Skip to content

Chapter 5 — Trigonometric Functions of Real Numbers

5.5 - More Trigonometric Graphs

§2.5 showed how to shift, reflect and stretch any function of the prototype catalog without redoing its analysis from scratch. The trigonometric functions admit the same transformations —and one more, proper to their periodicity, that no function of Chapter 2 needed.


The vertical shift —sin⁡t+D\sin t+D moves the graph DD units upward— and the reflection about the horizontal axis —−sin⁡t-\sin t inverts each value— are particular cases of the general transformations already proved in §2.5, applied here without any adaptation: neither the periodicity nor the bounded range of sin⁡\sin and cos⁡\cos alters those proofs in any way. They are not repeated in this section.

What §2.5 could not anticipate is what happens when stretching a periodic function horizontally: for the functions of that catalog —powers, roots— stretching horizontally changed the shape of the curve, but none of them had a period that this stretching could alter. With the trigonometric functions, it does.


The vertical stretch is the simplest to treat: it does not interact with periodicity at all. It is the content of §2.

Proposition. For A>0A>0, Rec⁡(Asin⁡t)=[−A,A]\operatorname{Rec}(A\sin t)=[-A,A].

Proof. Since Rec⁡(sin⁡)=[−1,1]\operatorname{Rec}(\sin)=[-1,1] —§5.3 §3—, multiplying by A>0A>0 preserves order and scales the interval: sin⁡t∈[−1,1]  ⟺  Asin⁡t∈[−A,A]\sin t\in[-1,1] \iff A\sin t\in[-A,A], and since sin⁡\sin attains every value of [−1,1][-1,1], Asin⁡tA\sin t attains every value of [−A,A][-A,A]. ■\blacksquare

Definition. The number ∣A∣|A| is called the amplitude of Asin⁡tA\sin t —or of Acos⁡tA\cos t—: half the distance between the maximum and the minimum value that the function attains.


The horizontal stretch, on the other hand, does interact with periodicity —and in a way that requires a proof of its own, not an analogy with §2.5—. It is the content of §3.

Proposition. For B>0B>0, T:=2π/BT:=2\pi/B is the minimal positive period of f(t):=sin⁡(Bt)f(t):=\sin(Bt).

Proof. That TT is a period: f(t+T)=sin⁡(Bt+BT)=sin⁡(Bt+2π)=sin⁡(Bt)=f(t)f(t+T)=\sin(Bt+BT)=\sin(Bt+2\pi)=\sin(Bt)=f(t), using the periodicity of sin⁡\sin —§5.3 §4—.

That it is the minimum: suppose that T′∈(0,T)T'\in(0,T) is also a period, that is, sin⁡(Bt+BT′)=sin⁡(Bt)\sin(Bt+BT')=\sin(Bt) for all tt. In particular, at t=0t=0: sin⁡(BT′)=sin⁡(0)=0\sin(BT')=\sin(0)=0. Since BT′∈(0,BT)=(0,2π)BT'\in(0,BT)=(0,2\pi), and by §5.3 §7 the only zeros of sin⁡\sin in (0,2π)(0,2\pi) are at t=πt=\pi, it follows that BT′=πBT'=\pi, that is, T′=π/B=T/2T'=\pi/B=T/2. But then, evaluating the periodicity condition at t=0t=0 is no longer enough: it must hold for all tt, and in particular for t=0t=0 it was already used; test at t=π/(2B)t=\pi/(2B), where sin⁡(Bt)=sin⁡(π/2)=1\sin(Bt)=\sin(\pi/2)=1: sin⁡ ⁣(Bt+BT′)=sin⁡ ⁣(π2+π)=−1\sin\!\left(Bt+BT'\right)=\sin\!\left(\frac\pi2+\pi\right)=-1 —by the central symmetry of §5.4 §2, sin⁡(u+π)=−sin⁡u\sin(u+\pi)=-\sin u—, which does not coincide with sin⁡(Bt)=1\sin(Bt)=1. Hence T′=T/2T'=T/2 does not satisfy the periodicity condition for all tt, a contradiction. No candidate remains in (0,T)(0,T), so that TT is the minimal period. ■\blacksquare

Definition. The number 2π/B2\pi/B is called the period of sin⁡(Bt)\sin(Bt) —or of cos⁡(Bt)\cos(Bt)—.


The horizontal shift —moving the graph to the left or to the right— is, on the other hand, exactly the general case already covered by §2.5, without any additional subtlety. It is the content of §4.

Definition. In y=sin⁡(t−C)y=\sin(t-C), the number CC is called the phase shift: the graph of sin⁡t\sin t shifted CC units horizontally —to the right if C>0C>0—, exactly according to the general transformation of §2.5.

No proof of its own is needed: it is the same horizontal transformation f(t−C)f(t-C) of §2.5, with f=sin⁡f=\sin.


Combining the four transformations —amplitude, period, phase, vertical shift— in a single expression, one obtains the general form that appears in almost any real application of these functions. It is the content of §5.

y=Asin⁡(B(t−C))+D(A>0, B>0)y=A\sin\big(B(t-C)\big)+D \qquad (A>0,\ B>0)

has amplitude AA —§2—, period 2π/B2\pi/B —§3—, phase shift CC —§4—, and vertical shift DD —§2.5—. Each parameter affects the graph independently of the other three: modifying one does not alter what the others already determined.

Worked example. Identify the amplitude, period, phase shift and vertical shift of y=3sin⁡(2(t−π/4))+1y=3\sin\big(2(t-\pi/4)\big)+1, and graph it.

Solution. Comparing with the general form: A=3A=3, B=2B=2, C=π/4C=\pi/4, D=1D=1. Amplitude 33: the graph oscillates between D−A=−2D-A=-2 and D+A=4D+A=4. Period 2π/B=2π/2=π2\pi/B=2\pi/2=\pi: each complete oscillation occupies an interval of length π\pi, half that of sin⁡t\sin t. Phase shift π/4\pi/4: the graph of 3sin⁡(2t)+13\sin(2t)+1 moved π/4\pi/4 units to the right.

Graph of y=3sin(2(t-π/4))+1 compared with sin(t) as a reference, showing amplitude 3 and period π Figure 5.4 — y=3sin⁡(2(t−π/4))+1y=3\sin\big(2(t-\pi/4)\big)+1 (solid curve), compared with sin⁡t\sin t (dotted curve). The amplitude triples the height of the oscillation; the period is halved; the whole curve ends up shifted vertically one unit and horizontally π/4\pi/4.


The general form of this section is, precisely, the one that appears when modeling any phenomenon that oscillates with regularity. A brief example closes the section before the exercises. It is the content of §6.

The depth of the water in a port, measured in meters above a reference level, varies with the tide in an approximately periodic way, with a complete cycle every 1212 hours, a maximum depth of 55 meters, a minimum of 11 meter, and the high tide occurring at 33 in the morning.

Model. Amplitude: A=(5−1)/2=2A=(5-1)/2=2. Vertical shift: D=(5+1)/2=3D=(5+1)/2=3 —the midpoint between maximum and minimum—. Period 1212 hours: by §3, B=2π/12=π/6B=2\pi/12=\pi/6. Phase: the maximum of Asin⁡(B(t−C))+DA\sin(B(t-C))+D occurs at t=C+π/(2B)t=C+\pi/(2B) —where the sine equals 11—; equating to 33 hours and solving, with π/(2B)=π/(2⋅π/6)=3\pi/(2B)=\pi/(2\cdot\pi/6)=3, one obtains C=0C=0. The resulting model:

h(t)=2sin⁡ ⁣(π6t)+3h(t)=2\sin\!\left(\frac\pi6 t\right)+3

with tt in hours since midnight.


Exercise 1. Give the period of y=cos⁡(4t)y=\cos(4t).

Solution. By §3 —applicable to cos⁡\cos by the same argument, replacing sin⁡\sin with cos⁡\cos—, the period is 2π/4=π/22\pi/4=\pi/2.

Exercise 2. Give the amplitude, the period and the vertical shift of y=−5sin⁡(t)+2y=-5\sin(t)+2.

Solution. A=5A=5 —the amplitude is ∣−5∣|{-5}|—, period 2π/1=2π2\pi/1=2\pi —unchanged, since B=1B=1—, vertical shift D=2D=2. The negative sign of AA does not change the amplitude, but it does invert the graph —reflects it about y=Dy=D—, according to the reflection transformation of §2.5.

Exercise 3. At which instants, according to the model of §6, is the depth of the water exactly 33 meters?

Solution. We seek tt with 2sin⁡(πt/6)+3=32\sin(\pi t/6)+3=3, that is, sin⁡(πt/6)=0\sin(\pi t/6)=0. By §5.3 §7, this occurs when πt/6=nπ\pi t/6=n\pi, that is, t=6nt=6n hours: at midnight, at 66, at 1212, at 1818, and so on every 66 hours —half the period, the instants at which the tide crosses its mean level—.


This chapter traveled, in five sections, a path that no elementary trigonometry course travels completely: constructing π\pi instead of taking it as known, constructing the wrapping function instead of assuming that “an angle determines a point on the circle”, and deriving each property of sine and cosine —Pythagorean identity, periodicity, parity, signs, asymptotes of their reciprocals, even the period of the tangent— from that single geometric construction, without appealing at any moment to a separate right triangle or to a memorized table of values.

The result is a family of functions as rigorously grounded as the exponential and the logarithm of Chapter 4, and with the same kind of honest limit: just as this book could not say in what precise sense the rate of change of exe^x is equal to itself —a question reserved for Calculus I—, neither could it say here how fast sin⁡t\sin t changes at each instant, nor why that rate turns out to be, precisely, cos⁡t\cos t. That connection —the heart of why the trigonometric functions are indispensable in the physics of oscillations, not only in the geometry of the triangle— will remain pending, with the same honesty, until the book reaches the derivative.


With this, Chapter 5 is closed. Chapter 6 takes up the trigonometric functions again to solve, in the inverse direction, equations and identities more general than those that this introduction could cover.