Mtg 13/26: Tue-25-Feb-2025

Outline for Today

Midterm prep and Chapter 6

Administration

Today

For Next Meeting

  • Good luck on the midterm

Wiki

Link to the UR Courses wiki page for this meeting

Media

Transcript

Audio Transcript

  • So I have A slow network. You
  • opposite. So how was your break?
  • I feel broken. That's how my break was.
  • It's
  • yeah,
  • no windows is poorly optimized, but I know Windows is poorly
  • optimized,
  • like mine graphics thing, so I don't really have a designated
  • graphics to
  • be designated for anything. Mine's
  • poor still running, still running, it
  • still works. Just replace,
  • yeah, those are good.
  • It was like a 19 $100,000 machine, so at this point,
  • wasn't like, and then I've upgraded price wise, works for
  • pretty much everything. It won't update Windows.
  • I'm kind of happy that is actually good.
  • Music. For other things, it can
  • be it's been a bit of a whole new chips, pbrt running
  • on my machine a whole do like
  • I'm just gonna look at the code as you
  • change GDR
  • testing, and
  • I'd rather focus on the 15% return and the reason being,
  • works better than
  • complete. And the reason being is four
  • used, kind of like, I Just understand we're actually, like,
  • video just past session. I
  • it. So any questions about the assignment that's due tonight?
  • Are there any specific kind of like textures that are kind of
  • like required because, like, I've used a plastic texture for
  • like one, and we use like a bubble kind of texture, a bubble
  • kind of texture
  • for the other. So is there any specific
  • requirement? No, yeah, I asked you to do different materials.
  • So the network is not cooperating with me. Even on my
  • phone, it's got a weak connection. So I'm going to I
  • switch to version on My laptop here I'm
  • so that'll give you a look at what I was planning to talk
  • about today.
  • Give me a second. Okay, I
  • about, Does anyone else Having network problems?
  • I Don't know what I cool.
  • I did a calendar of fractal images, or I edited it all. Here
  • are the links.
  • So I guess with my local, hosted version, I'm trying to access a
  • CSS file that's on the web
  • and it's not happy.
  • I think you might need to outline your text. I
  • I need to, I
  • need to move this, the CSS that's caused me the problems.
  • Make a local copy. I doesn't like
  • that. Never had this problem before. Every
  • single person
  • gives us the background was titled
  • in your Text, glitter.
  • I wish. I wish people remember pixel. I
  • Where do you get the fractals from? Can I get a copy of that
  • one I
  • yeah, I've been meaning to put the large scale resolution ones
  • up, so I will do that.
  • This is from 95 I
  • think so. This is a van Cook, snowflake curve I
  • spend it over a height field. This was Ray tree and
  • networks back.
  • Maybe, let's
  • Okay, so this is The assignment. So we're using materials.
  • So the main sphere starts out being with the dielectric
  • material. So add two more spheres with different materials
  • and
  • Okay, are there any other questions about the assignment
  • then my post yesterday,
  • answer some questions I didn't. There was no reply to it, so I
  • wondered.
  • So I thought better of getting into chapter six stuff today
  • based on the conversations we've had online. So I just focused on
  • the
  • the midterm that's coming Up next Meeting I
  • in Section chapter one, section two
  • introduces a lot of ideas that we're exploring This semester. I
  • so they talk about the light transport equation, also known
  • as the render rendering equation. I
  • So, what is the purpose of the rendering equation?
  • What exactly was the question?
  • What is the purpose of the rendering equation?
  • Total, realistic rendering. I
  • let's say it's to do with simulation as well.
  • So An indistinguishable simulation. I
  • So we're dealing with
  • light in the scene. Well,
  • so this video doesn't have any sound. It's about Three minutes
  • Long. I
  • information, Please Do
  • the best.
  • Practice your playback Speed up really High. Oh, my
  • God. It kind of like, because the SDR does not kind of like,
  • connect like correctly, whatever does that kind of like
  • accelerates it to some example, what
  • you're
  • saying is my cable might not be connected. Yes, does.
  • If it wasn't connected properly, wouldn't be able to see the
  • picture on the screen. I
  • I mean, if my phone is plugged into a charger, it can still
  • charge. If it's not plugged in properly, it's just it will
  • charge slower. I
  • THE END seems to be the right speed. I
  • here's the Wikipedia entry that
  • maybe gives A bit
  • more accessible presentation and
  • so the idea is, we're looking at the point x and seeing the light
  • that gets emitted from that point x in the direction and
  • W, o,
  • so we're dealing With the con looking.
  • I so why isn't,
  • let's Say, four samples per pixel. I
  • What's that? Number 4096,
  • yes, power,
  • I mean, n samples. It's just like when you're doing
  • statistics, you make more variance and stuff when you have
  • a small sample base. So, yeah, that's why you choose a second
  • to 30 should which should Set End to 30? I?
  • It so
  • it makes sense. The four samples per pixel is going to be as high
  • quality As 4096 samples per pixel.
  • What are some of the particular
  • challenges, if, if we have four samples so,
  • oh, you showed us the outer gamut pixels. So I imagine
  • there'd be way more of those. I
  • so all light Sources may not Be counterpoint.
  • At some point improvement you Get from adding more samples,
  • but
  • so if we look at Like the Image tool diff, we'll see the images
  • getting the subsequent images will have less variance, and
  • then when they're threshed, the variance is below a threshold,
  • We can say
  • this is a good number of samples per pixel to use And
  • yes, I Excuse me, Please Do
  • the idea was, they're supposed to open In Order. I
  • so this is the realistic camera, and I'm setting the shutter. The
  • shutter open time is zero, and the shutter closed. Time is
  • starting with 1234,
  • so that's shut or close At 10 and a 20 and a 3040, 50 and
  • 60, 7080, 9100, and that's 200 And it 200 seems like it's
  • overexposed. You
  • it almost seems like it's a percent, like at 100% it looks
  • good at 200% it looks 200% is bright. Like, too bright. It's
  • yeah,
  • I'm not sure what the units are on that time.
  • Usually they're about, like, some x over like a fraction of a
  • second. So like, you have, like, one over 60 is the 60th of a
  • second. And then if you have like, a time two, that would be
  • 200 over 100 so I'm assuming this is about 100th of a second
  • increments. So that'll be two seconds of exposure for the last
  • one,
  • unless you added it like aperture, but
  • what it's doing is it's adding all the light in that time. So
  • just we see stuff, and then we keep refreshing our vision,
  • whereas it's just like, what if we add all the light from the
  • last 200 milliseconds together? This is what it would look
  • like. Well, I mean, our eyes do the same thing, though. It's
  • like a running average. You
  • and so I think it's percentage and average, right?
  • Do? Okay, so David had a question last day, I believe,
  • before the break,
  • about why changing the position of the camera didn't make sense.
  • I don't have a phrase. It was weird. The
  • behavior was weird. Yeah, I
  • so can you look at this input file and see what might account
  • for that weirdness? I
  • camera.
  • Perspective, perspective,
  • so using
  • perspective, yes, it's perspective. So we have a
  • viewing frustum that's a truncated pyramid. So
  • that changes to the camera
  • would have to be normalized before we'd be able to properly
  • use the perspective.
  • Explain that please. What
  • do you mean for putting in any vectors
  • and stuff that we're putting in different cameras we're putting
  • in one particular view, so we need to normalize it to be able
  • to get it to be proper to the actual field of view and The
  • world of you, in regards to your frame, You
  • so
  • let's think about the camera frame
  • how
  • So a common and straightforward way to specify the camera frame
  • is to the to specify an eye point that's where the camera
  • position is going to be,
  • and then a look at point, the point you're Looking at, and
  • then another vector so the eye point look at point defines one
  • axis of the cord of the camera frame your out
  • vector is usually
  • defining another one. Yeah. So
  • when we specify look at or the I point, look at point. That's one
  • vector of our coherent system.
  • But we can still it doesn't fully specify the frame, because
  • the up vector can rotate around that, that I point look at point
  • vector. So once we fix the up vector, so we have to look at
  • the i point, look at point vector, the up vector, and then
  • we can find the third vector by doing a cross product between
  • these two vectors. I
  • call that on the right vector personal
  • reference. Okay.
  • So how have we defined the camera frame in this sample
  • input file? Well,
  • we set up, but we're also supposedly like, aren't we
  • looking down? I'm confused, because the I is above the
  • lookup point, so we should be tilted downward. So we're
  • declaring up as one, one. I guess that's so
  • the up vector is, is more so, like, it's like, you have,
  • literally, an arrow pumping up. You have an i, which is just a
  • point, yeah, it's like, you draw a string between here the camera
  • relative
  • to the Office of z, it's in the third slot
  • and one. So you just, yep,
  • that's, that's the thing, yeah, when we change the Z, it moved
  • backwards, or, like, it moved upward. And I'm like, that's
  • weird. That's the reason, right there, we're declaring what the
  • z is, our up. So that is our Y,
  • yeah, in
  • a lot of 3d space, they use Z is their up axis, as opposed to y.
  • I would say y is usually the up vector.
  • It's more common, but I've actually been reading about
  • this, oddly enough, over the break, of people using the Z as
  • the upward vectors the third axis, yeah, the third axis
  • becomes up because it's now, no, it's leaving the plane,
  • therefore it's adding the up. That's the way I've seen it a
  • few times. I actually this makes a lot of sense, but they're just
  • using it that way. It kind of seems arbitrary. Honestly,
  • I so if we switch the numbers from those two columns, it
  • should fix it for us, at least. Yeah,
  • you'd also have to change where the light source is as well to
  • make it relative to us.
  • That the Y in that file, yeah, so I've changed it here, just it
  • might be a good idea to change the light source so that the The
  • Y is 100 and The Z is 40. I
  • Let's do this One First and
  • any ideas about how it's just going to Look. I
  • feel like the light might be in a weird spot, but hopefully it
  • works. Oh,
  • it looks cool.
  • Is it has we then changed the world plane, where the like
  • world where the world plane is, because all those coordinates it
  • was given were relative to having the Y as one axis and the
  • Z is another. That's why I was saying we should swap that line
  • value. But if it's doing it for all the coordinates, you have to
  • make sure you're at least consistent on everything.
  • So so the simple pbrt file would be doing it in a different
  • direction too.
  • I that'll just move the light though You have to do both the
  • checkerboard, the ball and any other elements. Let's see what
  • this test first. It's just gonna move the light,
  • which might look cool.
  • Yeah. So now the lights casting the shadow downward, because
  • that's the good fraction We switched. Now the ground was
  • down there. Okay, okay.
  • Yes, you have to change it for every single thing. So I think
  • it's just better to accept the fact that it's using a different
  • system than wait a second,
  • it's got a translate
  • function. It's translating,
  • if we got rid of that,
  • fix everything, zero,
  • minus 101, negative one. But it would have to be zero, negative
  • one, zero. But then you see those points. They would also
  • like, would they a translation? Is not a multiplication. It's
  • not, it is not, no, it's just moving. It's like, it's saying
  • we moved it down, or, in this case, to the left. And then all
  • points are relative to that. So
  • everything's relative.
  • Say that. Why wouldn't we just wouldn't we just move the
  • cameras to one model view or one view?
  • Anyways, that's what I'm hammering saying. Everything was
  • in X, Z, Y, essentially. And so rather than switch everything,
  • should accept that it's using an alternative format.
  • So if
  • I were making a file from scratch, I would definitely do
  • X, Y, exactly.
  • I mean, choice of coordinates are
  • kind of arbitrary. It is, it is arbitrary. It's just you have to
  • know what it is, see, that's what happened. That's exactly
  • what it's expecting do.
  • Making sense
  • actually looks pretty cool.
  • Texture mapping over here is suspect
  • that's actually pretty cool. It's nice. It's nice to go
  • Yeah, every single one of those points you'd have to go like
  • negative 20, zero, negative 20, anywhere that there's a, x, y,
  • z, you'd have to switch the last two,
  • unless they're identical. That is, and then, yeah, obviously it
  • excludes things like RGB, which are not vectors. Per se, I think
  • the problem we're having is because of the setup. The other
  • way, yeah, exactly because I'm saying, rather than change all
  • the values, it's just might be better to accept it for this
  • file, any subsequent files we do make, we can set to whatever.
  • No, we were trying to modify this last week. Yeah. Sorry. So
  • the problem we get this got brought up is because it was a
  • problem because it wasn't working as We expected it to.
  • Yep, Now we Know why you
  • Fix the texture And
  • now the lights still in a weird position, that's why the shadows
  • off the side, but now it's a sunset, so that's Cool. Look.
  • So any other thoughts about the exam?
  • We're mainly focusing on the stuff we discussed in class, as
  • opposed to like, what do files and pbrt look like? Is that
  • correct? Or will there be some asking about the file structure
  • and stuff like that?
  • I I think I could ask a bit about that.
  • Will you, since today's quiz was all about midterm questions,
  • will you be sending out the list of the midterm questions? And
  • people suggested
  • Yes?
  • How? Much of the midterm will be based on questions that people
  • have suggested I
  • there won't so I would say that there's five chapters, two or
  • three questions per chapter. I don't want anyone
  • to hurt themselves writing, so I will take them as input.
  • So the questions will be based on your input and my sense of
  • what are interesting,
  • not just interesting, but
  • maybe will be a bit deeper than what's been suggested. So I get
  • a sense that you understand the concepts.
  • Does that make sense?
  • Will they all be short answer? Will there be a mix of short
  • answer, multiple choice?
  • I'll see if I can put in a multiple choice where two space
  • questions like transportation matrix.
  • There Yeah, I'm not
  • a lot of those you mentioned earlier in the semester. It's
  • more about learning the concepts, yeah, the details
  • of of everything.
  • Is that correct? Yes. So I was going through, I was going to
  • put the links to the Wiki pages, but there aren't any. I thought
  • people were signing up and it was going they just people just
  • didn't post the class discussions list. I was wrong.
  • You
  • Any final thoughts, yes, how long do you expect the start to
  • finish
  • for you or for everybody else? The slowest
  • person, I would hope it would be
  • such that you have time to comfortably review your answers.
  • And I'll commit to saying between 10 and 15 questions.
  • So you'll send us, I would assume you'll be reviewing and
  • sending lists today
  • or tomorrow. Yes, the last 30
  • minutes. Yes, although maybe you'll be done in 10, I don't
  • know.
  • Thank you for today. I'm gonna see about resolving these
  • network issues once and for all.

Responses

No Responses