Mtg 15/23: Mon-04-Nov-2024

Outline for Today

Midterm and Lighting and Shading

Administration

Today

Midterm
Lighting and Shading
- Chapter 6 code

For Next Meeting

Wiki

Link to the UR Courses wiki page for this meeting

Media

Transcript

Audio Transcript

I forgot
something. I just forgot something. Yeah, I
time. Okay, it's 232, 230
let's see if The projector is going to work today. You
I called them about this on Monday last week, and I got an
email today saying the ticket was closed. I tried they had
done
They had done something about it.
Okay, that's looks better. Fingers crossed.
So
has anyone heard of fractals before anyone hear that word?
Okay, so I've done a Web GL program I'll share with you a
bit later,
that generates shapes like these and
and I just wanted to share with you.
So this is
so these are screen captures, and this little icon is in the
middle and says, Apple thinks this is a plant.
So it thinks this is a Norway maple. And
and this one is an aloe vera. So
wrote, wrote, wrote a book called the fractal geometry of
nature. He talked about fractal shape which are self similar,
being very good at describing natural objects.
So they were being computer graphics and
some years ago,
and like I said, I'll share my web channel two program with
you,
just needs a bit of cleaning up.
So I thought I'd go over the midterm. I've started marking
it, but I'm not yet complete. Haven't completed that yet,
but I thought we could go Over the exam questions, does that
seem alright?
Post a questionnaire for you to give me some feedback about the
exam later today, I
so I tried to follow your suggestions about making short
answer Questions. I didn't do multiple choice per se, but I
but what do I want to say I took a lot of the questions are well,
some of them are directly from the midterm we talked about in
class last week. Other ones questions are rephrased the same
material that I was getting after.
So the window is 512 by 512 pixels and
what's the width and height of the clip space? Does it depend
on the size of the canvas?
Yes, no,
I saw one person put their hand up, maybe more. Yeah, it's not
related to the size of the canvas. Anyone know what the
width and height of the clip space are? Yeah? That's two,
because it goes from minus One to 1x
so the proportion is two divided by 512, one over 256,
I also so. People who did one over 256, squared. That was also
a k.
So we have a
the model in terms of
when we get to eclipse space, we have everything expressed in the
range minus one to one, minus one to one minus one to one. And
then we now
map that into Window cord. That's so
We're
Does that make sense? I
so if we had 16 by 16, then we'd have 25 by 16, one, eight.
So
we still have the
same range
between minus one and one, but we're dividing it much more
coarsely, so there's a bigger proportion. Of the clip space
per pixel.
Any questions or concerns about that? I
Do you have a question or Comment? Okay?
So we can think of RGB as three RGB as three coordinate axes. So
we go from black to white to
so this Hue Saturation and Value.
Looking from the top, we have Green. Red, Green,
Blue, Janta, Cyan and
I imagine those are equally spaced around a circle.
So from the side we go, value zero to one, black to light.
These are grays and
and we go from white to
the pure colors. And
so this is s the saturation and
so we go, if we have red, let's say we go from white to
different shades of paint, and we get more and more red until
we get to pure red color. I
that's just a bit of the intuition behind it, any
questions or concerns? Applause,
we create a computer graphics image, we are Taking
continuous renowned, I
discretizing it and
the fidelity of the represent of our discrete and
the fidelity of our representation depends on Two
Things You
so we can see the number of samples we're taking. We have a
camera with how many megapixels as you can.
Have 10 megapixels, so that's a lot of samples. And the other
thing is, how much storage are we allocating to recording each
sample? So if we just have one bit per sample, we get a very
high resolution black and white image. But if we store 32 or 48
bits per pixel per sample, we get a lot more,
A much higher quality representation.
Does that seem Okay? I
show between red, green, blue and sand, Bucha yellow, RGB is
additive. CMY is subtractive and
so we'll use RGB on displays and use CLI for printers and
bless you.
So why do we sometimes have, or oftentimes have black ink and
because if we want to print something in black, if we don't
have black ink, We have to print it three times in cyan, magenta
and yellow, so we in 30 will get Black. But
if we're using ink, the paper might get soggy.
It's much easier to do one pass with black instead of three
passes of full cyan, magenta and Yellow, Okay,
so the object space, we're thinking
of The representation of the models, cubes, spheres, whatever
an image space. We're dealing with,
pixels, screen coordinates and Z
buffer is an image space algorithm, because we're testing
the depth of points and clip space and
Ace, is that Okay? Any questions?
Triangle strips and this is the same question from the midterm
we looked at last week, except are there alternatives? I
more efficient than triangles because we don't have to specify
three vertices per triangle, we can link them together so
so Latin will convert the data into a float, three, Two array
that we can pass to Web GL and
okay, Then the next three questions related to This. No,
I so let's use For Visible Surface determination and
so we're going To be using the physical search determination,
it has The depth with each of each pixel. So if depth,
new pixel value
is less than the
store nine And
so if the new value that we have for position on screen, if the
depth of the new value is less than or closer to the viewer,
then the one in store e update that pixel with the new value,
And then we update the buffer.
Yeah, okay, does That make sense? I
so what three things do we need for the lookout, a way to
specify the position of the camera i
So i is where we're we have our where the camera is Located, and
so we have E and
A And then we specify an up vector and
so that gives us two directions, and then to do The cross product
of the vectors we can get. So
our moral Thank you, our third coordinate axis.
I Okay.
So I gave you some share Code and JavaScript here some
fragments and
so I recall that
when we talked last week, I About exam questions that.
Questions about, what does this code Do?
Were favored by some people. I
so this is a very basic vertex shader. It's just taking the
vertex position that we've set up in the JavaScript and
attached it to the a position variable. Then the fragment
shader is just setting a constant color.
And here's the bit for the JavaScript file where we're
setting up the vertex buffer and associating a position so the
after vertex attribute position with that variable that links to
the shader, the vertex shader. So
okay, so If we're going to make per vertex colors? What's the
first thing?
We have to change something in each of these bits in the vertex
shader, the fragment shader and The JavaScript.
Okay, let's start with a vertex shader. I
so we need an a color.
And so a color would be in, and B color would be out, and
does That make sense? Okay, what About the fragment shader I
so we carry the color we've created, we've assigned In the
vertex shader as an out V color gets passed out of the vertex
shader, and then we connect it. Here we say in V color. And
then, instead of assigning F color to be a constant red,
1001, and we say f color equals V color.
And so what do we need to do in the JavaScript? So Buffer. Yeah,
so we just, we're going to use the same structures we have for
the vertex buffer, except for colors, and we're going to make
it a four
black, four or four positions per color, and we're
going to link The
color buffer to a color I
so it's this basically, except
in the vertex attribute pointer
could be color location, And instead of two, it will be A
four and
does that seem okay? I
so I was being too clever with this next question. I
this is a bug that I encountered in my own code, the one that
generated the fractal shapes
so
what I could have said in question 11, I
three by three Matrix and
and dream do
so What the output? So I
get that point, since it's identity matrix, that's not very
interesting, but it's easy to do the math on the board for that.
So I have a VEC three that I'm pushing to
positions
now the
so I have three values for each position, because I've, I've
been pushing back three two positions. I
But what am I telling it to look for in the second last line? How
many? Yeah, two values.
And because I flattened it, so I don't have the idea that, Oh, I
just need two of the three values from the vector, three
that are stored in the array. And
so i get i
So what happens is, let's say these are guys are pushed to
positions.
And I'm going to get a
Does that make sense? So we have data for three
positions that were in we're
sampling or taking values off out of that vertex buffer two at
a time.
Does that make sense? I
instead of
let's do put them
out like this,
so 001, 0.5 0.5 the
so we get those two,
those two, those Two,
and those two. So this one is what you expect. This one too,
but these aren't.
I really want to have 00, 10, 0.54, 0.5 0.5
I've split that over two, so I should get 00, 0.5 0.5 0.25
0.25, and point seven.
Does That make sense?
Okay? I
Does anyone recognize that Matrix? I
Oh, this is in that matrices dot pdf,
so describe The effect of this transformation, a rotation and
you can
say more detail about how coordinates get flipped and
so how would I would you go about computing the
transformation matrix At centers the above rotation at point
five minus seven, two. So what do we need to do?
So we had our rotation, but we want to apply it to a point
that's not at the origin, yeah, I thought you had your hand up.
So what do we need to do?
Or Yeah,
so we can write that as a
so we're going to do minus five plus seven,
minus two.
Then we're going to apply the matrix, and then we're going to
go
five minus seven, two and to move
so we can bring that as multiplications as I
then we can do the rotations, And then you do
three. So we get, I think it's minus 314, three. I might have
the signs mixed up.
So how do we verify that we have the rotation centered at that
point, applied to 572,
and she had the same one back. Yeah, that's just like we did in
class. We took the center of the rotation. So because it's the
center of the rotation, it's the fixed point. So the image of the
center of rotation under the rotation is the center. It's the
same thing. So yeah, can
you go back to number 12? I think you might have skipped it.
It was on that previous page,
the last 20. Oh, yeah. How does it work? You
so I've noticed that some people are answering it after they've
made changes in question 10, but I would say it's it works
because we have a good set of defaults, because things are
lined up. And
we're looking down the Z axis.
So we don't need to do a model view matrix or projection
matrix, to get something to display on the screen.
So then the limitation part, well, what's the caveat if you
want to see something on the screen?
What's the limitation? You have to have coordinates and what
range and
I Wonder if we had a minus 10. I
would we be able to see something that has an x
coordinate of minus 10? No the
minus one work.
So the same in y, so the coordinates have to be in the
clip space if they're not they get clipped. We never see them
again.
That makes sense. You.
So here's, this is chapter Six, lighting and shading. I
so have a look at the code For shaded cube and
shaded cube to the
Okay,
any questions or concerns. Thank you for today. Have a great rest
of your Monday. See you on Wednesday. Take care. Thank you
for today. Thank
you.
You're welcome. You
You're welcome. You

Responses

What important concept or perspective did you encounter today?

Going over the midterm examination answers
The important concept we covered in the meeting today was reviewing the answers to the questions from the Midterm last week.
Midterm review - checking what I got wrong is indeed pretty helpful.
Midterm Discussion
Homogeneous cordinates
Midterm review. I may have performed more poorly than I thought based on the answers provided in class today
We looked at the midterm exam questions and solutions.
Went over the midterm and learned more about lighting and shadings
Midterm answers, Q11 vertex shaders and transformation matrix
Discussed the Mid-term paper
We review the midterm and the professor explain the answers

Was there anything today that was difficult to understand?

Was there anything today about which you would like to know more?

Site built: 2024-12-12 00:01 -0600 • Give feedback about page