1244
00:41:43,100 --> 00:41:44,633
put things on the face of that building
1245
00:41:44,700 --> 00:41:46,100
it reorientated my camera
1246
00:41:46,100 --> 00:41:47,466
and everything based on that building
1247
00:41:47,466 --> 00:41:49,866
so although it's not going to be a perfect
1248
00:41:49,866 --> 00:41:50,766
perfect mix
1249
00:41:51,166 --> 00:41:53,566
it is something that can be used
1250
00:41:53,800 --> 00:41:54,800
in that kind of situation
1251
00:41:54,800 --> 00:41:55,466
but ideally
1252
00:41:55,466 --> 00:41:58,100
if you want to make it the ground plane um okay
1253
00:41:58,100 --> 00:41:59,633
so let's see
1254
00:42:00,000 --> 00:42:01,066
um everything else in here
1255
00:42:01,066 --> 00:42:02,366
I mean I would definitely look at the manual
1256
00:42:02,366 --> 00:42:03,666
there's there's lots of
1257
00:42:03,666 --> 00:42:05,866
ins and outs and and tweaking that can be done
1258
00:42:06,133 --> 00:42:07,366
with the camera tracker
1259
00:42:07,366 --> 00:42:09,466
I did hit on the all the major points
1260
00:42:09,466 --> 00:42:11,500
and I would say 75% of your tracks are gonna
1261
00:42:11,500 --> 00:42:13,866
be able to be done with what I described today in class
1262
00:42:14,300 --> 00:42:16,300
I am hoping for lots of questions on the form
1263
00:42:16,300 --> 00:42:17,633
because this is new tools
1264
00:42:17,700 --> 00:42:18,200
not just for you
1265
00:42:18,200 --> 00:42:19,800
but for me as it just came out
1266
00:42:20,466 --> 00:42:22,400
and I look forward to talking more about it
1267
00:42:22,400 --> 00:42:23,866
and if there's more questions this term
1268
00:42:23,866 --> 00:42:24,666
I'll certainly
1269
00:42:24,766 --> 00:42:25,866
be hitting up
1270
00:42:27,066 --> 00:42:28,266
so those questions
1271
00:42:28,400 --> 00:42:29,200
not sure on the forms
1272
00:42:29,200 --> 00:42:30,100
but if there's enough demand
1273
00:42:30,100 --> 00:42:31,433
I'll do it in class as well
1274
00:42:31,500 --> 00:42:33,533
and just one last thing in here I didn't talk about
1275
00:42:33,533 --> 00:42:34,400
which I'm not gonna go into
1276
00:42:34,400 --> 00:42:36,100
you can also add user tracks
1277
00:42:36,200 --> 00:42:37,100
so although
1278
00:42:37,100 --> 00:42:39,066
usually Nukes auto trackers gonna work great
1279
00:42:39,066 --> 00:42:40,400
if there's any points that you know
1280
00:42:40,400 --> 00:42:42,033
or you can track well
1281
00:42:42,166 --> 00:42:44,066
and maybe Nukes Tracker just can't get enough points
1282
00:42:44,066 --> 00:42:46,166
you can hand track using the tracker
1283
00:42:46,200 --> 00:42:47,700
the regular 2D tracker
1284
00:42:48,366 --> 00:42:50,600
right here and import tracks
1285
00:42:50,666 --> 00:42:52,400
into your camera
1286
00:42:52,700 --> 00:42:54,400
so once you have some tracks done in here
1287
00:42:54,400 --> 00:42:55,433
you could literally
1288
00:42:58,333 --> 00:43:00,633
say add user track and you can import
1289
00:43:00,766 --> 00:43:02,400
um tracks into it
1290
00:43:02,400 --> 00:43:04,266
and you can also import user tracks and
1291
00:43:04,533 --> 00:43:05,600
use that to define the camera
1292
00:43:05,600 --> 00:43:07,366
so that can work as well if you have a really
1293
00:43:07,366 --> 00:43:10,100
really really difficult track and nuke just can't get
1294
00:43:10,100 --> 00:43:11,333
the auto tracker can't do it
1295
00:43:11,333 --> 00:43:13,700
so that's good stuff to know about and um
1296
00:43:14,733 --> 00:43:16,100
otherwise that's it for now
1297
00:43:16,100 --> 00:43:16,900
I guess and
1298
00:43:17,933 --> 00:43:20,100
look forward to hearing your questions in the forms and
1299
00:43:20,100 --> 00:43:21,466
seeing you guys give it a go with this shot
1300
00:43:21,466 --> 00:43:22,600
it's a lot of fun to do
1301
00:43:22,600 --> 00:43:24,066
a lot of stuff you can add to it after
1302
00:43:24,066 --> 00:43:24,800
once you have the track
1303
00:43:24,800 --> 00:43:27,100
so I encourage you to experiment
1304
00:43:27,333 --> 00:43:28,733
and that look forward to see you next week
1305
00:43:28,733 --> 00:43:29,966
where we'll be covering gizmos
1306
00:43:29,966 --> 00:43:31,366
and a little tiny bit of Python
1307
00:43:31,366 --> 00:43:33,300
that'll just be enough to get you started
1308
00:43:33,300 --> 00:43:35,200
and start creating some of your own custom tools
1309
00:43:35,533 --> 00:43:37,633
thanks for watching and see you in Week 3
95646
Can't find what you're looking for?
Get subtitles in any language from opensubtitles.com, and translate them here.