SlideShare a Scribd company logo
1 of 8
Download to read offline
Bezier Curves with Quadratic Control Point
Interpolations
Nicolas I. Miguel
June 1, 2016
1 Introduction
Bezier curves are curves that have many applications, from car contour design
to font recognition1
. The curves are exact sections of many parabolas and
conics, and are the result of many linear interpolations/line fittings, which are
correlated with the degree of the curve. Furthermore, the curves have control
points that define many of the aspects of the curve, such as the degree of said
curve. In addition, Bezier curves are essentially parametric equations.
Throughout the last quarter, I have focused on Bezier curves that solely had
linear interpolations. That is, the connections between different control points
would be linear. In this paper, I will be exploring Bezier curves with quadratic
interpolations, where the connections between control points are in quadratic
form, instead of linear. The driving question behind this analysis is: How does
the degree of the function of a Bezier curve change depending on the degree of
interpolation?
From my last quarter papers, we can remember that a Bezier curve with n
original control points will have a degree of n − 1. This was only tested with
Bezier curves that utilized linear interpolations, therefore, I don’t know if this
rule will hold for Bezier curves that utilize quadratic interpolations. Therefore,
the goal of this project is to find if there is a rule which allows you to find the
degree of a Bezier curve depending on the number of control points and the
degree of the interpolation.
Figure 1: Bezier curve with linear interpolation
1Source: Wikipedia
1
Figure 2: Bezier curve with quadratic interpolation
2 Curves with 3 control points
We start by observing a Bezier curve with 3 control points. With normal linear
interpolation, the skeleton of the curve (that is, the control points and the
connections between those original control points) looks like this:
However, we are looking at quadratic interpolation. Thus, the path between
two control points will have to be quadratic. To define these parabolas, we’ll
consider P1 (the original control point farthest to the left) as the vertex for a
parabola which goes through P2 (the original control point to the immediate
right of P1). Furthermore, we’ll consider P2 to be the vertex for a parabola
which goes through P3 (the point to the right of P2). We’ll have that:
P1 = (−2, 0)
P2 = (0, 2)
P3 = (2, 0)
We consider the first interpolation, which is a parabola where P1 is the vertex,
and P2 is a point the parabola goes through:
y = a(x − h)2
+ k
y = a(x − (−2))2
+ 0
y = a(x + 2)2
2 = 4a
a =
1
2
y =
1
2
(x + 2)2
2
We repeat a similar process and get that the parabola that has P2 as a
vertex and goes through P3 is of the form: y = −1
2 x2
+ 2. This looks like so
(note that I’ve restricted the domain of both parabolas so that they stay within
the bounds of the skeleton):
Now we need to find the parametric equations for the two points that move
along the two paths. We’ll call these points L1, which travels on the first path,
and L2, which travels on the second path. Starting with L1, we see that at
time t=0, L1 will be at (−2, 0), and at t=1, L1 will be at (0, 2). Thus, we
find that xL1(t) = 2t − 2, since at t = 0, L1’s x-coordinate is at -2 and at
t = 1 it is at 0. However, to find the y-coordinate we need to manipulate the
points a bit differently. Instead of only dealing with L1 at t=0 or t=1, we also
need to consider where L1 is at t=-1. This is because the y-coordinate of the
point does not stay at the same rate, rather, the rate changes because the path
is a parabola. Thus, we need three points to calculate the L1’s y-coordinate
parametric equation. We already know that at t = 0, y=0 and at t = 1, y=2.
However, since this is a parabola, we will consider that at t = −1, L1 was at the
point (−4, 2), thus at t = −1, y = 2. The following graphic best details what I
am doing:
Thus, we see that we have three points in the form (t, y): (−1, 2), (0, 0), and
(1, 2). Naturally, we observe that (0, 0) is the vertex, thus, we solve like any
3
other parabola:
y = a(t + h)2
+ k
y = a(t + 0)2
+ 0
y = at2
2 = a(−1)2
2 = a
y = 2t2
Thus, we have parametrics for L1.
L1 = (2t − 2, 2t2
)
We repeat the same process for L2, proposing that at t = −1, L2 is at (−2, 0),
like so:
We find that xL2(t) = 2t, and after setting up three points in the form
(t, y) ((−1, 0), (0, 2), and (1, 0), where (0, 2) is the vertex), we find that the
y-coordinate is given by:
y = a(t − h)2
+ k
y = at2
+ 2
0 = a + 2
a = −2
y = −2t2
+ 2
We have the parametrics for L2.
L2 = (2t, 2 − 2t2
)
Now that we have the two secondary control points, our curve skeleton looks
like this:
4
L1 is the red point on the left path, and L2 is the red point on the right path.
The next step we take is finding an equation for the parabolic path between
L1 and L2. Let’s consider that L1 is the vertex, and L2 is a point on that
parabola. Once again, we set up:
y = a(x − h)2
+ k
y = a(x − xL1)2
+ yL1
y = a(x − 2t + 2)2
+ 2t2
yL2 = a(xL2 − 2t + 2)2
+ 2t2
2 − 2t2
= a(2t − 2t + 2)2
+ 2t2
y =
1
2
− t2
(x − 2t + 2)2
+ 2t2
Once we find the equation of the path, we have (in red):
The final step is to figure out the parametric equation for the final control
point. The point, which we’ll call R, will travel along the red path, which we
just figured out. At t = 0, R = (xL1, yL1). At t = 1, R = (xL2, yL2). Figuring
out the parametric equation for xR is simple enough. We know that x goes from
-2 to 2 in 1 second, thus: xR(t) = 4t − 2. For the parametric equation for the
5
y-coordinate, all we have to do is plug in xR(t) for ‘x’ in the equation of the red
path (y = 1
2 − t2
(x − 2t + 2)2
+ 2t2
). We get:
y =
1
2
− t2
((4t − 2) − 2t + 2)2
+ 2t2
Ultimately, the parametric equations for the final control point R are:
(4t − 2,
1
2
− t2
((4t − 2) − 2t + 2)2
+ 2t2
)
The skeleton is finally complete, and the curve is traced out by R (the blue
point):
Now we solve for the path traced out by R by solving for t in terms of x and
then replacing that for the t in the parametric equation for y, like so:
x = 4t − 2
t =
x + 2
4
y =
1
2
−
x + 2
4
2
4
x + 2
4
− 2 − 2
x + 2
4
+ 2
2
+ 2
x + 2
4
2
y = −
1
64
x4
−
1
8
x3
−
1
8
x2
+
1
2
x +
3
4
The function that results is of degree 4, which is the degree of a linearly-
interpolated 3-point Bezier curve (2) multiplied by 2. Here we observe the
graph of the Bezier curve, which is in orange:
6
3 Curve with 4 control points
We repeat the same process as before, setting up a skeleton and solving for each
set of respective control points. Ultimately, we get a path traced out by the
blue point:
The final path looks like this (in blue):
The function of the path is of degree 6, which is the degree of the linearly-
interpolated 4-point Bezier curve (3) multiplied by 2.
4 Conclusion
With the two examples we have been given, we see that the final degree of the
curve is the degree of the normal Bezier curve multiplied by 2, which happens
7
to be the degree of the paths between the control points. Let’s consider the
paths between the control points in the curve skeletons. These are the green
paths in the diagrams, or the first set of interpolations. Consider looking at
them as individual Bezier curves. To get a quadratic Bezier curve with normal,
linear-interpolation, we would need 3 control points. Consider the diagram with
the 4 control points, above. There are three green paths, which connect P1,
P2, P3, and P4, like so:
Here, the black points and path become the first control points and linear in-
terpolation, and the green path and points become the second control points
and interpolations. Considering the original quadratically-interpolated 4-point
curve as part of a composition of normally interpolated 3-point curves allows us
to view how the curve operates. We see that we end up with 7 control points,
meaning that the final Bezier curve ends up with degree 6 (the blue function).
We can extend this to interpolations of any degree k between control points.
Ultimately, if you have n control points, p paths between those original control
points, and interpolation of degree k, we can write a rule for the degree of the
resulting Bezier curve (d):
d = ((k + 1) (p) − (n − 2)) − 1
Getting this answer is quite simple. We know that if we have a Bezier curve
with k degree interpolation, and p paths between the first control points, then
it can be split into p sections with k + 1 control points per section. With the 4-
point 2-degree interpolation Bezier curve, we see that the curve has 3 quadratic
paths, each quadratic path requiring 3 linearly-interpolated control points to
form. Thus, we would have a total of 3 ∗ 3 =9 control points. However, some of
these control points would serve as control points for two different paths, since
we’re treating the Bezier curve skeleton (the green paths) as an amalgam of
lower-degree Bezier curves. More specifically, of all the original control points
(the points directly on the green paths), only the outside two, or farthest left
and right points, are not functioning as control points simultaneously for two
different curves. Thus, we end up getting that the total control points for
the curve, if it were linearly interpolated, would be ((k + 1) (p) − (n − 2)). In
this case, it calculates out to 7 total control points, and with any linearly-
interpolated Bezier curve, that would give a curve of degree 7-1=6.
In conclusion, we see that Bezier curves can have many interesting patterns
when varying and modifying some of the “rules” of said Bezier curves, in this
case changing the control point interpolation from linear to quadratic.
8

More Related Content

What's hot

267 5 parametric eequations of lines
267 5 parametric eequations of lines267 5 parametric eequations of lines
267 5 parametric eequations of linesmath260
 
Linear transformation and application
Linear transformation and applicationLinear transformation and application
Linear transformation and applicationshreyansp
 
Linear vector space
Linear vector spaceLinear vector space
Linear vector spaceSafiya Amer
 
Question 5
Question 5Question 5
Question 5inner4zn
 
Chapter 4: Vector Spaces - Part 1/Slides By Pearson
Chapter 4: Vector Spaces - Part 1/Slides By PearsonChapter 4: Vector Spaces - Part 1/Slides By Pearson
Chapter 4: Vector Spaces - Part 1/Slides By PearsonChaimae Baroudi
 
11.1 Sequences and Series
11.1 Sequences and Series11.1 Sequences and Series
11.1 Sequences and Seriessmiller5
 
A geom ocultanasleisfisicas
A geom ocultanasleisfisicasA geom ocultanasleisfisicas
A geom ocultanasleisfisicaselysioruggeri
 
Cirlces day 1 and day 2 together
Cirlces day 1 and day 2 togetherCirlces day 1 and day 2 together
Cirlces day 1 and day 2 togetherjbianco9910
 
Solving recurrences
Solving recurrencesSolving recurrences
Solving recurrencesWaqas Akram
 

What's hot (20)

Chua's circuit
Chua's circuitChua's circuit
Chua's circuit
 
267 5 parametric eequations of lines
267 5 parametric eequations of lines267 5 parametric eequations of lines
267 5 parametric eequations of lines
 
Linear transformation and application
Linear transformation and applicationLinear transformation and application
Linear transformation and application
 
Linear vector space
Linear vector spaceLinear vector space
Linear vector space
 
Question 5
Question 5Question 5
Question 5
 
D04442327
D04442327D04442327
D04442327
 
Chapter 4: Vector Spaces - Part 1/Slides By Pearson
Chapter 4: Vector Spaces - Part 1/Slides By PearsonChapter 4: Vector Spaces - Part 1/Slides By Pearson
Chapter 4: Vector Spaces - Part 1/Slides By Pearson
 
R lecture co2_math 21-1
R lecture co2_math 21-1R lecture co2_math 21-1
R lecture co2_math 21-1
 
Character tables
Character tablesCharacter tables
Character tables
 
circles
circlescircles
circles
 
D4 trigonometrypdf
D4 trigonometrypdfD4 trigonometrypdf
D4 trigonometrypdf
 
Gaussian
GaussianGaussian
Gaussian
 
11.1 Sequences and Series
11.1 Sequences and Series11.1 Sequences and Series
11.1 Sequences and Series
 
A geom ocultanasleisfisicas
A geom ocultanasleisfisicasA geom ocultanasleisfisicas
A geom ocultanasleisfisicas
 
Cirlces day 1 and day 2 together
Cirlces day 1 and day 2 togetherCirlces day 1 and day 2 together
Cirlces day 1 and day 2 together
 
parabolas
parabolasparabolas
parabolas
 
Ford Circles
Ford CirclesFord Circles
Ford Circles
 
Week 7
Week 7Week 7
Week 7
 
Solving recurrences
Solving recurrencesSolving recurrences
Solving recurrences
 
R lecture co4_math 21-1
R lecture co4_math 21-1R lecture co4_math 21-1
R lecture co4_math 21-1
 

Viewers also liked

Zebrafish myocardial regeneration regulation
Zebrafish myocardial regeneration regulationZebrafish myocardial regeneration regulation
Zebrafish myocardial regeneration regulationNicolas Miguel
 
Majumder_B_et_al_Nature_Communications_2015
Majumder_B_et_al_Nature_Communications_2015Majumder_B_et_al_Nature_Communications_2015
Majumder_B_et_al_Nature_Communications_2015Joelle Lynn Kord
 
Christianity in east and southeast asia 2
Christianity in east and southeast asia 2Christianity in east and southeast asia 2
Christianity in east and southeast asia 2JamesMartin1216
 
Ikhsan firyanusa portefólio e currículo 02
Ikhsan firyanusa   portefólio e currículo 02Ikhsan firyanusa   portefólio e currículo 02
Ikhsan firyanusa portefólio e currículo 02Ikhsan Firyanusa
 

Viewers also liked (8)

enzyme (2)
enzyme (2)enzyme (2)
enzyme (2)
 
Zebrafish myocardial regeneration regulation
Zebrafish myocardial regeneration regulationZebrafish myocardial regeneration regulation
Zebrafish myocardial regeneration regulation
 
Modulodecontabilidadbasica
ModulodecontabilidadbasicaModulodecontabilidadbasica
Modulodecontabilidadbasica
 
Majumder_B_et_al_Nature_Communications_2015
Majumder_B_et_al_Nature_Communications_2015Majumder_B_et_al_Nature_Communications_2015
Majumder_B_et_al_Nature_Communications_2015
 
Christianity in east and southeast asia 2
Christianity in east and southeast asia 2Christianity in east and southeast asia 2
Christianity in east and southeast asia 2
 
Ikhsan firyanusa portefólio e currículo 02
Ikhsan firyanusa   portefólio e currículo 02Ikhsan firyanusa   portefólio e currículo 02
Ikhsan firyanusa portefólio e currículo 02
 
fundicion-de-oro
fundicion-de-orofundicion-de-oro
fundicion-de-oro
 
Realidad virtual
Realidad virtualRealidad virtual
Realidad virtual
 

Similar to FINAL PAPER!!!!

Transformación de coordenadas
Transformación de coordenadas Transformación de coordenadas
Transformación de coordenadas leonelfernandez33
 
Class 14 3D HermiteInterpolation.pptx
Class 14 3D HermiteInterpolation.pptxClass 14 3D HermiteInterpolation.pptx
Class 14 3D HermiteInterpolation.pptxMdSiddique20
 
NESTOR ARTEAGA GEOMETRIA ANALITICA
NESTOR ARTEAGA GEOMETRIA ANALITICANESTOR ARTEAGA GEOMETRIA ANALITICA
NESTOR ARTEAGA GEOMETRIA ANALITICANestorArteaga5
 
Arango angie geomatria 20%
Arango angie geomatria 20%Arango angie geomatria 20%
Arango angie geomatria 20%AngieArango12
 
Transformacion de coordenadas
Transformacion de coordenadasTransformacion de coordenadas
Transformacion de coordenadasEduardoJRamosDiaz
 
Transformación de coordenadas
Transformación de coordenadasTransformación de coordenadas
Transformación de coordenadasAndres32698
 
Erika vidal 20% transformacion de coordenadas
Erika vidal 20% transformacion de coordenadasErika vidal 20% transformacion de coordenadas
Erika vidal 20% transformacion de coordenadaserikavidal14
 
Section 1.3 -- The Coordinate Plane
Section 1.3 -- The Coordinate PlaneSection 1.3 -- The Coordinate Plane
Section 1.3 -- The Coordinate PlaneRob Poodiack
 
Chapter1polarcoordinatesandvector 150105021140-conversion-gate02
Chapter1polarcoordinatesandvector 150105021140-conversion-gate02Chapter1polarcoordinatesandvector 150105021140-conversion-gate02
Chapter1polarcoordinatesandvector 150105021140-conversion-gate02Cleophas Rwemera
 
Applied Calculus Chapter 1 polar coordinates and vector
Applied Calculus Chapter  1 polar coordinates and vectorApplied Calculus Chapter  1 polar coordinates and vector
Applied Calculus Chapter 1 polar coordinates and vectorJ C
 
Yuxnei mora slide share ing petrólera
Yuxnei mora slide share ing petróleraYuxnei mora slide share ing petrólera
Yuxnei mora slide share ing petrólerayuxnei mora
 
Chapter 12 vectors and the geometry of space merged
Chapter 12 vectors and the geometry of space mergedChapter 12 vectors and the geometry of space merged
Chapter 12 vectors and the geometry of space mergedEasyStudy3
 
CrespoSampleof Student Correspondence
CrespoSampleof Student CorrespondenceCrespoSampleof Student Correspondence
CrespoSampleof Student CorrespondenceIrma Crespo
 

Similar to FINAL PAPER!!!! (20)

Transformación de coordenadas
Transformación de coordenadas Transformación de coordenadas
Transformación de coordenadas
 
Class 14 3D HermiteInterpolation.pptx
Class 14 3D HermiteInterpolation.pptxClass 14 3D HermiteInterpolation.pptx
Class 14 3D HermiteInterpolation.pptx
 
Plano numerico
Plano numericoPlano numerico
Plano numerico
 
NESTOR ARTEAGA GEOMETRIA ANALITICA
NESTOR ARTEAGA GEOMETRIA ANALITICANESTOR ARTEAGA GEOMETRIA ANALITICA
NESTOR ARTEAGA GEOMETRIA ANALITICA
 
Arango angie geomatria 20%
Arango angie geomatria 20%Arango angie geomatria 20%
Arango angie geomatria 20%
 
Transformacion de coordenadas
Transformacion de coordenadasTransformacion de coordenadas
Transformacion de coordenadas
 
Transformación de coordenadas
Transformación de coordenadasTransformación de coordenadas
Transformación de coordenadas
 
Erika vidal 20% transformacion de coordenadas
Erika vidal 20% transformacion de coordenadasErika vidal 20% transformacion de coordenadas
Erika vidal 20% transformacion de coordenadas
 
Section 1.3 -- The Coordinate Plane
Section 1.3 -- The Coordinate PlaneSection 1.3 -- The Coordinate Plane
Section 1.3 -- The Coordinate Plane
 
Chapter1polarcoordinatesandvector 150105021140-conversion-gate02
Chapter1polarcoordinatesandvector 150105021140-conversion-gate02Chapter1polarcoordinatesandvector 150105021140-conversion-gate02
Chapter1polarcoordinatesandvector 150105021140-conversion-gate02
 
Applied Calculus Chapter 1 polar coordinates and vector
Applied Calculus Chapter  1 polar coordinates and vectorApplied Calculus Chapter  1 polar coordinates and vector
Applied Calculus Chapter 1 polar coordinates and vector
 
Yuxnei mora slide share ing petrólera
Yuxnei mora slide share ing petróleraYuxnei mora slide share ing petrólera
Yuxnei mora slide share ing petrólera
 
Curve tracing
Curve tracingCurve tracing
Curve tracing
 
Conic sections
Conic sectionsConic sections
Conic sections
 
Plano numerico
Plano numericoPlano numerico
Plano numerico
 
Plano numerico
Plano numericoPlano numerico
Plano numerico
 
Chapter 12 vectors and the geometry of space merged
Chapter 12 vectors and the geometry of space mergedChapter 12 vectors and the geometry of space merged
Chapter 12 vectors and the geometry of space merged
 
Planos numericos
Planos numericosPlanos numericos
Planos numericos
 
CrespoSampleof Student Correspondence
CrespoSampleof Student CorrespondenceCrespoSampleof Student Correspondence
CrespoSampleof Student Correspondence
 
6869212.ppt
6869212.ppt6869212.ppt
6869212.ppt
 

FINAL PAPER!!!!

  • 1. Bezier Curves with Quadratic Control Point Interpolations Nicolas I. Miguel June 1, 2016 1 Introduction Bezier curves are curves that have many applications, from car contour design to font recognition1 . The curves are exact sections of many parabolas and conics, and are the result of many linear interpolations/line fittings, which are correlated with the degree of the curve. Furthermore, the curves have control points that define many of the aspects of the curve, such as the degree of said curve. In addition, Bezier curves are essentially parametric equations. Throughout the last quarter, I have focused on Bezier curves that solely had linear interpolations. That is, the connections between different control points would be linear. In this paper, I will be exploring Bezier curves with quadratic interpolations, where the connections between control points are in quadratic form, instead of linear. The driving question behind this analysis is: How does the degree of the function of a Bezier curve change depending on the degree of interpolation? From my last quarter papers, we can remember that a Bezier curve with n original control points will have a degree of n − 1. This was only tested with Bezier curves that utilized linear interpolations, therefore, I don’t know if this rule will hold for Bezier curves that utilize quadratic interpolations. Therefore, the goal of this project is to find if there is a rule which allows you to find the degree of a Bezier curve depending on the number of control points and the degree of the interpolation. Figure 1: Bezier curve with linear interpolation 1Source: Wikipedia 1
  • 2. Figure 2: Bezier curve with quadratic interpolation 2 Curves with 3 control points We start by observing a Bezier curve with 3 control points. With normal linear interpolation, the skeleton of the curve (that is, the control points and the connections between those original control points) looks like this: However, we are looking at quadratic interpolation. Thus, the path between two control points will have to be quadratic. To define these parabolas, we’ll consider P1 (the original control point farthest to the left) as the vertex for a parabola which goes through P2 (the original control point to the immediate right of P1). Furthermore, we’ll consider P2 to be the vertex for a parabola which goes through P3 (the point to the right of P2). We’ll have that: P1 = (−2, 0) P2 = (0, 2) P3 = (2, 0) We consider the first interpolation, which is a parabola where P1 is the vertex, and P2 is a point the parabola goes through: y = a(x − h)2 + k y = a(x − (−2))2 + 0 y = a(x + 2)2 2 = 4a a = 1 2 y = 1 2 (x + 2)2 2
  • 3. We repeat a similar process and get that the parabola that has P2 as a vertex and goes through P3 is of the form: y = −1 2 x2 + 2. This looks like so (note that I’ve restricted the domain of both parabolas so that they stay within the bounds of the skeleton): Now we need to find the parametric equations for the two points that move along the two paths. We’ll call these points L1, which travels on the first path, and L2, which travels on the second path. Starting with L1, we see that at time t=0, L1 will be at (−2, 0), and at t=1, L1 will be at (0, 2). Thus, we find that xL1(t) = 2t − 2, since at t = 0, L1’s x-coordinate is at -2 and at t = 1 it is at 0. However, to find the y-coordinate we need to manipulate the points a bit differently. Instead of only dealing with L1 at t=0 or t=1, we also need to consider where L1 is at t=-1. This is because the y-coordinate of the point does not stay at the same rate, rather, the rate changes because the path is a parabola. Thus, we need three points to calculate the L1’s y-coordinate parametric equation. We already know that at t = 0, y=0 and at t = 1, y=2. However, since this is a parabola, we will consider that at t = −1, L1 was at the point (−4, 2), thus at t = −1, y = 2. The following graphic best details what I am doing: Thus, we see that we have three points in the form (t, y): (−1, 2), (0, 0), and (1, 2). Naturally, we observe that (0, 0) is the vertex, thus, we solve like any 3
  • 4. other parabola: y = a(t + h)2 + k y = a(t + 0)2 + 0 y = at2 2 = a(−1)2 2 = a y = 2t2 Thus, we have parametrics for L1. L1 = (2t − 2, 2t2 ) We repeat the same process for L2, proposing that at t = −1, L2 is at (−2, 0), like so: We find that xL2(t) = 2t, and after setting up three points in the form (t, y) ((−1, 0), (0, 2), and (1, 0), where (0, 2) is the vertex), we find that the y-coordinate is given by: y = a(t − h)2 + k y = at2 + 2 0 = a + 2 a = −2 y = −2t2 + 2 We have the parametrics for L2. L2 = (2t, 2 − 2t2 ) Now that we have the two secondary control points, our curve skeleton looks like this: 4
  • 5. L1 is the red point on the left path, and L2 is the red point on the right path. The next step we take is finding an equation for the parabolic path between L1 and L2. Let’s consider that L1 is the vertex, and L2 is a point on that parabola. Once again, we set up: y = a(x − h)2 + k y = a(x − xL1)2 + yL1 y = a(x − 2t + 2)2 + 2t2 yL2 = a(xL2 − 2t + 2)2 + 2t2 2 − 2t2 = a(2t − 2t + 2)2 + 2t2 y = 1 2 − t2 (x − 2t + 2)2 + 2t2 Once we find the equation of the path, we have (in red): The final step is to figure out the parametric equation for the final control point. The point, which we’ll call R, will travel along the red path, which we just figured out. At t = 0, R = (xL1, yL1). At t = 1, R = (xL2, yL2). Figuring out the parametric equation for xR is simple enough. We know that x goes from -2 to 2 in 1 second, thus: xR(t) = 4t − 2. For the parametric equation for the 5
  • 6. y-coordinate, all we have to do is plug in xR(t) for ‘x’ in the equation of the red path (y = 1 2 − t2 (x − 2t + 2)2 + 2t2 ). We get: y = 1 2 − t2 ((4t − 2) − 2t + 2)2 + 2t2 Ultimately, the parametric equations for the final control point R are: (4t − 2, 1 2 − t2 ((4t − 2) − 2t + 2)2 + 2t2 ) The skeleton is finally complete, and the curve is traced out by R (the blue point): Now we solve for the path traced out by R by solving for t in terms of x and then replacing that for the t in the parametric equation for y, like so: x = 4t − 2 t = x + 2 4 y = 1 2 − x + 2 4 2 4 x + 2 4 − 2 − 2 x + 2 4 + 2 2 + 2 x + 2 4 2 y = − 1 64 x4 − 1 8 x3 − 1 8 x2 + 1 2 x + 3 4 The function that results is of degree 4, which is the degree of a linearly- interpolated 3-point Bezier curve (2) multiplied by 2. Here we observe the graph of the Bezier curve, which is in orange: 6
  • 7. 3 Curve with 4 control points We repeat the same process as before, setting up a skeleton and solving for each set of respective control points. Ultimately, we get a path traced out by the blue point: The final path looks like this (in blue): The function of the path is of degree 6, which is the degree of the linearly- interpolated 4-point Bezier curve (3) multiplied by 2. 4 Conclusion With the two examples we have been given, we see that the final degree of the curve is the degree of the normal Bezier curve multiplied by 2, which happens 7
  • 8. to be the degree of the paths between the control points. Let’s consider the paths between the control points in the curve skeletons. These are the green paths in the diagrams, or the first set of interpolations. Consider looking at them as individual Bezier curves. To get a quadratic Bezier curve with normal, linear-interpolation, we would need 3 control points. Consider the diagram with the 4 control points, above. There are three green paths, which connect P1, P2, P3, and P4, like so: Here, the black points and path become the first control points and linear in- terpolation, and the green path and points become the second control points and interpolations. Considering the original quadratically-interpolated 4-point curve as part of a composition of normally interpolated 3-point curves allows us to view how the curve operates. We see that we end up with 7 control points, meaning that the final Bezier curve ends up with degree 6 (the blue function). We can extend this to interpolations of any degree k between control points. Ultimately, if you have n control points, p paths between those original control points, and interpolation of degree k, we can write a rule for the degree of the resulting Bezier curve (d): d = ((k + 1) (p) − (n − 2)) − 1 Getting this answer is quite simple. We know that if we have a Bezier curve with k degree interpolation, and p paths between the first control points, then it can be split into p sections with k + 1 control points per section. With the 4- point 2-degree interpolation Bezier curve, we see that the curve has 3 quadratic paths, each quadratic path requiring 3 linearly-interpolated control points to form. Thus, we would have a total of 3 ∗ 3 =9 control points. However, some of these control points would serve as control points for two different paths, since we’re treating the Bezier curve skeleton (the green paths) as an amalgam of lower-degree Bezier curves. More specifically, of all the original control points (the points directly on the green paths), only the outside two, or farthest left and right points, are not functioning as control points simultaneously for two different curves. Thus, we end up getting that the total control points for the curve, if it were linearly interpolated, would be ((k + 1) (p) − (n − 2)). In this case, it calculates out to 7 total control points, and with any linearly- interpolated Bezier curve, that would give a curve of degree 7-1=6. In conclusion, we see that Bezier curves can have many interesting patterns when varying and modifying some of the “rules” of said Bezier curves, in this case changing the control point interpolation from linear to quadratic. 8