VideoPoints

• ••• • •• • ••••• • •
GEOL6395 Petroleum Seismology - Dr. Zhou 4
Help Tour
© Distribution of this video is restricted by its owner
Transcript ×
Auto highlight
Font-size
00:01 Let's look at Oh, Although I years there. Okay, so this
00:12 a chapter three. You should have older slides. By the way,
00:16 will get the I will get the week material very fast. Uh,
00:25 maybe Tuesday evening. My problems are me very busy. I'm teaching to
00:31 classes. So Monday I have to a whole day teaching. So
00:36 uploading the material for next weekend. a 456 Okay, 456 What are
00:44 next week? We will wear Okay. There's also a good time
00:49 . Review the schedule. We can the tops here. So yesterday we
00:55 introduction. This morning, we talk basic off data processing data analysis.
01:03 afternoon we talk about modeling inversion. ? In detail. I will spend
01:10 section on modeling and then you was dekom because in worse off filtering,
01:18 , filtering conclusion economies undo conclusion. one more time about your physical inversion
01:26 we do kill three. We finished before five. For sure.
01:30 Style of time. Similar to today. Okay. Oh, on
01:37 next week time flies 29. We have war logging set me water.
01:47 many of you guys have background Well, logging anybody? Are there
01:55 here who has a background in? , logging? I'm sure assessing how
01:59 ? I will go a little Little bad. Okay. Okay.
02:05 that's fine. So someone has I'm not so many your first
02:09 Everybody's first time. But you have . That's helpful. Okay, The
02:15 one I want go very deep Talk about mostly about basic ideas.
02:21 how to see sand from San You know? You know, blogging
02:26 . The general idea of that Then I will talk about some of
02:31 water. So many things within Like a coalition conclusion you're seeing again
02:38 . Water. Okay, this is very fast paced classes. So before
02:44 know it, we have example what example Why will be basically along the
02:51 off 1 to 4. Right. all four topics, including blogging and
02:57 on on. Describe your open So you have You should know more
03:02 time. We should just try to sure you know what you're doing.
03:09 afternoon. We go from 1 to . Probably our guessing by the example
03:15 . Probably around, say, 3 Next Friday, three PM Around that
03:21 we finish lecture already. So we started by exam. So in a
03:28 , my lecture will be shorter than because I'm doing two sections. But
03:33 you want, we can have 20 time off. Half an hour time
03:39 . If you wanna have any something like that, I'll be happy
03:44 go through issues. I just wanna sure you will get what you you
03:48 , you need to know before text . And then about 3 30 I
03:57 show exam. I'm a star. will have until the next morning the
04:05 , which is Saturday. Tony, Is that why? So you have
04:12 whole evening if you're nice work or , but it's a tedious because next
04:18 , Saturday have hold the electoral game and that's what they said. You
04:24 run, you can hide right so will catch up with you. But
04:28 that's why it's important for you to get involved on now you look at
04:34 . The screen looked at me and to follow. But I think the
04:38 way is a two way communication. to rest questions. Other revisiting
04:44 as a issue to delay elaborating. right. And that's how the pace
04:51 go. I'm trying to diversify the risk. So instead, off in
04:58 top of the crease, I have , see here 10 times three crazy
05:04 reading. So I'll give a couple here. Okay, so I Q
05:10 Q two all the way to kill . So they're seven Crease, are
05:14 three readings? Okay, so today show the first reading, which is
05:19 here. Basically, this is the I mentioned, uh, name
05:25 Maxim Burger. You know, he was my colleagues Exxon years ago.
05:32 here this is a classic paper about about predictive dekom. Okay, Pretty
05:40 . One of the techniques we're gonna in section topic today have no on
05:47 technique is most useful for zero upset based on the pure period obesity multiple
05:55 city repeating off the timing off But when you have a foul set
06:01 to stretch the fall set, the work is working less effectively. It's
06:07 not working. Eso Schoenberger and Those two guys in exile in that
06:13 , they developed a way to do is through something called a stationary
06:17 Okay, so your job is to the paper, which is in the
06:25 , and then write something. We're . Answer those five questions you can
06:30 once. And this is great in mall collaboration. Fine. Right.
06:35 I'm just going to make sure you those aspect. Uh, the reason
06:41 assigned you doing this way have two . Why? You understand locally what
06:49 doing about predicting dekom. So your about practical issues involved in predicting the
06:57 . Okay, that's one off difficult for this paper. Uh, there's
07:03 other opportunities. This is a very example. Off writing style, although
07:10 still arrow here and there. But think they're writing excellent eso. I
07:16 you can learn a bit more about writing in geoscience, especially to your
07:22 . What? What people typically for example, motivation versus the
07:30 how they're presenting stuff. All so I'm trying to focus on those
07:35 aspect any questions or Commons. I a question. Is there? I
07:47 been able to get onto teams for class. Is there a code or
07:52 that we can use, um, join on teams? That's why I
07:57 don't know. The answer I hear your question is about Team. I
08:01 I'm assuming, uh, I think might be the last one joining,
08:08 . Ah, look, team right do you break? Maybe can add
08:11 as a member. Okay. So we'll try to break. Let
08:16 see. You can do that right . Yeah, that's important question,
08:21 I was worried about people. Hard accessing blackboard. So I'm trying adding
08:27 to the teams because I can also files and they're in teams.
08:32 Okay. Thank you. Yeah, problem. Let me let me do
08:36 . Let me do that during I'm not very fast in doing
08:40 okay? After break, we'll talk this, okay? No problem.
08:45 , because I'm not saying it's Yeah. I mean, the way
08:50 do it, you You tell Uh, I think when we
08:54 you tell me I'm going to try . I'm gonna compare your name with
08:58 I see under enrollment, you name missing. There. I'll try
09:03 do. Look at. Actually, , I'm not sure she's, uh
09:07 not sure she is listening here because has the recording into her own things
09:12 greeting. Maria considered also. I think there's something to management.
09:20 you able to check in team to everything there? Maria? No.
09:26 . Members, uh, there are members. I don't know why you
09:32 know there. I'm looking at right . Actually. Let me stop sharing
09:35 . I can show you what I on the TV. So you see
09:45 team here. Thank you. I'm this incoming members. This is what
09:50 my view. I say owner with Maria, the daughter of a new
09:56 . But if I don't see your here, then you are missing.
09:59 you see them here. There nine . Angela, you're there. So
10:06 are a member. You should give accidents. Okay. I'll check
10:12 Yeah. I mean, this is by your village. They basically based
10:16 enrollment, the automatically toe get through . In another class yesterday someone was
10:23 so I can add I can remember . Remember key here? I can
10:28 , but if you were there. cannot aid it. Could be your
10:34 . Different email. We have wise , remember? They're gonna ask me
10:38 , you know, searching for whatever , for example, students, student
10:44 . It may be your I'd somehow different. I don't know what's the
10:48 reason? Okay, but the Yeah. What's the worst we can
10:56 ? Harmony Investigator. We can ask guys about what's going on. Your
11:00 there. Your first one. I what Sarah saying is at the top
11:04 the search bar, there's an activate . Yeah. Could you click
11:10 Um, nine students were waiting to that. Maybe the reason you're saying
11:26 , see, that's kind of that's of teamwork, right? Teamwork on
11:32 . E got it now, That's the reason. So you're not
11:36 , Everybody. Everyone couldn't get So why do you go there?
11:40 can look files. You should be to see the files that you can
11:43 the the stuff, you know, on the file. That's why.
11:50 talk about the quiz last time? can get it there because you don't
11:55 don't have access. So now 1233 have through their got 33 All
12:02 good. So we're making good This is actually time, time saving
12:07 than time wasting. So let's stop this one. Switch back to sure
12:12 hear, uh, the slides. chapter three. Okay. So you
12:18 what? We're go Are the good through this exercise due Toa ascend to
12:25 will become Wake up now. We're actually more now. Good.
12:30 let's continue the presentation mode or the mode. So let me get my
12:41 color. I'm not sure which So it's always hard to see the
12:46 . I don't know why they got twice. Okay, here we
12:50 Modeling. Okay. I'm going to a general description about the modeling.
12:56 air section from my from my book textbook. Okay. So below
13:03 Defined modeling. This is how I people the way of a definitive definition
13:08 versus emerging. Okay, So just you example of how we define
13:14 but I like using cattle showing So this is the first time we
13:18 about it. So this morning, show you that simple model with the
13:26 ? Uh, coming together and talk the data space. The model
13:31 Okay, so here example Off, , data space model space. So
13:39 have to space. There are spaces we're observing data. We analyzing signal
13:45 and so on. So forth. space is where we deliver result.
13:50 have ah, profile model or sweetie or whatever Model restaurant model, any
13:57 . So let's look process on the side. In the form modeling
14:02 we start from real world geology. , We start from real world
14:06 We try to simplify into some kind model. So, actually, the
14:11 models we have many possible models with flying. Really, geology. This
14:17 super creation process is a really critical geophysics because we not only have considering
14:24 this realistic. We also have a . Is this favorable? Okay,
14:30 those of you that are working on , uh, dissertation. You need
14:36 know that when you proposed a most people, the managers or
14:42 they care those two aspect. Why significance off quality? The other one
14:51 physicality. Can this be done in given time? Resources. If you
14:56 , someday you're gonna worry about that the time, right? So we
15:00 to be certified enough. Let's be enough on deficiency now. So then
15:07 will be able to have some Like if you have a restaurant models
15:14 model, you should be ableto forecasting you will see by means you can
15:21 geochemical E. You're physically, Any way to measure something about this
15:29 , the model response that you can that you can verify in the real
15:34 . So that's the idea, So, through that theory or constraints
15:39 studying, predicting from that model from one model or many, many model
15:45 you're doing, for example, I survey human making a model out of
15:50 densities, subsurface like and forecasting the corresponding by that, uh, entity
15:59 that your theory clearly gonna be the law. Newton's gravitational law theory.
16:06 , but there are always so far party can predict. They're probably cannot
16:11 . So in normal, how to it would be something you can
16:16 Something cannot, therefore tradition, off noise as well as the so
16:21 output or the something you can really . So this approach, a modeling
16:29 keep trying. Different model. Modified to see. Can you predict compared
16:35 data, something ready? Meaning your is right or wrong. This is
16:41 approach. And there are people doing , uh, scanning approach. They
16:47 scan all the models automatically and kind scan for some something they're looking
16:52 So that's a modern approach. In , the other side is universal universe
16:59 investment in emerging ideas. Go Okay, in a way for modeling
17:05 called Also called the model driven model the universe is images more data
17:13 So June data, you start from . You say they're choosing signal on
17:18 rest of the noise. Right. have some signal or signals you tryto
17:23 your signal. You using the data towards your answer, but data driven
17:30 considering signal and noise You can also this a signal driven. If you're
17:34 33 signal, can you derive the for your sub surface geology?
17:40 this usually is more challenging because you've more better about theory and constraints connecting
17:49 When you have three university, you have multiple answers again. Floral.
17:56 , So you have a different models . They're gonna interpret. I can
18:01 with geology, right. So you talk about the good bad about both
18:07 . You're you're speaking. Modeling is because you can easily for modeling inversion
18:13 a little harder in theory and also in the algorithm. Okay, but
18:18 people in geophysics prefer inversion. The is that they think emerging is more
18:28 . Okay, Because in a it kind of fair. You
18:32 you see the reason why you work front of the answer Where are modeling
18:37 kind of can be cherry picking. picking 30 models you like. You
18:41 be ableto to try older models. this is the general idea about the
18:48 . However, uh, people can say they're also substantive part in the
18:56 because you decided where to invert for my model on the What's the
19:02 So it's suddenly is ahead in okay, hidden inside. You also
19:06 some subject to part. But in , people thinking versus more,
19:13 more elegant. Theoretically, that Okay, So initially, usually we
19:20 modeling because regularly, But when we very, very much better understanding of
19:25 we might try inversion. In geophysical work for inversion is always challenging
19:36 you have lots off noise on a to use in every aspect on the
19:43 between the data and the solution. model. Usually we call the EU
19:49 or call that a poorly constraint. you have a lot of possibilities,
19:54 know which ones were wrong? I don't have complete observation. You tell
19:59 a limited your time is limited or Time coverage is limited. So those
20:04 the reasons. All right, This a reason why they need a good
20:09 . Scientists like you guys to figure how to choose a bad answer.
20:15 ? So let's look the model. took a picture of Grand Canyon up
20:21 . Okay, that's really Earth Really? Earth library. But some
20:26 , My jaw. I model this Derek plan model, maybe something like
20:33 area. You know, the uh, very grateful model at your
20:39 will make this other model. For , this is because this layer model
20:44 call this a great model. match greater bottle. Okay. But
20:50 looking assessment sections in three D volume the one b all this one
20:57 I can see some stylish similarity because is all those who are buried within
21:03 rock volume. You can see the . You can see the meandering because
21:09 small, you know, faces off or even some Southern Channel, different
21:14 , different levels. So you see whole bunch of stuff. So that
21:18 you feel that I mean, what model? They can be very geological
21:22 we carried off to our, making a modeling approach us realistic as
21:30 as we can. Of course, balanced by the, you know,
21:34 ability. Like I said, visibility there. So we try to make
21:38 all so easy to manipulate eso. one is actually very high quality is
21:44 image image volume this one, the quality ones, but even put 100
21:49 it has still seen channels here. ? That's that's a channel there,
21:53 . It's not a very high I'm just give you a spectrum of
21:57 politics of quality. But in there are two class off,
22:03 dialect is called them and member models always try using any member model toe
22:10 are are thinking process because that intuition the left one's color take model.
22:17 . The writer one is called a model. As I said,
22:21 E and, uh, a great here. Later model here.
22:26 Those three models so you can see modeling on their good and bad on
22:31 sides. If you're doing computational they always like this kind of great
22:36 . Gaza's machine computers always integrated, that's kind of a fair in the
22:43 of greedy. That's unfair because geological never be like that way. That's
22:48 a great model. Geology models more the layer model. Okay, or
22:52 . We should follow everything. But later model are very hard to.
22:58 called called Addressing a Leveling in the . If you ever compare language,
23:03 are two parts. There is a under the addressing another bookkeeping on that
23:10 of very complicated. Okay, if doing modeling in, will fall.
23:16 , if using this kind modern the method will be like we call
23:20 funnel element methods and write a one call the finite difference. F T
23:31 F E. Final elements find a difference. I'll try to elaborate with
23:39 detail, but for this class, just trying to give you a
23:42 okay? Because every topic here can whole class I This is not
23:47 I don't want to overwhelm you, I think I already over with.
23:52 , here's my example. I just to show you example. Modeling forward
23:58 , if you've done properly, can very powerful. Coming very powerful.
24:03 one example The demonstrate. Okay. was a mine in Utah.
24:10 Okay, actually, people die from money is a collapse off money shelf
24:18 being killed. Well, my one my friends. Well, actually,
24:24 was four years after me in the graduate program. I wrote a paper
24:30 science. This is as a modern in science. Amazing, right?
24:34 think science papers always something big? ? Well, this is the first
24:38 people using sex, mammogram, toe mining disaster. That's why it's a
24:43 . So the science paper, His is Dr Kermit. I mean,
24:47 sorry. Uh, Dark Drager, professor in UC Berkeley. So this
24:54 paper payload. I think the good showing modeling so on the left upper
25:01 is map of Utah. Utah is of the few. State was almost
25:06 . Almost handle exactly this This Okay, so the mining locations right
25:12 in the star you see around Triangle six sessions, for example.
25:17 o g. Right here in the G. His first station. So
25:22 station there are three components. Data . Buddy, pull. She's transfer
25:27 ours radio. Okay. In if you have a you have a
25:33 you have or hear the earthquake. a map earthquake travel towards the d
25:37 T station like this way, coming way on that direction is called the
25:44 . I'm sorry are ours for The direction from South receiver direction on
25:52 direction perpendicular is called a T t's transfers These vertical with the three
26:03 Okay, so t is trans Our radio or south to receiver V
26:11 vertical. So there's three companies. a long period datasets on in the
26:16 . I don't know where I can it. There's one curve. Is
26:20 other curves More, more gray Okay. Why, instead, up
26:26 is the position from model. usually the data is more complicated than
26:32 model. You can see which ones here to my the great one is
26:37 little more common here. So most that's the data. The dark ones
26:43 Most likely the model the original was color actually showed color here Too late
26:48 . So what happening is a He using modeling this Professor Drager to show
26:56 the only solution from, uh, this data is solution off. We
27:02 a crack model is a is a their collapse. So the significance of
27:10 study shows that you can using such grand feeding on holiday. The monitoring
27:17 huh. Collapse your cavalry cavern You can model that. So that's
27:25 the one of the idea of this also off the paper. If you
27:28 , I can show this paper you find in science, so I
27:33 so modeling can be very useful if know what you're doing. Uh,
27:37 is case off off using since my Liddy toe, uh, monitoring what's
27:45 on in mine? Disaster. by the way, this is also
27:49 same way the monitoring nuclear test in days. Why the neutral test monitoring
27:57 Exactly what we do. While I doing my PhD on first few years
28:02 age in the early nineties, I some study about Russian high side.
28:08 really surprised to see this is Okay. The number of sites,
28:13 have three nuclear tests. Yeah. don't know if you guys heard about
28:18 or not. I mean, if look the matter Russia. My
28:22 it's looking very touched. Nuclear all of our place. It is
28:30 . Okay. Anyway, so I'm toe first described detail here about the
28:36 remodeling using re theory. The theory is very intuitive. Were simple in
28:42 sense. Since you guys are most of you, some guy's elaborated
28:46 bit. But even in geology, do the same thing if you
28:50 For example, optical technology you doing the microscope, You talk about our
28:57 index right? Using refreshing index to out what's the reflection coefficients determined.
29:03 kind of material Property? Their article . Right at the same principle,
29:08 called snows Law. The same same physics. Okay on. Then
29:14 can also treating the race by 1.2 3.1 point, Like shooting out
29:19 Source. Two point flung Salto station three point is go from source to
29:26 on to the receiver on their some names for this thing. Shooting,
29:32 economic equation I saw so forth. , so, my citizens, this
29:38 of papers, I can see a for my name. John with
29:43 He's Professor U C L. Um uh, my classmate. He's
29:50 a member off US Academy of So we're famous guy. So let
29:57 show some refusing here idea model give and choosing race on final child
30:02 Re past This is actually very This actually the main thing behind the
30:08 the migration methods Kerkhoff migration we gotta about Took on migration, chapter eight
30:14 seven. So we will talk about . Okay. I'm just giving
30:19 So this illustration off shooting on bending , plant source. You just try
30:25 , not law to see where they're . But if you follow not
30:29 you might not end up in the . You might go somewhere else.
30:32 you're trying to trying to keep Abandon is you started from ST Array
30:38 Shadow receiver. I have been the until the really very satisfied physics,
30:42 is not the law, some other . This is kind of detail about
30:47 . Okay, you can also go a network. This is exactly the
30:52 principle you guys using for your cell . Your cellphone feel that you could
30:59 China. How do they connected signal you for the nearby network off stations
31:04 your cell phone. So they do same thing. This is called the
31:08 path or so this time on the principle. So use that same principle
31:14 segment. Which is it? In years, Really? Sometimes even even
31:20 the Chilean model. Okay, there's lot of principle in detail here,
31:25 I'm not gonna kinda give you guys much of this. Their issues
31:29 Okay, I just give some illustration . This is just a numerical
31:33 This is, uh, the network or the shortest path methods. Put
31:38 software put down the receiver. Here surface are zero R one R
31:44 You can see the re pass. . I put some also model in
31:47 . Also put some reflectors in black . One reason I'm showing this to
31:52 to showing that even the Superman, way many taxable concepts, Uh,
31:59 a black and white anymore. They're . For example, in textbook,
32:04 no difference between reflection under a fraction O hara waves reflection. You know
32:13 difference. I'm sure. Right. in reality, they might be
32:20 They might be between all right. example, this, uh, let
32:26 see. I can move my stuff . This point here, you
32:29 Yellow re pass. This is real . Okay, this is refreshing.
32:38 , this is a reflection or So in reality, the different between
32:46 . Reflection might not be that black okay on for for small offset is
32:58 flashing. I have a very foul the difference become become, you
33:03 smear. It's true, even for ways. For example, body wave
33:12 surface data, we'll have a very offset. The difference between the two
33:20 smear. We have extremely wide Your body will become just except when
33:25 combine. Okay, The physics, continuous process. But what we conceptualize
33:34 we often kind of simplified. Just I force you guys to give answer
33:39 in the crease. We're often the is not another one. Exactly
33:43 It's somewhere between. Okay, So really earth Also you have global model
33:50 can see re pass, uh, because the center of the earth the
33:56 the other core in the core the of the Earth have very high velocities
34:02 . Okay, so this phenomenon This is an example of Regency you
34:08 re path. There's a low velocity here. No velocity. I'm showing
34:13 last model. But we lost the not only three layers, but also
34:17 lateral. So this low last If you're shooting the first rivals,
34:23 see shadows loom over survivals. but however first are overshadowed doesn't mean
34:30 no wave inside here. They're actually inside there. Which letter?
34:38 in fact, first rival are not in terms of and produced. If
34:43 doing migration, you can see that off shallow. You have very low
34:49 . This is a way for Okay, Uh, this is one
34:52 my students I was teaching you he remodeling in this study. We have
34:57 time for for exercises he didn't. modeling modeling world. Okay, But
35:03 can see that similar phenomena, For . I'm just giving example. I'm
35:07 drawing here like conceptual. If you away from getting from left, try
35:15 way. Traveling. Okay. This where we're traveling. If you have
35:20 model, homogeneous will keep going on . Keep going as planned.
35:26 Okay, Just nothing changes. Nothing . No problem. But if you
35:32 lost the e commerce in 19, me tell you. Paint color.
35:37 , I have. Ah, uh I have an army here.
35:43 I'm here. Okay. My background will be zero. Okay, now
35:52 here is velocity slower. Press Slower now. He's also but he
35:59 slower. Both the same amount, slower. I don't I don't care
36:02 much about Okay, What happened to were from, so I'm the joy
36:09 . Make sure your comments, the comments about the idea. Okay,
36:13 you were from will keep going on change color slightly until here you will
36:18 still going the same passion because you see anything. Army. All
36:23 keep in mind the area inside. . So what happened? Why were
36:27 good here, going to throw down little bit. Also slow down.
36:31 here on all sides. Still same on skills throwing down here,
36:36 You will from gonna have this effect they're slowing down. So this is
36:40 traveling 10. They won't go fast you go further. Similar idea.
36:49 see the concept now the issue is you have a current room from what
36:57 to the energy. Okay, if choosing one units in the world from
37:04 energy in the world, fast, cause the energy. All right.
37:10 when you have a slowing down slow global off the area, the
37:15 from direction traveling will recovering convergence, be convergence on waves. Right?
37:24 us? In the other relatively higher . Because also higher than the BS
37:32 will from the diversion, because you . Okay. So in terms of
37:41 off waves up to the waves, go through low velocity? Your amplitude
37:49 to increase one goes through high Um, political decrease is the
37:56 because the convergence and divergence this is reason for we talked. Uh,
38:04 called the Velocity. I mean, wave guide a wave trap, travel
38:10 or guide of wave Blame trial. , this is what I'm not is
38:17 important. Let me give you an example. You've seen people uh
38:25 shepherd or say somebody, uh, have horse. You have a stick
38:34 now you have web on top of . You can swing that. They
38:39 swinging wrong, right? So your off web is high velocity. The
38:47 itself is slow. Well, once high once Law, you ask
38:53 why use my hand here about my holding here? I'm Sui Mai my
38:59 . What happening? I'm shaking this . You can see a much bigger
39:05 on your your your sweet your sleep this. Your stick only taking
39:13 I'm just finding your physics okay for same amount of energy, the
39:19 different amount The phenomenon is in the area or high impedance. You're shaking
39:26 a smaller in the area of the impedance a local. Are you guys
39:30 more just like you have? rocks contains the fluid inside. When
39:39 rocks, the rocks is a So they don't move as much but
39:44 to move a lot more because they're salt. All right. Same
39:49 So I go from very simple convergence from in the cartoon showing here,
39:55 here. So the my sweet as shown here, illustrate to you
40:00 slow lost area actually contains more There are more energy here in low
40:06 than the high velocity, but they're later on. So, internal
40:12 the high amplitude energy are where they ready to lower incidence. Early
40:19 actually, the low amplitude. Any ? I wanna make sure you guys
40:26 the phenomenon here because a lot of is about interpretation, but I want
40:30 assure you see the phenomena. This morning I described using three triangles
40:39 metals, model way three. You , the simple model for one shot
40:45 , but I was using first If you have lost these linear
40:49 this is linear greater model. You see the last model of local like
40:52 one deal ingredients. If the job this model will be like last year's
40:57 of the Okay, this is The what's model Businesses Constant,
41:05 Brilliant. This value here is that on top This is the zebra.
41:12 confession is greeted is the slow You can prove in theory, if
41:19 model is linear ingredients now, you passed off first. Arrival will be
41:23 part off a perfect circle like those . Part of each circle is exactly
41:28 right path for that. For that model, we can derive the
41:36 Those kind of question was the first to buy a famous guy. You
41:40 know, uh, you probably heard Mojo Mojo, this community model.
41:46 ? Which? He's the one first got discovered lower bottom of the crossed
41:53 the mojo. But his knowledge geology seismologist. Okay. And he studied
41:58 re past. That's how they found mobile. His one study those
42:04 So this is how we computing the path. Also, how we estimate
42:08 far we can estimate in the last . Which is why this morning Why
42:13 those three triangles? I'm saying the is you cannot go very deep
42:18 Hardy was deep. Well, if offset, give maximum offset. Roughly
42:24 tips about maximum one half, usually one third. That's how far you
42:29 . Okay, so I can get role of some principle from those simple
42:33 . I'm just showing something example of doing riches. And this is one
42:36 exercise. The tracking is the in kind of model he tracked this,
42:41 , are off re past locally ingredient . This is three D model.
42:46 was working on this long time ago the real data Example. Car abandoned
42:50 your examples You can you can you see the part in those models on
42:55 arrivals is another issue. Why don't letter of origin velocity? The re
43:01 become multi multi patting, which means source toe anywhere you'll see more than
43:09 arrival the same time. They cannot it's a publication that those areas
43:15 They all have a saying All this screen cup with us are repairs coming
43:20 same time. So other not the off other cultures. So you have
43:26 than arrivals. That means they're diverging so on. Just I was
43:32 here a moment ago about the convergence , right? How about convergence?
43:37 here. Remember those divergence convergence. we have a convergence. You have
43:43 rival Also multiple rival coming Same So three passes, no unique.
43:49 can coming across every direction give you same arrival. So multi pattern is
43:54 the issue that makes things complicated. also give the opportunity you have multiple
44:00 . That means they are responses off medium Geico marginality. So you might
44:06 able to use them to figure out going on. Okay, I'm sure
44:12 parts here so they can curb you wrong. Okay, let's move on
44:18 something else. May be more This is Ah, well,
44:22 I rotate everything but 90 degree. , not me. The guy.
44:26 is the same Sugar Mike Bloomberg did lot of different things. So rotated
44:32 this way. So the first uh, won t log.
44:37 log A. So, you I can give imprudence and give you
44:40 refreshing condition. So here's conditions. study, Mike on levin, they
44:47 to demonstrate the issue about multiple Okay, what's multiple trace?
44:55 even along Well, logging just One world war. If a layer
45:00 earth model, you can still have multiple inside. So most people doing
45:06 sensible grand, They're assuming they can the multiples, right? Because that's
45:12 . For example, we showed how do convolution. We come always late
45:17 the ref activity. Why do that ? You're assuming you only care about
45:23 reflections because we don't think about the falls. Right. But within the
45:29 media, you can have a bouncing multiple times. Right? So the
45:35 here means the contribution from multiples. is a top toe, bottom to
45:42 right, so you can support their multiple. But if you're going on
45:46 time of death, they were getting and more energy. So there's a
45:52 increasing multiple amount with depth. That if you go very deep, you
45:57 ignore that. So then I have to your well logging forecast. There
46:08 other complications. Okay, so this the modeling study using re path compared
46:13 will form in the will. For , like this d here.
46:19 You see real forms. The problem real form is it's very hard to
46:23 where the wind from came from. they go through, you see a
46:28 off way field, but where they from, you don't know. So
46:33 often combined will form with three paths give us some better understanding about about
46:40 going on. Okay, this is miracle for modeling the model because it's
46:46 model there. Let me show you thing they call a double bonds the
46:52 . Okay. S o. I know where source. Let's say let's
46:58 this is all sorts here. He going down, down, down pounds
47:03 , up here. Okay, both . So here's receiver Just a
47:14 What we call that. That's not multiple, right? What is
47:18 What's the double bounce away prismatic See? Somebody knows. Answer.
47:25 guys hear him? What is called where inbound. It had multiple bounces
47:36 the subsurface counts more than once. , but I heard that. But
47:41 is kind of prismatic away? Because shipment prison, what was the result
47:48 any way? Because maybe it simulates optics in the in a prison.
47:57 r. S m. Look the of this. This thing is that
48:02 like prison. Crystal Wright. That's the name came from. The ship
48:10 at prison. You can have a situation. For example, you can
48:15 I a niche in the model. glossy here low. What do you
48:22 ? Hi. I'm sorry. This coming. So you have women
48:27 Ability. Any color. Let me . Color will quickly here. So
48:34 my sauce here, you sort of . Okay. Guess what happened.
48:47 problem. See that ping pong ball the war? That's also prismatic way
48:57 it can bounce in multiple times. is a crazy waves. It's like
49:00 with God inside. This is also example. What about the wave
49:05 Will track energy is multiply inside. is a theoretical, but they're real
49:12 off this very, uh, big for human life. Uh, none
49:18 you guys were born before 1985 I'm gonna I'm gonna expose you my
49:27 . My daughter was born at a , actually. Did I live?
49:35 . My daughter was born before Let's see, I better know that
49:40 daughter will be very unhappy. I forgot her. Your bowl. She
49:44 born that year? Yes. Okay, I'm good. Bad.
49:47 remember she was born that year. , he was born a year four
49:56 . Okay. Somebody admitting his age . Similarity. Good. Yeah,
50:04 , there is a big event that in seismology. There was a huge
50:08 happening in Mexico, The in Mexico . The earthquake epicenter is about 100
50:16 south of Mexico City. Still, the 10,000 people die. Imagine this
50:25 , Houston, if you see 10,000 die in one day by earthquake,
50:31 think that that the whole city more 10 million people at the least,
50:39 million people scared to death on the , somehow only small promotion their
50:46 But most people was scared. You , No home, every damage building
50:51 . You know, no life like war is a disaster. Shoot a
50:58 . I was a psychologist in and I was very busy in a
51:02 of news media interviews. But I'm you, that's example off with
51:08 So the city was sitting on top this. Ends in the lake is
51:14 change in the lake. So the subsidy mints beneath the city. Many
51:19 city on earthquake In the selection of , that's where the sauces, but
51:24 God trapped energy. So lots of Well, certainly there they are very
51:31 in two reasons. one. They're high altitudes to demotion. Also,
51:36 Besar. Lots of low waves, bouncing waves. So there are very
51:40 component of lateral motion. Modern Almost every building is against gravity.
51:47 build a building toe against vertical. , so you have very cold.
51:53 down. Your really usually okay. building works for building most buildings.
51:58 motion. Sure emotion while strong share building can stat. Okay. I
52:07 , I even heard certain Amber seismologist even heard horror stories long time
52:12 In one respect in China Tangshan Earthquake will say one guy sleeping the higher
52:18 off story building. Not very high , maybe only five story high.
52:25 during the earthquake his bed was throughout the window, the buildings shaking.
52:32 still in the window. When they on the ground, he landing on
52:37 of the roof. Survived He was out a window building collapse Hey,
52:45 on top of beauty. Can you ? It is crazy phenomenon. But
52:53 people died that that earthquake even killed , Many more, I think was
52:57 a quarter million people killed. Yeah, So it's a big
53:03 which is one reason kind of motivated to study of seismology. But
53:07 for you guys, I'm just passing for the modeling. Okay?
53:13 Maybe I mean, I watch my . I don't wanna be telling too
53:18 story and were carried away. Now, this is a good
53:22 Show you modeling. Now this state a very lousy This is, uh
53:26 don't know how many How many? . Japan's island. Northern Hokkaido.
53:32 , So here's the Sebastian. You see the trench. Okay. Political
53:36 Pacific Ocean Trench. Here. Here's . Those number here, like S
53:41 s A S five s four. those are ocean bottom. Such grand
53:46 . B. S. Okay. the data here is showing. Sees
53:50 . That has a very hard to . Anything is very lousy. So
53:54 only thing when when you have a of data set, you cannot do
53:58 working anything, you only only doing . So these predictions are modeling from
54:04 model on model on this side. from this model, they're predicting
54:10 uh, events. No. Only rival, only first rival a
54:17 three path. Reducing the parts in model, you can see they
54:21 Forecasting First rival Wiggles similar pattern as first viable of data. So comparison
54:29 those two. Give them verification off model off B. So this is
54:36 study. So modern studies very useful you have a good idea. What
54:42 what's going on on your dad has lousy okay, their laws, they
54:47 getting played a model because this is bottom. There's the oldest ocean
54:51 ocean. There's so much you can there. There's no observation for deeper
54:56 . This is a very deep This is a cross. Okay,
55:01 Area and second Graham Understanding some background you Very careful observation. I'm just
55:10 you one detailed example on give you break. This is where the model
55:18 model. Okay, on top aides . So we have a veggie
55:23 There's there's low. Last year's over . Right? Wedge. I will
55:28 a sauce here. We put a of the receivers. Okay, I'll
55:33 you to so small ground on B coming together. What is a
55:38 She is also the same glider but heart of the components. So
55:43 there since. More ground by meaning are wiggles off time record on
55:51 So what it calls time? This a problem. We call this data
55:56 , right? State of space. you cannot tell what the model looks
56:01 even though you're not model. for example, I'm asking you see
56:04 wiggle here? What are the What the What are those Wiggles? You
56:08 know when you see the model? we're often we have to spend a
56:13 of time toe in detail. Toe the snapshot. Okay, so I
56:20 you here below in P E f de en er to time.
56:26 This is 18 seconds. This is seconds time We talk about what you
56:33 at that moment on the snapshot off way progression, for example, 1.8
56:41 . What do you say in the component? You see a rival events
56:45 like this. Okay, if you with 1.8 to 3 point all,
56:50 can see the same events advanced. ? So that one that one that
56:56 are the ones to here. So those actually are reflections events. Ours
57:06 . T is the transmission or refreshing dso depend which wanna talk about
57:11 I mean the numbers index of the . So you have toe looking those
57:16 forth, But the main point I to show you Yes, this dash
57:22 here in the n B. It's same location asked. That's not
57:28 Indeed. So intercession here. Others in the same point. Okay,
57:38 this is how we figure out what's on because we fire, even though
57:42 two D model recorded here is So the result is a three D
57:47 three D volume of data on what I'm saying is a bi on
57:53 intersecting each other their slices off this volume you have X, Z and
58:01 . So depending which slides you cut cake, you see intersections. Now
58:05 how we figure out what's going We human beings, we can see
58:11 very easily like a profile on D . Those profile. But we cannot
58:17 this data spacing at all because we . We don't use to see that
58:24 live in the physical war. We only a snapshot. We don't see
58:29 time history of things. Let's Let's all our genetics consider. But
58:36 have a train. If you have technologist, you have trained sell to
58:40 a lot of modeling together. You start with a simple model.
58:46 have the software to immoderate. You simple model. You're observing what's going
58:50 . It's actually mostly geometrical. It's very difficult. Okay, I'll just
58:58 you a test of what's going I'm not require, you know,
59:00 this stuff. I'm just expose you those phenomenology. Okay. For
59:05 your system stands out here. Here's modeling. So this is basically by
59:10 will form. You can make a model. Consists off three D
59:16 Assuming waters constant media, constant media That's fitting. Wave here Threw
59:23 most Islamia Butoh, Sebastian earthquakes. you know I'm joined profiles about,
59:30 know, Here's ocean floor subduction. , here's trench. Here's back
59:35 So your whole slab going down like . And if you look in the
59:41 so you're for the motion will be is good for you guys. You
59:46 are geologist. So you select on . You're gonna be ah, thrust
59:52 , right? Makes sense. So deep emotion off. This water about
60:01 can be through the air because this off this thrust motion, squeeze the
60:07 water upwards. Just squeeze water upwards that the water from the whole area
60:14 , spread out is a trench squeezing entire bottle Water. Keep in
60:20 Let's try very deep. This is . You talk about eight kilometers
60:26 Be squeezed. Operators. I think center part of that. Maybe you'll
60:29 . Huh? I don't know. m high water, 1000 ft high
60:35 . Okay. Through the air. Saudi tsunami happening. All right,
60:43 we have models by way for There are many different ways. I
60:48 wanna go. Detail I mentioned Find out a difference. Which is
60:52 great model. Find the elements, is a layer model. There are
60:56 other friends that we kind of This is basically doing the same thing
61:01 the have case Maze. We've number space. Okay, I'm sure.
61:08 what? I did a long time with students, uh, different supporting
61:12 and shown real forms. Well Which was done this way. Five
61:18 . Basically five different in the Basically using approximation for a different
61:23 Instead, off differentiation, right? equations, differentiation. That was just
61:30 you the numerical answer. Here's some This case People try to handle the
61:37 artifacts called a Barbary reflection. So 32 ways or handle this.
61:44 Here's another study about student PH. . This is PhD by this
61:49 Yeah, on her entire PhD Got modeling. Okay, so I'm
61:55 example people can read it. This a good school. This is the
61:59 school. So I got my Caltech. Okay, It's PhD
62:03 but it's the bottling, very challenges has a very lousy this. Is
62:08 a God comments? Alligator. But is across the images. They tried
62:12 model area in California and she I'm not sure her final model this
62:19 final. Her final result. The models for final model. Okay,
62:25 a profile of the last line The source this north, that's that's
62:31 a lot of force control. So was able toe keep the first rival
62:37 like you see here. Andi, this is Hadera. So the the
62:42 here is the data. Everyone, think like if you come here?
62:46 15 comments are gators. She picked for her all the time that she
62:52 doing. Careful modeling. Well, modeling. Here I divide this kind
62:57 model on top of their data fitting sort of the one sort of Caroline's
63:03 deadlines for position from this model. modeling can also produce good results.
63:09 in case there's no data. No data, Only one data.
63:17 There's no room for your version or processing. Just modeling. I'm gonna
63:23 him, or I think we're done this as I can also do.
63:27 official modeling. I was a promising a break, but I keep
63:32 so I better give a break. could official modeling. You can join
63:37 . Here's one example for small. mean, I think I'm done for
63:40 sector. Okay. I cannot fly here. I forgot all time.
63:46 . Forgive me what time we have to seven. So let's give you
63:52 longer break. How about what do like? 15 minutes? 30
63:58 Right. So 2 to 20 we'll again to 20. Thank you,
64:06 . Right. Yeah. Thank So we're gonna talk about the the
64:17 to answers. Okay? Every day's with the two of those. So
64:24 we go. Uh, don't take answer off. Absolutely. Correct
64:33 Because this is always issue we can . Okay. So we can discuss
64:37 this. So go ahead. And , are you sure to record this
64:43 you are uploading the everything on the point. If you want toe upload
64:49 who is this as well or because you are recording that. This
64:55 Are you recording is old. You are recording this box. It's
65:00 . Recording this party. Okay? big deal. No problem. Thank
65:03 for reminding me. E understand what saying. Yeah. Yeah. We
65:08 worry about people recording and then Okay. That's why we way talk
65:12 this. Uh, this is basically quest to answers. I'm gonna go
65:19 following the style. So the first is a false because all the correlation
65:29 normalized, although coherence. So, speaking, they're not exactly the
65:36 Okay, The only case they're the wonders through the both waves involved has
65:43 been normalized. Eso That's the reason this exercise number two multiple reflections are
65:51 faster. The stacking velocity, the That's false Because other Iraq primary reflections
66:01 velocity than the multiples. Because they're deeper. Whatever will stay in the
66:07 heart. So they're lower lost. speaking third one nearly old feel they
66:14 processing, producing both signal on That's always true. Okay, for
66:22 , statics correction is less effective than statics. Okay, The key is
66:28 I mean by effective. Okay, video usually means a big, big
66:35 . So, uh, it is Okay, Trip, Lismore, but
66:41 can be wrong. Okay. Effective the sense off to be difference,
66:45 you can be or do it. were the wording here is tricky.
66:50 don't like this effective. Next time attending toe big change or something like
66:54 , it's a bigger change, Five stacking velocity profile smoother.
67:03 Now the interval velocity. That's because average assumption involved Six other correlation off
67:11 sweep is record as false as we in the slides. This is Claudia
67:18 . Which one question here. Computer language. Okay, Riker clutter.
67:25 to go off stacking. Refreshing. , Improve, Snr. And
67:30 I mean, covering the events. , there's no new surface aesthetics problem
67:36 processing Brenda hard. That's a false even though there marine marine means anything
67:41 the water in the ocean, there shallow water. They have a strong
67:46 problem that God Mexico show a lot very big problem with the static
67:52 Okay, next to a compare for differences Colijn co evolution. Yes,
67:59 different in formula. But the meaning the main difference. The meaning correlation
68:05 the similarity between the two time but the conclusion is the multiple multiplication
68:11 you, Cherie filtering very different from . Okay, so they're they're physical
68:17 very different. Make her versus Carter bricks, both zero face.
68:25 but the former the recur is the future velocity record off impulse, which
68:32 import response. A letter is Thea Correlation off sweep. Okay, like
68:37 sweep. So there are very but they are origin. A
68:42 Why is the impulse response off spike otherwise just came around the correlation or
68:54 last one? I know. How you have found completing the metrics?
68:57 you forget your refresh, your linear , Maybe that's the way to look
69:03 . So here's what I draw Okay, Put a P three in
69:07 metrics. A pre yellow color here highlight. Right. That's the one
69:13 given on the A four vector. we become involved with multiply.
69:20 while multiply magics with better if the roll a time, multiply with corresponding
69:27 The Adam together Okay, for 0.5 multiplies with minus 10 minus
69:36 That's that's why I'm as a pawn because everything else is here. Zero
69:40 of all, old zero, The second role multiply the first one
69:46 zero. So it's medicine one times 11 0.550 point 0.25 So this is
69:57 adding one plus 0.25 So you follow I'm talking about here. One by
70:03 , you get this answer is massive point 25 but at 1.85 1.71 point
70:11 1.711 I also applaud this so you see this result of plot and this
70:20 ah convolution. If you ask you same question, but for cross
70:30 What do you do? Okay, you ask waiting, because this time
70:34 of them are asymmetrical. April peacefully as medical because it's reverse the order
70:41 all that. For example, you the same metrics. Then change the
70:45 here into reverse order. You can one minus 3.70 point five and minus
70:53 negative result that resolved the cross So across colleges are very similar.
71:01 convolution just reversed order of the while futures involved reversed order. You get
71:10 most efficient. Yeah, sure. , I guess the question Because you
71:17 a rusty with the linear algebra. , when you multiply metrics with a
71:24 that music are doing in the you do one draw a time 100
71:31 into a top product. You have factor that product, which means you
71:38 correspondent elements, then soon appear For example, if only you have
71:45 here. In this case of a 0.5 multiply With this, the first
71:50 0.5 rest off them 00 cannot get . So only the first one non
71:57 multiply with the first one in the minus one that give you minus
72:05 Right. The second hero in the is the second column in the
72:11 which is minus 10.5 and older So there to non zero elements,
72:18 first two elements, non zero. they're gonna multiply the first two elements
72:23 the vector off a four. So is one times Venezuela is Why
72:29 Why? 0.5 times 0.50 point 25 , the 1.25 Because to act
72:39 remember the product the 30 roll not this 10.543 elements non 0 to
72:49 . So you gonna multiply one by with the first ray off a
72:53 So one first of all elements one minus one Give you minus one signal
72:59 Madeline times 0.5. Give you minus , which is a total is a
73:07 of 1.5 31 0.5 times minus So minus 0.75 So that together give
73:17 minus one point. If I makes . So you just one by one
73:26 and then some of them. This kind of thought of product to better
73:32 . One by one, multiply that them together. Be careful with plus
73:39 . So make sure to sign. is that Clear it out. Anything
73:48 , folks? Now you can see lens. That means there's one thing
73:55 the length of the first one is number, the lens of second
73:59 One number in this case, the 1463 at together, that is not
74:05 past 37 The total lands of results be the dire lands total of the
74:12 to Manus one. So the lands results will be six in this case
74:19 four plus 37 minus 16 So I see you can see I've got a
74:24 number here, right? My result six numbers. Non zero. If
74:31 have a two vectors, a two serious multiply in sense of collusion or
74:38 correlation. The length off the result same length is the length of the
74:43 one plus one. The second one one. No. Okay, they're
74:52 always more detail. But again, is how you quantify the result of
74:58 tary. If no more questions, gonna stop sharing. Go back to
75:03 chapters. Okay? Now, talk the chapters. Uh, we are
75:14 . Move on to the section two this afternoon or chapter three, Section
75:22 . Okay, so I just finished mom go about the first one.
75:27 about modeling. Uh, there's some we can talk back. That's not
75:32 worries. We're gonna talk about in filtering or decomposition. Okay, so
75:42 is my starting point. Uh, slides off section tool. The first
75:49 is what we say this morning about conclusion. A reflectivity X, come
75:55 with the filter. Give you which is your data sets. But
76:01 we want to do is wrong. wanna undo the process. Because when
76:05 started in, we have data. is why eso we start on why
76:11 want to remove the future. We to get it back. Relativity.
76:17 is how we try to enhance your . Okay. So, to begin
76:21 , I want to give you some about why I do this. Here's
76:26 . Why do this? Here's a of migration. Okay, for one
76:32 good examples I have to. Somebody here just showing the area because somebody
76:38 wanted to enhance the resolution because they toe the Georges. What's a more
76:44 ? So they start using that they using resolution. I mean, Deke
76:50 to remove a minimizing the regulator so can see better detail. Okay,
76:57 is before that. The convolution Here's I'm gonna fly back, for you
77:04 see it. So those are the . Why do this? People call
77:13 migration. The conclusion opposed migration. conclusion. This is one way to
77:18 resolution. Okay, if you know . So me one point told me
77:24 same upper left because oh, me before this is after. Okay,
77:32 always more detail you can see. right. So those are motivation.
77:41 people do this. Okay. By way, uh, while the most
77:46 methods of doing this it's called a big car, which is really,
77:51 is the reason why I send you first reading exercise. So from today
77:57 Friday next week, you can read paper on and answer those five
78:03 It's all tied together Okay, so gonna talk about those, but let
78:08 look in the metrics again. So metrics. Okay. Remember how we
78:14 our exercise? A moment ago, did my message way. I put
78:19 web late as a column vector shifting space. I have a reflectivity here
78:26 my conclusion. So you can accomplishing this way. Right? So
78:34 multiply the tool. This way you also make the other one. You
78:38 , inside metrics. Three other directors back off the record outside. So
78:43 is this is the manager. It's to see the convolution operation. Allow
78:49 elements in the metrics. The vertical are the same way late. I
78:55 stagger them, shifting one step of . So the square metrics in the
79:01 . All right. Um this is consistent. Everything. No issue,
79:07 ? Everybody happy about that? What call this is stationary. Why?
79:11 mean, what you statuary stationary in symptomology means the energy doesn't change with
79:21 . So if you say we go because I'm all right side more or
79:25 the maximum attitude the moral is the with time. There's not much increase
79:30 It's about the same level. That's the words stationary. Okay,
79:37 works. That's nationally. Um, we care about that? Because there
79:42 something of a non stationary. What more non stationary? Well, here
79:48 no stationery. You see now your decay with time, right? Like
79:54 GC. You talk about gain You know the decay because the geometrical
80:01 . Now keep in mind our Our model says go back to our
80:05 invention. Ramos says the future the Earth relativity, give us
80:12 The earth did not change. The did not change. So when we
80:17 the phenomena off non stationary with our , something wrong with our future,
80:26 same earth this edge part of the be didn't change the previously with these
80:31 earth. Now the same earth we're everything Changing the data is because the
80:39 they feel they have been changing. the one we didn't exercise is same
80:43 across now, the future being changing as you go with the further you
80:50 , color members. So Okay, is kind of what we think.
80:55 is our interpretation or assumptions about what's on right. And then my
81:02 Yeah, because subsurface during our survey change. It's just our feuding processes
81:09 , which is why we're thinking this . So what we do agency or
81:13 we do way basically change the The future, Although mathematically the to
81:20 the two time cereals are in Poland , they're different. Well, that's
81:24 we want. The other one is future we want to remove. We
81:29 to get in the reflectivity, which the earth structure from based on our
81:34 were using the tool toe assign one the other, right? So I'm
81:41 to show some teacher mathematically. But I do that if you I want
81:46 , review something this morning I should I didn't I was lazy, So
81:52 need to I do I said this to myself. You can run,
81:58 can hide. So I thought that go through that you wanted to go
82:02 , So I'm going back. I'm material. Well, this morning there
82:06 slide. I didn't go through. this slide here. Remember stopping
82:13 Okay. To point to this morning talked about this one. Talk about
82:17 one I talk about this one? , that's given to slide Escape that
82:22 slides. No, no, I this slide. I skip this
82:26 No, I have this one. what we escaped. Here's why we
82:31 . So this to slide. I a little bit. Okay? They're
82:36 detail The reason I'm trying to minimize detail. But still, I cannot
82:41 some detail in orderto give you a for explanation about the convolution. This
82:48 the reason now. I don't like . I do it only the why
82:52 had to do it. Okay? like I don't know. We're lucky
82:57 us. The January if your you know, over April, Guess
83:03 ? Text time. Nobody want to . Text, right. Alaskan give
83:07 money back. But it's on cutting We don't like is that detailed mathematical
83:12 . So I'm going to go through little bit. I'm not going to
83:14 to all the details. Here's a bit. Why? I have a
83:19 future like the one we're using time . Any of those future can be
83:26 . We call the irrational forms love point of view. So the rationale
83:34 . Okay? original form just means ratio is that these transformed something on
83:40 . This has something about him. what expression gonna be Okay. It's
83:45 because there's little complicated. Especially this , downstairs one. Then you don't
83:51 downstairs. That's that's the best Okay, that can happen downstairs.
83:58 we call this now. Reclusive now . That means there's no recursive just
84:05 just simply a time serious. uh, that that often the
84:11 we'd like to see that. So the only thing to say here.
84:15 example Here we escape. Don't worry that. So we just like
84:20 So now same expression here, remember Monster. So, Ted har reflectivity
84:27 earth go through pewter. The future upstair downstairs and stands for numerator He
84:37 for din operator the ratio of numerator top youngster about So you basically the
84:45 by shooter by this guy in the give you output right? So that's
84:51 same expression I'm just rewriting them using and d we also wanted Toby Mineral
84:59 dumpster as the minimum face. This the reason why we discussed minimum
85:03 This is also a reason that the require minimal face. I talked about
85:09 , but I didn't go deep. know why? Here is the reason
85:12 it is not minimum face. You be stable. Okay? What?
85:17 mean by a long, long be . What? Mr Bean's human
85:23 What did the convergence? Okay, means your result ability will be
85:27 very long, too long for your . Okay, there is no I'm
85:31 scared to hold that. So it's , right? So let me again
85:39 the whole thing rather than put this here. I write it out.
85:42 give you example here. I'm saying lobster has three terms here. Just
85:48 your exercise. Three volumes. This This is called Aziz transforms. He
85:53 a symbol for the location. They the sequence. Okay. The really
85:59 thing here is about the coefficient. in one and two. Those are
86:02 confession upstairs on D one d. other coefficients are downstairs. You know
86:09 you are here the first time downstairs why? Well, the reason
86:15 afternoon. I think you guys are a little bit positive. Okay,
86:20 one here because doesn't matter if there's one here which is normally divided everybody
86:24 upstate and surveyed about it. You have. I want to hear every
86:29 by two. That's still so for dance their first time. Always want
86:34 normalize that. Okay, divide everybody that term. Now you say that
86:44 be zero. Yeah. Zero can zero. Also, this is superhero
86:50 . So anything you have this we don't divide. Okay, Keep
86:56 . So because mathematically this irrational then we can just like we do
87:04 . We can multiply both sides Equation advanced there. So we have the
87:10 one plus C one c plus into square. We multiply that downstairs thing
87:15 need a collaborator on both sides of equation. I want to do that
87:21 the left side. We have See that? Let me change the
87:30 . Maybe you can see my point clearly. This is a good thing
87:36 online. We all on the same we all force ourselves into. I
87:42 using different color this time. all of different cardinals. Uh,
87:48 battle blue some color. Just a to light. Okay. I can't
87:57 see it all right. So why down stair on both sides? That
88:05 this whole thing gonna turn on left . Multiply with why, right?
88:11 result. Gonna be one times why wise time serious. Okay, Why
88:18 ? So I'm gonna focus on Whitey . Why? It is about criminal
88:23 . Serious, for example, the y Z because toe is this transform
88:31 be y t which is accomplishment power toe power t plus Why he plus
88:44 the p plus why our continue on before that one will be the he
88:58 of the while on the why he is one See my point and they're
89:06 coming from So the folks here on term just I'm just showing one what
89:13 on one term. So this is simple here. Why txt off
89:20 However, here's Acti. They're whitey means conditions off the wide serious Y
89:29 So Y Z, this is why he here? Is that why?
89:32 he okay? This is a That's why I'm saying mass require your
89:40 . Just beware patient with it. when you multiply this, what do
89:44 see when you have one, multiply , multiply with with whitey. You
89:51 this term, that's what That's Why is right like he next term
89:56 T one was e d WASI comes Robert Ghazi here the means making your
90:05 one other lower. Why? Because got X Y z. So it
90:10 this other term they want Why you give you power of tea? Because
90:16 got already got to see here. why this was a D. Y
90:22 the the human. Because you see is your location change? Because
90:27 gotta keep collecting term or saying our power the two terms the one is
90:34 plus T minus one Give you tea here the next term to on white
90:41 to give you a total power of . This is this The result off
90:48 the dance Their future with the wide Give you this left side Okay.
90:56 upstair the numerator the sweeter multiply with same same fashion besides the same
91:04 Okay, zero times Y t x I'm sorry. X t total zero
91:10 t equals t and one s t . One total is also one.
91:17 sorry. Total authority. The last and two times X T minus two
91:23 . Also, T this is a off upstairs in the filter multiple.
91:28 was X'd. So that's why we're at this moment off power the power
91:36 those who started equation holes. I'm just purely to mathematics here.
91:45 , this is just algebra accounting. is accounting and a boring stuff.
91:51 . Just prove that everything is It's kind of a logical proof on
91:56 proof. All right, now let's what we talk about it. X
92:03 x is the reflectivity on this numerator this stuff here inside of the
92:08 Okay, Why are they observed on ? The problem is, now you're
92:14 with the question. Why t and things here there's three times. There's
92:18 terms in what he means. That time moment. Remember, those are
92:22 grams. Okay, So what Like wiggles, right? We go
92:28 down. So there's one term. y t minus one. There another
92:31 whitey next time T plus one. ? So Whitey is term.
92:37 current time. While two minutes one a moment. Last step. While
92:44 tools to master So the sweet stop the in the in the wiggle wiggle
92:49 right. So that at the same , So I want to see it
92:53 right now. Okay, that had move them. I simply keep
92:58 Here are left side. Um would toe term there. I moved over
93:05 the other side of invasion and move to the back. Actually, they're
93:10 now, right? So they go now, right? You can see
93:20 . Right side and zero and one two. Then I move the left
93:25 too. There. Okay. What say is, uh this current moment
93:31 , Christina? Data space Your value not only the future of input.
93:38 X's reflectivity. It is the upstair the future. You also have danced
93:44 future control with the with the previews put Why he Marcella is all for
93:52 moment ago. Why two minutes to step ago? Those air produce output
93:59 is reason why can't we call this able feedback Okay, those are
94:09 If anybody means you all put become again. So this is complication off
94:15 filtering of time. Serious because time Not just one value is many,
94:19 value. So they produce output from future. Gonna join again as the
94:25 step Input. That's why it's called recursive sheltering. So whenever you have
94:32 , which is stance there no one can have this process unless you don't
94:36 to see one people term. Then just like going upstairs and no recursive
94:43 , that's well, that's my review . Why talk about the review
94:47 I think in a moment to talk the con. You can't require your
94:51 here. I'm trying to minimizing the . I'm just trying to go or
94:56 ? I have to go over. thought I can skip this morning.
95:00 can, so I'm catch back. . I'm glad you have about I
95:05 your attention span here at least any ? So this is basically recursive description
95:13 that page. That's the me. you mentioned when I want to give
95:17 on, then we're ready to talk the EECOM. What's deacons about?
95:21 , so now I was there about stationary, non stationary filtering. Now
95:28 can talk about dekom. Okay, appoint Let's appointer. So what we're
95:34 here is that I have white Um, this is this is one
95:39 here. There's some multiple math when have a wide noise. I thought
95:44 was right here. No, I have that slide. It's okay.
95:49 have white noise? Their auto correlation white noise will be like that
95:55 This is one definition for all Why? Noise? White noise means
96:00 no relation between one part of next . There is random. So I
96:04 to coalition. When they overlap, gotto. When there's no overlap,
96:08 is no money. Their attention zero one non 00 No, no,
96:16 times. Three times zero location. nobody know applicants. Okay, explanation
96:24 expectation operator with cross correlation. So here's all the correlation across
96:31 Okay, This is one way to that the same thing. I'm sorry
96:34 the same question twice. Won by , one by equation here in that
96:40 . This is basically off correlation. XP, I'm saying, actually,
96:45 reflectivity is random. You'll get this function. Okay. Now, this
96:51 this is a background for the for protectee conclusion. There's some heavy theory
96:58 involved about some scary names. But worry, we don't have to deal
97:02 them. The two guys found this long time ago now, Kocin.
97:08 . King King may not Kitchen Um, okay. These two
97:13 Iran is a long time ago. first applications biggest, most famous one
97:19 the application for the radar. Us ableto win the World War two largely
97:25 large, because we as technical of , especially radar technology. A lot
97:31 theories. Redox technology is divided by guys. Conclusion conclusion. The detection
97:38 . Basically, Andi, it says following. It says that Oh,
97:45 spectra equals through the Fourier transform auto function. It is. It's we're
97:52 , but it's let me. I've using pain again. I'm sorry.
97:57 need to draw something there. So scared of a color now.
98:04 Color. Maybe Right again. A rate. Maybe I don't wanna be
98:08 invasive. So this is the same spectra because to the four a transform
98:15 the code cart function Okay. Looks short about how to understand what I
98:20 . Okay. Basically says that if give the other correlation, then the
98:25 more for transform you get a Okay, The two steps to get
98:30 spectrum. So diamonds I can take previous result. If relativity function like
98:40 , I can just just get a . Okay, uh, by the
98:46 . I mean, this is all car. This is auto all the
98:53 , and then I do the special the Florence. But by last,
98:58 diffusion last page, I should get terrible function because the white relatives,
99:03 wife. Okay, so it's constant wide. Okay, I'm not sure
99:11 follow me because this is a mathematical . I'm not sure. Follow
99:14 Okay. Let me using I geological description about what's going on. What
99:21 saying is following we say, why a small ground you start since my
99:26 . Okay, so no ground okay? I'm just trying to assess
99:34 . We go. We're saying this has two components to part of contribution
99:40 factor here in the Bago. One is that the earth relativity earth structure
99:47 other part is available eight. we know that already. But the
99:52 deal we say here is, actually, what we're saying is,
99:56 , the Earth Relative party is very random compared with the regulated.
100:04 you agree with that? Whatever. the earth structure is always running
100:08 If you look in the relativity I show you I'm just get back
100:12 . This is a perfect example, ? We're saying this earth relativity is
100:18 random. That means you have no about where gonna be higher.
100:22 We're gonna be small. Big is one part of off the TV on
100:28 parties. Cross correlation. There's no . However, everybody was covered with
100:35 waves laid right. This part of glamorous bakery late is late. Every
100:39 of the same way late. So , there's something about winning.
100:45 I mean, that means if you a spectra, the spiritual of the
100:50 Here they are, is the same special off, the off the
100:56 Nothing to do with the earth of . Okay, what I'm saying,
101:02 we're, you know, confusing very . Let me show you a
101:06 This is why I'm show you this . Okay? This is a picture
101:11 got relative here, right? Pretty . This time, though,
101:15 here's special Ever reflected. Right side a special I am special on the
101:22 I got away blade. This is a sign way. I just drove
101:25 women. I got a special sideway the right side. Now I come
101:33 the fertility with regulate I come All them I derive the conclusion.
101:44 Now looking the special conclusion as spectral the regulated. Did you see
101:50 They're the same. Almost then. , this picture is not very
101:58 so I'm going to back out. going back. Hold I This time
102:06 keep I'll keep the junk because I another picture here. Later on,
102:10 can see you show you better The picture. I don't have a
102:14 Not is there for a bit Now let's go backwards. I'm not
102:20 about that one I'm talking about. is my reflectivity. Okay, on
102:33 was my favorite. When I come with with reflectivity and all this this
102:39 notice my special off conclusion Samos my . This is a two confused,
102:47 I don't I don't want to jump . That's why modified to the last
102:52 . This is just color working. , Same thing. So I go
102:55 to my saved notation Point of So this is my seat notation part
103:03 you. Let me get my pain again. I better use a different
103:07 this time I'm came back forth, back. So this is the foundation
103:13 our process. Let me let me the problem one more time.
103:17 so we're better present star. So all problems. We have a
103:26 Assad's here. We see. Uh, we want to go from
103:34 to here because we know these two can evolve. Give me the
103:39 Our trouble is we know only one of three. Okay, a come
103:46 with the edge. Give us the . We only know T. We
103:52 to hurt. Determine both the Okay, we got a big goal
103:58 . Too ambitious. This is almost you are someone. What's the result
104:06 ? Two factors multiply. Give you seven once those numbers. Well,
104:16 you ask me the result of multiplies . What's the numbers? I don't
104:21 . I mean, any number you , right? Number seven is the
104:29 . One time. Seven. Two times 3.5 or whatever. Any
104:36 , Any number you. You give the multiplied. The sevens?
104:38 Unique. So we want toe giving resolution. But we only have data
104:45 don't have with Blake. This is force us to the assumption, our
104:52 Years, We say the earth is , okay, versus random and opportunity
105:02 be customized special data because of because we notice that the special off
105:08 data in this case the lower one the same as a special off the
105:17 regulate. Then we can take a , assume the data we've laid.
105:23 mean that a special is the village , then from there to go back
105:27 the village. That's our approach. why it's called a spectral factories,
105:36 meaning starting from spectra, to figure what is willing to give that
105:46 The challenge is that there are many eight different face given the same
105:52 because this special power special we didn't the we didn't fix the the first
105:58 . We do not feel special, is why we are so many in
106:02 face. They were saying, the wise a minimal physical one. Although
106:08 you go through a special tribunal you have multiple choices so unique.
106:14 , you can choose the one which minimal face the only one little
106:19 that company, the story. So terms of operation will take a few
106:26 are they come through the, special. Now we say this specialize
106:32 special of the privilege. And from spectra, we figure out which ones
106:38 face living on that we say. , that guy is the true
106:47 Okay, so let me using the that out here maybe can see better
106:52 many of us jobs always keep that mind. Again, our problems will
106:57 on data We tried to turn me relativity, Relativity. That's impossible because
107:04 , unique, you cannot give me result of medication and asked me for
107:10 are the two multiplied. Give you answer. That's just crazy,
107:12 Any number can give you. But said, Yeah, I can do
107:16 . Why can't do it? Because is true. Relativity is weather,
107:19 is most time true. Then the off ability off the future the
107:27 A special of the data without the . So from data, we can
107:32 a special from spectrum we can choosing which is remove face the way
107:40 Now I get it. Okay. not sure you guys follow me,
107:45 I hope you do. But if have any questions, speak out.
107:51 ? My fear is the level. super from you, you know,
107:55 me anymore that we're in trouble. question. Okay, there's some detail
108:05 . I will just describe Monica maths the detail here is that the public
108:10 . Some details show you what that . Our Here's all the tolerance,
108:14 is same. A special. The serum says all the car is
108:18 fully transformed. Give a special so don't even go for Children. Just
108:22 to the correlation. We got that here, we ought to call it
108:26 we former metrics, we solve the , and that solution is ready the
108:32 we're looking for. Okay, this is my medical part of I
108:37 believe the, uh, jumping around little bit, not jumping around,
108:42 escaping the detail here, but idea describe to you. We start on
108:49 that we solve the Theis mattress way solving for someone will answer on then
108:56 can do other answers. Okay? feel if too much deeper level math
109:02 , maybe more confusing the health. I I'm not that much detail,
109:08 you can look it here through my . There are other books on these
109:12 . You know, I I was the literature and interpreted this way.
109:18 my example. I'm just give you example of a few. Get out
109:22 marine case, This one taking from by monkey. Uh, air gun
109:28 . So what? Probable air You have this ghost. Okay,
109:33 refreshing. Here, Spike, this . What happened to those is your
109:39 level here. Now, you aired fire here were gone. So you
109:46 going down. But they're also wave going up, hitting the seawater.
109:52 going down. So you have to because of water is like perfect
110:00 So the signal one slightly delayed, the same energy as the primary.
110:06 this double the source give you to the primary on the coast.
110:13 I p and G's goes. So this case, this guy tried toe
110:20 goes using this, uh, idea the conclusion idea. Okay. Using
110:28 square feeding. Okay. And once apply that, they give you this
110:36 that this is a kind of the off de conclusion, you can
110:43 or somebody and somebody called people's same . The Eagles off he goes
110:54 he goes, Changed operation. This surprised. Coast. You see,
110:59 ghost of surprise here before was Not much reduced, but like many
111:05 your physics, you gain something, lost something. The losses there's long
111:11 here before was very clean. Now got this long tail. This is
111:17 result of this square inversion. The energy in this one result this one
111:25 they're they're very similar by this. you kind of smear the ghost
111:32 Okay, but people feel this is because you kill that. The double
111:40 second peak is ready to go. Much of that is gone now.
111:44 is good. Okay, additional It means de ghosting, right?
111:53 is the ghosting by become, this is also same approach behind this
111:58 showing before about the change of the between zero face on the minimum face
112:04 this coming maximum rate. I applaud out of Iran was very strange.
112:08 is time zeros, all right? a problem. The same way.
112:12 mean, I don't know why. , because of my printing rotation,
112:16 it looks very strange. Okay, this is what I showed you
112:20 Basically same thing I call a match future. But it's also your
112:24 Same time. So traditionally come from same idea you can go this
112:31 Uh, Susan Model. I can this. This is where my patient
112:35 this about a white noise white noise result coalition will be like that about
112:42 . Okay. Any question? I this is very dry material, so
112:47 not sure you guys follow me, I'll give you post time. Anyone
112:53 any questions? Right around too I'm worried. Anyone has issues.
113:09 ? I'm not sure you guys with . That's why that's why I
113:13 This is challenging for online class. don't see official expression. I don't
113:17 you don't see me clearly. So I'm really scared about I'm running all
113:23 you. You guys say okay. . Talk about. Keep going,
113:28 , going. It is very Yes, it is very hard electoral
113:34 , So I'm going to switch something little bit more interesting. Maybe giving
113:38 motivation. I'm gonna say something more . Okay, let me give you
113:42 very busy picture. Very pretty but its informal. This is a
113:50 , big, heavy book. All humans is a big,
113:55 uh, assessment Processing. Very, heavy issue volume. I told him
114:00 heavy New Orleans behavior. First time saw the new book Compound off when
114:08 killed my you killed my rest. going on? As I was reading
114:12 book like this, I got Okay. Can I follow suit for
114:19 ? Okay, so feel smalling, ? It's very big book. A
114:24 of examples. Uh, some techniques for him. I'm using this for
114:29 purposes. Though Different parties described in is a book. So I'm eligible
114:36 you the point of war, The on the convolution. You can see
114:41 process. You also see the issues the limitations. Okay, so this
114:46 important. One. Uh, very one here, I better I better
114:53 it. Get it back to my mode. I don't know what's the
114:56 way. Maybe maybe less. Our , you have to draw our
115:01 So this is the productivity of the earth. Because it's the impulse response
115:07 call this earth Relativity basically is a rotate 90 degrees vertical right now.
115:13 it's a really earth relativity. you know, broadband, high
115:20 Now we don't. We want to that A But we have. We
115:24 no other way other than using So using a very broad sweep,
115:29 ? I said this morning, the can be very long. So total
115:34 I cannot see here very well, I think the lands off the
115:39 maybe only, like, two seconds something. Maybe three seconds, but
115:43 becoming, I don't know, 10 beyond low frequency and I go to
115:48 frequency, You know, I'm wondering . Well, physically, you have
115:53 be this way. This is almost you drive a fancy car.
116:00 Do you know how much power House offensive? Half What? The maximum
116:08 . Like a wiper. You 600 horse power. How about space
116:16 engine? You know, rockets? , even you have the rockets.
116:21 cannot go from zero toe highway speed 00.1 seconds. That's too short.
116:26 see the time? Same thing with rial. Physical viable size You can
116:31 go slow. Low frequency gradually cleaning . Okay, so this is a
116:37 long time toe 100 years to be hurts me The cycle per seconds.
116:44 means every second there £10.10 10 shaking is too fast So you you have
116:51 sweet here from 10 to $100. , if you come off websites with
117:00 the fertility Here's answer So this is from websites Very, very long Because
117:09 what your Your signal. Your ladies long Remember I said the length of
117:15 is the lens off the two together as one right so very long What
117:23 do the first is we do uh we can change privilege We say
117:31 might not be middle face because the has a function Not this website You
117:37 elect electronic signals Viber size You're heimer not I electrical heimer? I mean
117:45 can't electric heimer but your signal from heimer has his own function Because you
117:50 ah heimer physical dimension. You hit ground. So you know one more
117:57 , which is the minimal face south . We're saying you're heimer, has
118:03 viruses. Heimer has his own which is, you know,
118:08 So you're really you're really observation data here which has three components. The
118:16 of earth First one, the sweep two the third wide new face Software
118:24 The three of them come all together you the data observed. This is
118:30 for modern people. Data makes sense you're too signals is a spike that
118:41 already have your face off South country signal is most likely to signal sweep
118:47 sweep. Come always the human face This is really can always the earth
118:52 you the data e I didn't get minimum phase. So sweet report isn't
119:01 mean isn't the sweep is our source In this case, sweep is the
119:06 the signal. So everybody wanted sending the Bible size. But virus has
119:12 a signal of hit the ground. your hand moves like this a
119:18 But the signal from the ground to anywhere else is not that anymore.
119:23 something and on top of it. , so it's It's like the but
119:30 not that. But it's not the response. And it's not The Jew
119:33 response is no, no, it's soldiers. That's why I saw the
119:37 . It okay, for example, my source is a smaller or bigger
119:48 electronic signals still the same way. my earth Southwell is different.
119:56 This is actually the big issue because on the biggest problem between landed on
120:02 data is the land data even use same source either websites or are
120:10 Because your local geologic different your your source is different. Yeah,
120:19 way to look. This is the about how big will it is.
120:24 surface. We typically using the velocity water as the average velocity. A
120:31 of water is, uh, 1500 per second. That's the P wave
120:35 water. Okay, which is the , but it can be two
120:39 I say 15 100 m per the near surface velocity. If you
120:46 this case, 10 to 1 20 60. If 60 hertz, you
120:53 the 150 1500 m by 60 that you something like, you know,
121:02 certain meter. 25 m. 25 . Like 100 ft. Almost very
121:09 area. You don't have that kind size of time at all.
121:14 that means you're you're really sauce. the rock peace be with your
121:23 It's like you're kidding. A Bill bill. You're hitting with heimer.
121:29 the bill, that location on the is the rial source. If you're
121:35 a stronger part a weaker parts off part, you're some different. That's
121:41 we we played piano, right? have same finger about piano. Sounds
121:50 . So you have that power because talked about the total thing. This
121:54 physical. You gotta come in So you're really detail here. That's
121:58 I give you the example this morning collusion. Many, many things.
122:03 here are minimizing to those three The next one is magic is the
122:08 for the emphasize here that are very . What you do is you do
122:14 thing. This is a sensible cross . You cross college. What?
122:20 hole? Later. This is where not using my panel and pan
122:27 Of course. Oh, Where's my color? I didn't want you guys
122:34 eyes. Okay? The city is okay. I'm can all be.
122:41 here with E is here. So guy, these two guy cross correlate
122:51 ? I'm using this line for cross When my across color sweep with data
122:59 suddenly go from here e put Is that amazing? As a
123:11 we were just one step You go Mito F What happening? Well,
123:19 Colijn off Be with an e beings There's the old to correlations I'll be
123:35 right Access Because he complains be he three things come altogether a be de
123:48 b d Come out together give you So when I come all the auto
123:52 across currently with with B again That inside there's auto correlation or B and
124:03 makes sense on dso This is what was talking about The reason for sleep
124:10 Go back home I slides about sweet , I don't have here this
124:22 right. I talked about Claudia. the chapel too. China to I
124:28 the have the wish I can It's okay in chapter two this morning
124:35 show you the all two Korans uh, sweep, which is a
124:44 ? A Really? I told it's a very, very small,
124:47 small were tiny. That's a good . Okay, so that's why you
124:53 . That's why the Web sites and . Big discovery, loss,
124:59 biggest revolution in reflection industry. so now this is color. So
125:05 we're saying is, uh, dog there here's collar clutter is the other
125:14 of sweet. So that's why it toe. He hopes that goes
125:21 Can always be piece that big sweep you a scholar. Okay, but
125:31 is zero face. You can also at zero uto minimum phase, which
125:36 the edge as a minimum face you covered by I imagine Future are ratio
125:44 future. Match them. Okay, me go. Let me stop
125:51 Say something here. That profound looking at versus original A Basically we're
125:59 f is a lower solution version off retroactivity. Okay, if you look
126:10 the special, I will be like to 2100 years. It's very similar
126:15 the relatively. Already is our big step just by on cooperation across
126:23 were down. Big step The rest the steps to enhance resolution through decon
126:28 evolution. The rest of them just to do something calm. Okay.
126:34 , so, for example, we because it's 91. The Claudia is
126:38 face, so zero because symmetric zero . If you want to make this
126:44 tow, minimum face, you simply with this future here because that filters
126:51 fielder matching minimum faced with the color . So that's why you come off
127:00 tool. I and F Okay, is the result here in the
127:06 And I see. So this to , this guy, this guy Kamal
127:11 you here. What we say is F is the zero face version
127:18 The data on gear is the minimal version on the data because we can
127:27 the zero faced aluminum face. Are still with me? And then we
127:36 do to the car one for The spectacle off F, which is
127:43 minimum zero face version. A spectacle J, which is the minimum face
127:53 . I'm just compared for you on last one is the to answer,
127:58 is the initial a futures row. to $100 dampers filtering. So this
128:06 , uh this is the true Onda Key and I k and l
128:12 answers. You gotta buy com. . I'm sorry. For the small
128:18 . You can look at the original . I don't have electronic copy,
128:22 I scan it. I talkto He doesn't have the original either.
128:30 , but you can see the the . Here you get the most energy
128:34 in location. However many, many party are wrong, by the
128:44 Okay, Because many point of Wrong. I know you. I
128:48 see them. But I see many of large Iran between the color again
128:54 of different issue here involved. Keep color. Okay, like 200
128:59 Last one here is probably one but its negative here. This one
129:04 not in our location, even the locations. Wrong. Too much from
129:11 , Andi. So you can see also plus minus bonus so their priority
129:18 be wrong. These are issue usually something something to do with the
129:23 . So even in theory, their in priority you do again resolution.
129:30 there's always problem. Talk about How much? Matching with reality.
129:37 , I might give you too much now, but I'm always do
129:40 I guess too much. So we We can move further, give you
129:46 example here to relax a little This is a very terrible the asset
129:50 nothing there. I call this, , that one. I call this
129:57 TV screen. Okay. Well, not signaling the TV screen. That's
130:03 you see, but amazing. If do, I President, you could
130:09 this signal. It's still not I can see stuff. Okay,
130:15 is a data flow offshore. California aquatics citing surveyed by us years long
130:22 assets. But we can study and structure. Okay. You don't see
130:26 . There is a lot of echo what opposed to strong? Okay,
130:34 are other methods. This is the popular one. I just finished.
130:38 prolific are I think you have a reading reading by Schlumberger and Houston.
130:44 a good paper to read. I'm elaborate. That paper later on,
130:47 time. So the closest this section , I'm gonna show you another easy
130:54 . What happened if you know one them? I said, Earth relativity
130:58 through computer, give you the but sometimes you're lucky. You
131:02 lately, you know, we were divide them. So, you
131:07 data Just no privilege if you know . Okay? Just remove activation so
131:13 can divide them by division. Don't about detailing math. Okay? You
131:17 divide because you know them mathematically. can. This kind of frequency
131:21 Um, that means you're not with you can you can You can minimize
131:27 . You can divide them. The is that you don't want divide by
131:31 is you're gonna define something called at level, so make sure you don't
131:35 to go below something level. So you have a water level so you
131:40 want divided zero downstairs. 10 So can. This constant here are to
131:46 division, right? Okay. That's on. You can do by.
131:53 some other special issue here involved. don't need toe go to the
131:57 but I will show you my This example off This kind of become
132:02 in by the reserve functions. If you're working on crustal seismology,
132:09 are very popular doing the show It's basically the evolution as the solution
132:15 simple. One dividing the related by signal. So they're just using the
132:22 off a letter face letter for, , you know, convert debate s
132:30 s well, blue color. So . Incoming letter P didn't come
132:35 So the cover to cover the wave by that then toe determine where the
132:44 coming more whole, out of Okay, so I'm not going to
132:49 here to tell you that. Isis conclusion. Metals in detail, That's
132:55 . Okay, I'm done with this here. I'm not sure I Everything
133:03 is a little bit mathematical. Very . I'm not sure how much you
133:09 , so let's not a few minutes . You can you want you want
133:12 guys want toe elaborate or anything? . And there are people very,
133:22 much like to get this thing done increased resolution little bit, very expensive
133:28 opposition. You have a mall receiver, mall, you know,
133:34 in putting the opposition. But there's because because attenuation it's very hard to
133:39 out resolution. So because probably want only way now for all these
133:47 but still has old issues. a D example show you here like
133:54 first motivation example. This was a examples. Sometimes they're not this
134:00 But even this example, as you from theory, we cannot trust the
134:05 they can be from. Polarity cannot something like HBO, and it's all
134:11 . Maybe party reversal are not in same location. It's always issue about
134:16 it's true or not true. This ? Was it spiking deacon or a
134:21 ? Isa predicted economy. So it's . Yeah, you can call the
134:29 become, but the technology is Uh, most times people say
134:34 maybe they just go to the improved . But when This way. But
134:39 a Catholic. How that means that have some solved constraints. They don't
134:45 need to be just a spike. can be basically anything repeating themselves.
134:50 want a real one way, this of revolution, basically to remove
134:57 we are repeating themselves over and over over. I don't think they're putting
135:01 You want to take that way, is another way to say the
135:06 which is the reflectivity is random, contrast to the regulator. Okay.
135:17 , can you Sorry. I was was going to to shift something
135:23 Maybe you go ahead better. I'm gonna imagine that, uh, I
135:31 to myself, what should I show ? The paper by Should murder?
135:35 . Now, if you look that which you can download from the blackboard
135:40 in one, you can see their off the multiple they're using. This
135:47 pretty become to surprise multiple now, is a lot of the biggest challenges
135:53 the industry because lots of the really multiple problem before they can be.
136:00 , predicted calm, but which is very effective only for zero set,
136:06 that I could get more and more , you know, pre stack with
136:09 fall upset, you know, it work. So there are other methods
136:14 have nowadays, like s army and T mobile methods. But there's always
136:20 . They're always issue there. So will hear this If you're looking
136:25 even your geologist, go ahead. you haven't thought right or anybody
136:31 Yeah, I was just going to so. So when we do the
136:34 we We input to the process. input the circulation. The spectrum of
136:39 creation of the articulation. And the here is that both are white.
136:46 a random The noise and reflectivity Yeah. Okay, so then they
136:54 all what we get from the as you said, is the is
136:58 spectrum of the wheel. That's That's right. It also means in
137:04 on the wrong. On the better , the wrong side means the random
137:09 contains both relativity and the noise. , so that's a that's a
137:19 So even though the result is you still have noise, for sure you
137:25 . It also contains some, color noise you can remove anyway,
137:30 you don't care. So actually called can be removed. Recognize will not
137:34 removed. Yeah. We usually tend do a lot of my situation to
137:43 the colored noise before the dick. you're ready in a racist, this
137:48 mold. To some degree, the is easier to be removed and ready
137:55 . Green allies is very hard to . Yeah, Yeah, that's the
138:00 you cannot predicting how they go. , he's only assumption to some
138:07 but in reality they're always summary between don't follow your assumption privately, so
138:12 a lot of real world problems. that's that's opportunity for you to see
138:17 . I know you want people Follow everything I said today. But
138:21 not trying to give you Summarize the off all over the finish before you
138:25 next week. I think my first actually here. Just give you a
138:29 about the, uh, modeling, is the back background seismology. Plus
138:35 new version filtering off com, which echoed to what I said this
138:41 You said small ground you see in data conclusion of manufacturers, if you
138:49 industrial patrolling your finishes off your you want you care about sub services
138:57 on the property. So everything because noise so you want to remove them
139:03 while the method of using it's teak , uh, inverse filtering.
139:08 But when someone doing, for monitoring subsurface magma monitoring earthquakes, that
139:19 , I think the structures noise, want to focus on the source off
139:23 earthquake, even in the industry later , if you have a better
139:29 we can monitoring micro seismic. Beautiful . Then you can go beyond the
139:37 . You might be able to matting radiation pattern off medical cracks.
139:43 you know, fracturing since that that . But right now, we're not
139:48 we're not that far because our technology our instruments are not that good.
139:53 you guys are young, You I think I believe in the
139:57 Gonna be up or down, so always gonna be that way. I
140:00 they're gonna be a bright future in petroleum industry, regardless because gonna be
140:05 for for fossil field, no matter , but at the same technique can
140:11 applied toe apply to other field. was recently about toe organizing a special
140:19 on imaging using reverse time migration by O. Uh, I'm trying to
140:26 Tell them that that's the image ing processing methods useful for both petroleum.
140:36 swell, a swell as green even some other things. Like a
140:42 . I think human being has a need formatting sub service, whatever
140:49 using technology or geophysical methods making So technically, multiple high resolution,
140:58 physiology, always in demand, always demand Okay, So I think for
141:06 guys to learn this Your cause, goal right now is going toe to
141:10 in the industry. But if you , follow the intuition on the physics
141:16 argument you might be all find usage the principles. Even if you tell
141:21 field to something else. Maybe, know, civic engineer Lee, who
141:28 detective, natural disasters and a lot those issues involved. Industry, of
141:38 , always want They're actually very few in the industry now. Now that
141:42 attacks about the conclusion you always need . Okay, it's time for one
141:50 break. Right? How long you want? 3, 15, 12
142:02 ? Thank you, Maria. So continue now. Let's using lost
142:07 Get it down. I guarantee you good sleep tonight. I'm sure you'll
142:11 so tired. I guess in your you imagined that Keep mumbling and
142:19 Alright. Any any questions? But far, everything or comments or any
142:24 things? Interest about it. I'm considered drag on, so we
142:35 about modeling this inversion part. so now it's about 3.
142:44 I think we'll make sure we started like about half an hour from
142:49 or maybe 40 minutes. Then we almost on time. Okay, so
142:55 good thing is, we don't I have to go through a quick three
142:58 . I just heard today. So ? I started history. You guys
143:02 be on your own and take your on the guitar with it.
143:08 so this is the last part. there are always more detail, but
143:13 think there's some detail you might want know. For example, you heard
143:16 talk about the the squares. Maybe multiple, like STD generalizing, Virgie
143:26 even fancy stuff like bottles. Gilbert been there forever. But you can
143:30 for it s que are so maybe much stuff. So I'm just focused
143:36 the concept, even though this has a lot of questions, but I
143:41 wanna give you idea about concept want what they after What are their
143:47 ? Okay, very often. If are looking for oil company, for
143:54 , I had many students in pastor got a mass degree and find jobs
143:57 the oil company, for example. call them portfolio holders, so they
144:04 portfolios man exploration or production or So their job overseeing their job,
144:13 the operation Understand enough detail so they manage in this for the company we're
144:20 , jailer deciding where to drill are some aspect and so on. So
144:27 , so you want to know the principle? That's that's the first thing
144:29 want to know. So I'm going give some motivation about emerging.
144:35 so it's just wanted. Make your . I'm just give you this
144:41 Now, by now, you're very . This is how you do the
144:45 or filtering detected while the future into metrics applied to data. So
144:53 I'm sure you even though this morning by, you know, scratch some
144:57 thing. This is actually how we in computers, right? This is
145:04 we do. Because overall that our discrete data sets their time streams.
145:10 this is how we apply even a time ago. We do this
145:13 we still all right. So I'm review several issues, even dinner,
145:19 linear emerging means like this. The can express as a linear system.
145:27 in a leader, um, use powers one not, like, go
145:32 non in here, okay? It's PETA expansion in the first term is
145:38 linear relation between all now and the on a smaller parameter nonce data.
145:48 . Yeah. Okay. You can . Write this as a mathematical
145:54 You know, like the linear system . You can simplify in tow.
145:59 what? I call it the metrics if you call that. Okay,
146:05 a a is the metrics X is vector of am now on these
146:12 Okay, A a is metrics. like the last last example showing the
146:18 here. A Here's the Here's the . Okay, So depending on what
146:24 looking, so this is you can this is this is basic, leaner
146:28 system. You know, this may high school time. A middle school
146:34 , maybe. You smarter? Didn't Grandma Grandma school? I don't
146:37 . They're jealous here. You somebody did this grand kindergarten. Maybe
146:41 so simple, but anyway, the is in geophysics. It's not It's
146:48 a linear. It also very Huge problems in a seismic,
146:54 industry. Our honor can be million . All that can be, you
147:00 , 10 million points. That is huge problem, right? It's a
147:05 problem. So we know some detail . I'm not gonna community to show
147:10 principle. For example, the square the square. These carvings. You
147:16 using data here through the model through metrics, Colonel, to determine the
147:21 . Which is the ex, The parameter. Okay, Andi, the
147:27 is, uh, if you if you provide a solution, then
147:31 can use that solution toe forecasting what will see in the data, your
147:36 , your prediction or forecast result compared two data. Compare the difference.
147:43 Miss Fade that the misfit is minimized the least square sense. Police
147:52 like arm s basically so your position fade data, you know, in
147:58 square sense best feet, the minimal the miss faith. The square sense
148:06 that's the goal. Were driven. . It's the most common type of
148:10 for a linear system inversion. There many reasons why we like the
148:18 Uh, number one reason is that square assuming something called a normal
148:28 pdf, maybe Let me not using fancy words. Let me say this
148:32 layman's worst, where, assuming the function, Follow the Gaussian distribution.
148:38 that, the misfit. If you of miss faith, they are spread
148:44 a caution bill. Okay. Bill, The reasonable caution bill.
148:49 what, you have many random process with big ocean. Statistically, it
148:55 out when the Miss fatal error authenticity different type of situations than the measurement
149:02 era will be different. So this is the past for God you
149:10 That's the first reason. The second is, uh, it happened to
149:14 for, uh, if using the square feeding. There is a in
149:20 , analytical solution for your answer. , there's a new theory, and
149:28 have to answer for your answer. , other than now, another at
149:32 where one you can you don't have linear. You don't have a
149:36 Okay, Even give you right So if this is a four problem
149:40 solutions X in theory, your solution x at least correct. Because tool
149:50 first multiply both sides by, a transport. So a trance,
149:58 times the left side stayed off the size trackballs multiply with a on ex
150:06 rewriting the whole thing, but adding transport in front. But you multiply
150:13 , operation linear algebra, the left right side are differs because the dimension
150:20 . So I will multiply something from Everything from left. Alright, Everything
150:26 right both sides has been consistent. , so I go from here.
150:31 first question I get this, this one on. Then I can
150:36 further. I can go say I find in worse off this. This
150:40 was a multiplied inwards on both sides will give you this work. So
150:46 you on the right on top, gonna be is hunt balls Times a
150:56 the universe. You come on the . Okay. Assuming it exists,
151:02 is in worse on times that toe transpose on data. Okay, that's
151:14 visual solution. Unethical of, you , the score solution Now, how
151:20 ? Well, how come you know ? Well, you can see
151:22 Okay, We're saying they think the . So you find in version of
151:28 multiply on both sides. That's why go to that. They're in
151:33 This whole thing here go through the , right? Everything consistent. So
151:46 assuming this is a solution. This the theory can prove this is solution
151:51 minimizing the least square era. I tell the answer already and the rest
151:57 the equation to show you how I prove you. That's true. So
152:01 is the way it defined the we the misfit function or somebody call this
152:07 of function object function. I'll call the area function. Okay.
152:11 Functional air function. So air which is a dead, huh?
152:17 . Tradition is the man aside, . This is Victor transposed by the
152:23 thing. So which is, this is this is basic rms off
152:28 Miss Faith. This part is This part is prediction from the model
152:34 Models X here. Okay, that's Arab function. Your job to minimize
152:43 . Okay, by minimize the function need toe differentiate a 60 So why
152:49 minimize the air function? You what you for differential that with respect to
152:55 variable in this case variable is model inspector model. I sit at
153:02 That's how you minimize mathematically. So I do that. I spread
153:10 the metric from here. I spread in detail here. You know,
153:14 you want a detail that I'm gonna should with the model, which is
153:19 next pitch here. Okay, I'm I'm just showing that zero,
153:26 There's multiple here. But if you through that process, you see the
153:31 you organize everything you come with the . So here is my answer.
153:34 show you on top is the least solution. Okay, But I just
153:42 that cheating way in front and show that the answer. But you can
153:46 derive just by minimizing the city So this is the only one has
153:52 solution if you do some other you know what? This is also
153:56 l to the squares because the state a square power of two so called
154:01 two. But there are other For example, l y Okay.
154:07 phone number for, uh, our . So you can you can have
154:11 measures they want. They don't have analytical solution. This course only lies
154:18 another solution. Okay, Now, you have this issue around here,
154:26 have next issue, Which is how you find reverse of the magics?
154:32 an inverse of metrics. Never be , because magics not all of them
154:38 inward Table. What? I that means many, many of
154:43 especially. Why in geophysics they're not not inevitable. Mathematician called them.
154:49 post they there, you post. not immutable. Okay. I
154:55 there's just too poor to be Mathematicians in orderto we totally invulnerable to
155:01 with. It has to be a magics to begin with. By the
155:07 , while you multiply any metrics transports the original one. That's a square
155:16 . But I want to make sure guys follow me. I'm not sure
155:19 not rusty with math. You know a transposed me right? Transport transport
155:28 means to switch the role of the , switch the road column.
155:35 Raw column switch. Okay, I . Elements goto 80. I
155:46 the role column switch that the Argonaut move right. I think I make
155:52 because you and I think you guys tired now. Even something. Not
155:57 you don't want to ask. You wanna get over with it.
156:01 maybe not true. Okay, maybe know everything. This is awesome.
156:03 know but I don't know why, I'm not sure you guys follow
156:08 Okay, so, uh, we to find out what the universe so
156:14 Mathematical is not inevitable, but we invert it using something SPD Okay,
156:21 we d is a recommendation on Any measures is and by an committee
156:27 to three messages multiply. And the thing is that every model three are
156:33 . Therefore, result is invaluable. , I m is the number of
156:39 on the number of columns. So is the action. Actually it's
156:44 It's like Handler is not even Bye, honey Handycam holes this
156:50 the Middle World Cinema is the Aga like this? Now they are gonna
156:58 album. No inevitable. Anyway, other was really incredible because there's off
157:03 album zero so emerging. Just lower bottles. Okay? The other tool
157:15 square messages, please. Other square they call the injury you need
157:22 Your message is always incredible. Therefore there are good mattresses on their emerging
157:29 the transport. So you in worse toe your transports the slice b So
157:42 SPD is discovered in 19 seventies It very powerful way to show mathematically any
157:49 scanty from most on the one which you can define what have some zero
157:54 in the diagonal. So zeros if non zero, that's that's a totally
158:00 messages into your physics. Most time have zeros. So there, your
158:06 , we call those air No And you l hell no. No
158:12 . Okay, so there, Like this way. Then you can
158:18 example you can do in cross Gautama within version like that. Okay.
158:24 can also show example off the number the rial example. They're real example
158:29 publication by Catalina. Uh, this . Okay, this is volume of
158:34 signable articles. So the value in metrics in middle those s Y s
158:41 the singular values you can see example . A simple example. The single
158:46 a lot of here against number Not a single value numbers. So
158:50 of the non zero, but we're . Video practical use zero.
158:56 you can read this paper if you . You can see the detail,
158:58 I'm not going to that detail. give an example. So, you
159:01 , this is not not. You , something unreachable. So the universe
159:07 universe off the S single values notices about zero. Okay, if you
159:14 a zero, you cannot convert. U. S. Zero bc
159:18 So we're cheating them. We Okay, zero, don't touch
159:24 So this is cheating because you should our zero. We don't. This
159:28 called a general. Generalized. The we avoid zeros. Okay, zero
159:33 still zero, which is blocked This is a practical way. So
159:39 call it Sterilizing Virgin. Same same idea. This is my least
159:44 solution. Okay? On my This way, I'm just how this
159:50 thing here as the as the operator cartoon is emerging on. Then I
159:57 define generation version as the decomposition by . The last one is called the
160:03 signal about the combination activity. The best video already. Then go
160:07 a CD. We can define the off the operator g. Then we
160:14 showing SPD off there. Now we the words between relays. We keep
160:21 . We don't doing one hour zero . Don't do that because that's the
160:27 . We don't do that. We this. We don't know. This
160:30 . Go to some. Some level say stop beyond the artist came
160:36 Okay, now that's called terrorizing You know that you can t o
160:42 virgin, and then you can get solution, so they're probably They're probably
160:47 . What happened to those? Why in the death house space of model
160:53 corresponding to the zeros in single Because there's no space they cannot mapping
160:59 so much. Also coming for you zero on the U zero are the
161:07 in model space space that I blocked the zero singular values. Okay,
161:14 is a no small It doesn't Okay? They're They're blocked. So
161:20 no space, which means they're bridge . The other side is burned,
161:25 broken. They cannot go out of other side. What? I mean
161:31 world. Okay, let me A survey is in Houston. I
161:35 know where you got to leave. leave you sugar. That's why I
161:40 your your campus. But somehow I my crew the wrong place. I
161:44 a survey in Shula. I got back. I cannot using that I
161:49 land to determine property under you Always . That's example of no space.
161:56 have no control what's going on because I don't matter that right.
162:02 one example that's example of a B . How about example off you
162:09 Well, I didn't graduate survey in location. I have great. Imagine
162:14 . I measure them a measure I want to check in a lot
162:17 times. Same location, Same at same time. I found what I
162:22 changing. What will keep changing the occasion? What's going on?
162:28 that's no space, right? There no consistency off my measurements. They
162:35 running around anyway. I'm just giving concept of examples. So next one
162:42 something called bottles of Gilbert immersion. is commonly used in gravity data or
162:50 data about Barca's. Gilbert is very guy to professors. U C San
162:55 George George, uh, buckles Freeman . Okay, so their idea is
163:04 say, not worry about the inversion . But let's worry about where you're
163:09 for. It's kind of like constraining . They constrain the so called resolution
163:15 is in a way to determine where want to determine okay. I just
163:23 describe the idea. I didn't give enough detail. Okay, so for
163:28 , What do you mean? What you mean by the resolution? Measures
163:31 okay. Uh, let's say you understand relations. Okay, There's two
163:39 . There's tomorrow, but they're they're earth. There's model you found
163:47 Okay? Their model you found right , University model, whoever. The
163:51 making might not be exactly the same a true model. Maybe. Maybe
163:56 so good. So we're saying the you found are you inverted by modeling
164:02 inversion? Has a relation with the model. Through these measures is multiply
164:09 with your true model with model farm be the true model. What?
164:15 are wrong. You say the to future. You will be your model
164:19 that. A few different matches is the resolution measures is so resolution mentions
164:25 that is the future between your model the rial Answer. So what you
164:33 is that a few conversion of your ? So this guy's idea off.
164:38 Gil a bottle Bill Gilbert Emergency is or manipulating on the resolution so they
164:45 a resolution. The viable. Very . For example, Oil,
164:50 gas, reservoir our fault or So not a worry about mathematical
164:57 Just say, what's the idea? is the idea. This is a
165:02 idea for you Ever deal with certain of diversity? Certainly they are very
165:09 difficulty because they only have a service traveling. How can you use in
165:12 a way different Deep earth, which similar to gravity. So they're using
165:17 . Gilbert idea. They've manipulating our . Mattresses. Okay, I was
165:24 , you know, worry about And then they will answer over the
165:29 . You can see how much We don't have time for that.
165:33 , this one we also escape. just mentioning the idea. L s
165:37 the square. Cures mathematical operation. s Q R. S idea to
165:44 the SPD practically physical. The problem SPD thinking about the accommodation is you
165:53 holding everything into the memory of computer of computer has two parts,
165:58 Hard disk on Iran memory. Do know what I remember? Stand
166:04 I am Who? You guys still me last 10 minutes. Uh,
166:18 Stand for random access memory Okay. means any time we're used, you
166:23 it. This is like you asked like this this guy, Professor,
166:28 , you know, do you remember year? This day? What are
166:30 doing? I said no, I remember. But do you remember?
166:34 did this morning? Yeah, I . So the good memories, Rather
166:40 . The problem is, the remember, is very small. We
166:43 a big, bigger, hardest So this approach rescue are utilizing the
166:50 disk instead off using the random So instead of taking the whole data
166:56 , this actually scanned it up by piece of time so you can run
167:00 problem with us. You are, if you have a small computer.
167:06 . Years ago, I talked about ago. About 30 years ago.
167:11 inversion off about two million animals. I first arrived, your wage 1989
167:19 . That computers allow the computer. state is 35 years ago. Come
167:23 . I can do a version off than two million honors the emerging ones
167:30 like, six months. You got pageant waiting, right? So that's
167:37 sq are. I'm not gonna But this mathematical here's some example of
167:41 solution. Now I ask you are in California. There's some other issue
167:49 constraining version on some other issue about waiting function. So much, so
167:55 details. So I'm not schedules. last talk today is actually talk about
168:01 Virgin. Okay, I'm gonna burn . Find off. I don't want
168:08 go beyond. I think your stuff little bit. What is George
168:16 Well, is a technique that you a different viable. You wanna combine
168:21 example, I urge structure. Here's surface. Here's a my structure area
168:27 . I not only when they determine also politician intensity. So how using
168:36 ? Because using gravity, right? I combine both together? Gonna come
168:45 both because there's geology, same This is actually one area industry has
168:52 been doing too much enough. All right. So do you doing
168:58 together better? I'm not trying together opening in my opening years. Even
169:04 if this joint are complimenting because the emerging has a problem off.
169:12 they have a limit for example Now , their strength, their weakness.
169:17 the two part of this joint help other their weakness a strong, strengthened
169:24 the other methods. Then join them . That's what they call the one
169:28 one greater than two. You just putting together that's that's useless has to
169:36 more, you know, Addition. better has to be complimented. So
169:43 just give you one example. this is example. I used a
169:47 time ago. 1996. This is example. Earthquake location. Okay.
169:54 I show you here is something called Earth. Like a focal mechanism.
169:59 , you guys a geologist. So know what a focal methods. I
170:02 , these are to focus fears, of them. I mean, this
170:06 two focal noodle plans. Well, Israel fault, for example, if
170:10 really fault your operation be like This is no if that that was
170:16 fall or something like that. So will that Israel for what it is
170:20 , you know, axillary plan. a low hemispherical projection. I found
170:25 time that one, the location of wrong. So will be the
170:31 Okay, focus mechanism. So this a lot of detail here, but
170:36 don't want to mention one here, example. This is X Y Z
170:41 on Move them by one kilometer. Y z ok minus a plus.
170:46 was the minus plus minus on most the extra allocation of earthquake by one
170:53 . This is big area. A 100 kilometers Notice of my mechanism might
170:57 saying or change. Middle is my earthquake mechanism. This is my true
171:04 . The other ones might move. example, I moved here. I
171:08 here sometimes. Don't change but this says a lot. Okay, for
171:14 same That my mood location there Okay, sometimes you don't know what
171:20 is because determine earthquake location. There is four parameters, not three.
171:26 X y z. I wasn't there the time of origin One earthquake older
171:33 trade off between origin time, earthquake the depths, their trade off.
171:38 you have many, many answers. , the mechanism of earthquake only depends
171:44 the location. They don't depend on origin time of earthquake. So I
171:50 the idea to join together We do several Chinese inversion off for earthquake occasion
171:56 capacity, location on the origin time the mechanism combine. When combined,
172:03 get a better result. So you join emerging. So there's a lot
172:08 detail here. I don't need to to hell. I don't wanna tell
172:11 . That's the color. Better Okay, on. So the bottom
172:17 is, since you guys tired why joined something together, you gotta make
172:22 join the complimenting each other. You get a better result that way.
172:33 . I cannot rush you through. can see from the response. Normal
172:37 . Normal comments, people to say much. Too much. Yes,
172:42 stopped a salami. Okay, let's a quick review or any questions or
172:48 before we do our crease. It's a long day since we start.
172:53 morning is more than how many hours did have a break, right?
172:58 the talking in the break. Any ? Anything. Today you guys want
173:05 talk about it about scheduling. Don't take the material. I forget about
173:14 tomorrow because is someone next Friday covers of four topics is open book.
173:22 still you should know the basics. . So all of these lectures will
173:34 loaded on teams so that we can back and review them. Even the
173:41 . Yes. Yeah, I assume Maria still there. Yeah, that's
173:45 plan. Excuse me. Not on on team or blackboard on video point
173:53 the team's video. Beyond teams? , on video points. Oh,
174:00 . Okay. You please to the . Tell them where you Lord.
174:05 . Yeah, because Maria wanted to other other sides. She familiar?
174:11 , tell everybody where it iss. , I appreciate efforts. Actually.
174:17 tryto load everything onto teams. Another , it takes forever e will
174:24 Excuse me. I will send the . R L their email, and
174:28 can access to the video files very . Don't worry about that. My
174:38 . Yeah, we're in good Always like professionals like Maria. Thank
174:44 . Yeah, that's what the Because I almost forgot. Yeah,
174:48 of course, I think the best wise again is the best time for
174:52 to learn during the lecture. Slow down. And also, uh,
174:57 on the point you want to learn your time is limited, even though
175:01 might think that more time. But in mind, this is like,
175:05 . Yesterday is really three hours Another 76 hours. That's only nine
175:10 . How many hours you have? flies, right? So you better
175:16 questions during the lecture. Also previewed . Maybe this this first top Because
175:23 don't have time, little material. now you see the style now use
175:28 . So we should be better prepared next week. Make sure you really
175:32 like four. You go to class I can bring question the class.
175:38 don't have to really carefully. You everything. Just scan through. Just
175:41 sure there are few slides. You questions. You know, learning is
175:47 process off. Getting something new materials foreign now is very troublesome. I'd
175:57 to choose the material fits my For example. Typically about 60%
176:02 30%. 0, no. You , in between that's better for
176:06 But if too much I know I bored. I can't see anything.
176:10 don't know. Anything is too Everything is now. It is boring
176:16 . This is why, for your , wise for the previewing previewing the
176:21 . I'm going to try to make happening. Loading the slide for at
176:26 on Tuesday. This is my Maybe even before. Maybe can do
176:30 on Monday because I'm teaching Monday, I'm not even trying Monday, But
176:34 should be able to see the the topic next Friday, on Tuesday
176:40 before Tuesday. Hopefully, you can loading the topic of 56 Also,
176:47 way you can have time to say you're gonna see. Maybe it's so
176:54 . To be honest with you. wouldn't be surprised someone if I'm doing
176:58 class like you guys. Um I only one listen to part of I
177:02 know. I know I can get other class materials or something else.
177:06 , that's fine, But do asking toe, verify your learning. That's
177:16 you go. It's a good Small class. So we can do
177:19 . So, are you ready for quiz? Yes, I hope.
177:24 , sir. I'm gonna try something . I was ready to see the
177:29 , but I'm gonna I'm gonna put Queen on the team. I'm testing
177:34 you. Make sure you can see on t. I'm putting out there
177:39 . Even this talk about it says already skeptical team name is called the
177:45 S. Q three. Also put P s Q 212 There. Now
177:55 have asked the team. So should all see all the other creeds
178:00 We have a Q one Q 2123 , I mean, I can see
178:06 . Good. We're very good. . Thank you. So I'm going
178:11 us Usual. This plan was meaning the history. In case anybody have
178:17 questions. No panic. Okay, you don't see it, I can
178:20 the email. Don't forget when you , please. Sending with Maria.
178:31 , it gets into me. I'll her. I did yesterday. But
178:34 is better because she is greater. going to go where? The answers
178:40 week on Friday. So once you the question, no more questions can
178:47 on that. You don't need to the stay with me. I can
178:50 it easy. Okay? Usually I I gave time limits. But if
178:55 want to believe the delay, I think Maria penalize anyone yet. Of
178:59 we want you finish it on But if you don't, you know
179:02 five minutes, 10 minutes I don't speed deal. Okay? Yeah.
179:10 . So let me let me see queen Three. Queens three. Here
179:16 go. Here is the sharing off Okay, here is this. So
179:27 don't want to trouble you with more college anymore. I just doing the
179:31 as the same as yesterday. so I gave it 12 for the
179:36 question. I'll give you two The difference. I'm not sure like
179:43 way or not, but I don't We gotta do something quick, easy
179:48 anything speaking, but I need to your really? So let me bring
179:52 quickly from first one. Joining version two variables is preferred while a solution
180:00 each other. That's the first one apply. More constraints, Maxine Solution
180:09 unique. Okay, this is a toa. I didn't say too much
180:13 that in lecture, but we have and talk about that. We're saying
180:18 or not. Is not Is a term unique means fewer choices. Last
180:26 make more choices. Okay, unique. Listen to three de convolution
180:35 zero face lately because of his best four least squares is working. Assumes
180:46 is misfired. Follow Gaussian distribution. low velocity layers are difficult to be
180:56 by first rival tomography. I didn't to teach you about tomography today,
181:01 I just wanted Let's just think what on. The problem is that we
181:05 a lot of us last layers. first rival will not try to stay
181:10 from that. So the coverage there the law six predictive the conclusion Surprise
181:21 more effectively. Are the five set year offset? I cannot lose this
181:27 . A little bit. I didn't example, but this is actually
181:30 What? I mentioned this. Why described this. Okay, seven resolution
181:39 . Off. Assessment inversion depends largely the simple, pronounced ratio of
181:46 Is that true? Now, this about Resolution 87 imaging in petroleum
181:55 I'm all three. Based on reflection , nine Such modeling is guided by
182:04 such as whether the model is geologically . 10 some measures is cannot be
182:15 by S VD 11. Usually, is a superior to modeling. One
182:24 data coverage is very poor that our also means you don't measure.
182:32 so that's not gonna measure. Talk last one. Usually such modeling is
182:39 opportunity. That's that's being inversion. you see, vague, unclear
182:50 Not sure, right. You're doing right. All right. Uh,
182:59 comparison straightforward. We compare difference between dramatic are under producing. What a
183:08 between those two. They could be in the assumption or difference in the
183:14 the process. Okay, last one version was the buckle skill, Bert
183:21 . I just want to give you conceptual comparison. You don't have to
183:25 anything too much in here as long see the main point. All
183:31 folks, this will clear now at the now. No major error.
183:36 if they're issues like I said beside circle principles, you can also write
183:41 about What are you thinking? Even case of a mistake you can maybe
183:46 can give you no listening for. have problems or something, I
183:55 All right. Okay. It's been long day. I'm amazed we're still
184:03 . I was still No, I'm teaching me. No, I'm still
184:07 pain here. Yeah, I squeeze my here. I feel people.
184:11 , I'm still alive. Okay, , it's easy. I appreciate you
184:20 efforts. It's not easy, but work together. We can learn together
184:26 , uh, any issue between Next Friday should be email,
184:32 Maria Mia
-
+