Horizontal stretch and compression: divide the input coordinates
Quick review (2 minutes)
Imagine playing a recording at double speed.
g(x) = f(bx), b ≠ 0, sends (u, v) to (, v). |b| > 1 compresses horizontally; 0 < |b| < 1 stretches horizontally. If b < 0, also reflect across the y-axis.
1. Find b, the multiplier of the input before f receives it.
2. Set bx = u for each known old input u, because this recreates its old height.
3. Divide the old input by b, including the sign. A fractional b calls for division by a fraction.
4. Keep the old height unless an outside change is also present.
5. Move endpoints, corners, and several other points before drawing the curve.
6. Substitute the new input into bx to check that it returns u.
Move old point (4, 2) of to g(x) = .
This asks you to change the coordinate named by each mirror move and keep the other coordinate.
Calling f(2x) a horizontal stretch because 2 is bigger than 1. The multiplier is bigger, but the coordinate must be smaller: 2x = u gives x = u ÷ 2. Also, f(2x) and 2f(x) generally give different graphs.
1. Identify b and verify b ≠ 0 before using the coordinate division rule. This prevents dividing by zero.
2. Set bx = u for a known old input u. This equation finds the new address that reproduces its old height.
3. Divide both sides by b, including its sign. For u = 4 and b = −2, x = −2, and substituting back gives −2 × (−2) = 4.
4. Keep the matching old output. For old point (4, 2) of , this gives new point (−2, 2) on .
5. Move the other landmarks using the same division, and rebuild the Domain from the allowed old inputs. In , require −2x ≥ 0 to find valid inputs, then divide by −2 and reverse the comparison: x ≤ 0.
6. Check a new input inside the formula and use only known table entries. For g(x) = f(x), g(12) = f(6) = 7 is known, while g(2) needs the unavailable f(1).
Transformation
A change to a Function that moves, flips, stretches, or compresses its Graph.
Function
A rule that gives exactly one Output for every accepted Input.
Input
The value you give a Function before its rule runs, usually labeled x.
Output
The result returned by a Function, written f(x) or y.
Graph
A picture of a Function's accepted Input and Output pairs, plotted as points (x, f(x)).
Horizontal shift
Replacing x by x − h moves the Graph right h; a negative h moves it left.
Inside change
A change to the Input expression before the Function runs; it changes horizontal coordinates.
Outside change
A change applied after the Function returns its Output; it changes vertical coordinates.
Domain
All Inputs a Function accepts without an invalid calculation.
Range
All Outputs a Function can actually produce from its accepted Inputs.
Vertex
The turning point of a quadratic U, or the corner where an absolute value V changes direction.
Endpoint
A boundary point where a graph or interval starts or stops.
Horizontal reflection
Replacing the Input by its opposite, so the Graph flips across the y-axis.
Reflection
A flip across a line that places each point the same distance on the opposite side.
Horizontal stretch
Multiplying sideways distances by a factor greater than 1. In f(bx), use 0 < |b| < 1; a negative b also reflects.
Horizontal compression
Multiplying sideways distances by a factor between 0 and 1. In f(bx), use |b| > 1; a negative b also reflects.
Reciprocal
One divided by a nonzero number. Multiplying the number by its Reciprocal gives 1.
Point rule
The coordinate recipe for moving each graph point through a Transformation.
Square root function
The toolkit Function f(x) = , returning the nonnegative number whose square is x.
Axis
One of the two reference number lines on a graph. The x-axis is horizontal; the y-axis is vertical. Together they are called axes.
Coordinate
One number in a point's address. In (x, y), x measures sideways position and y measures height from the Origin.
Square root
The nonnegative number whose square equals the number inside the square root sign. A negative inside has no real Square root.
Coefficient
A fixed number multiplying a letter or expression.
Magnitude
The size of a number without its sign, measured as distance from zero.
Substitution
Replacing every occurrence of a letter by a chosen value or expression.
Set
A collection of values specified together.
Nonnegative
At least zero: zero and positive real numbers.
1 In plain words
Imagine playing a recording at double speed. Something that used to happen four minutes into the recording now happens at two minutes. The event has the same content, but you reach it sooner. In g(x) = f(2x), the old function runs through its input addresses twice as fast. Its graph therefore takes half the horizontal space. This is a Horizontal compression. With f(x), the old function runs at half speed. Events need twice as much horizontal space, a Horizontal stretch. You find a dot's new position by undoing the multiplication inside. Multiplying the input expression and multiplying the dot's coordinate are different jobs. Keeping track of which job you are doing prevents the most common scaling mistake.
2 Picture it
Reminder
Reminder: Dividing by a fraction, Solving for the new input, Flipping an inequality
- Dividing by a fraction. Divide by by multiplying by 2: 2 ÷ = 4. Check by multiplying back: × 4 = 2.
- Solving for the new input. To reach old input 4 in f(2x), set 2x = 4 to find the matching address. Divide both sides by 2 to get x = 2, then check 2 × 2 = 4.
- Flipping an inequality. A negative divisor reverses order. For , solve −2x ≥ 0 to find allowed inputs: x ≤ 0. Input −2 gives inside 4, which is allowed.
3 Why
Suppose f(u) = v. For g(x) = f(bx), the same height v appears where bx = u. Dividing both sides by b gives x = . Thus the new horizontal coordinate is the old one divided by b. If b = 2, the coordinate is halved. If b = , division by a half doubles it. A negative b also changes the coordinate's sign, reflecting left and right. The size factor is , the reciprocal of the coefficient's magnitude. You cannot divide by b = 0, so this point rule requires b ≠ 0.
4 The rule
, sends (u, v) to .
|b| > 1 compresses horizontally; 0 < |b| < 1 stretches horizontally. If b < 0, also reflect across the y-axis.
The same idea, five ways
An inside multiplier changes where an old height appears. Faster input changes squeeze the graph; slower input changes spread it out.
For g(x) = f(bx), divide each old horizontal coordinate by the nonzero coefficient b and keep the old height.
- bx = u gives x = , with b ≠ 0.
- (u, v) becomes (, v). Width factor: .
- |b| > 1: Horizontal compression. 0 < |b| < 1: Horizontal stretch.
- b < 0: also a Horizontal reflection. |b| = 1: no width change.
A recording at double speed reaches the same event in half the time.
The same idea, other ways
For , the old square root point (4, 2) appears at (2, 2). The height 2 survives, but its horizontal distance from the axis is halved.
Twice as fast means each old event takes half as long to reach. Half as fast means each event takes twice as long. The graph's width measures when events occur, so the width changes by the reciprocal speed.
As a recording
bx = u gives x = , with b ≠ 0.
To make x equal old input 6, solve x = 6. Multiply by 2 to get x = 12. A smaller inside coefficient therefore requires a larger horizontal coordinate.
With a fractional coefficient
bx = u gives x = , with b ≠ 0.
The new function must receive exactly u inside f to reproduce f(u). Writing bx = u and dividing gives x = u ÷ b. This equation forces the reciprocal factor for every point.
Why it must be true
bx = u gives x = , with b ≠ 0.
Inside multiply, coordinates divide. Check the phrase with old input 2 and b = 2: the new coordinate 1 feeds 2·1 = 2 into f. This numerical check takes only a few seconds.
A memory device
bx = u gives x = , with b ≠ 0.
| Old input u | Old height | New input for f(2x) | New input for f(x) |
|---|---|---|---|
| 2 | 1 | 1 | 4 |
| 4 | 3 | 2 | 8 |
| 6 | 7 | 3 | 12 |
| 8 | 11 | 4 | 16 |
Part 1: Horizontal compression
Squeeze the left-to-right distances toward the vertical axis. Each old height arrives closer to that axis because the inside multiplier reaches old inputs faster.
- Formula: f(bx) with |b| > 1.
- Width factor: .
- For b = 2, every old input coordinate is divided by 2.
- A compression changes distances, while a horizontal shift adds the same amount to all addresses.
Counterexamples
What it is not, and the tempting wrong moves.
Tips and tricks
- Use the displayed landmark or known column as a check before drawing any additional points.
Part 2: Horizontal stretch
Pull the addresses away from the vertical axis. A half-size input multiplier means you travel twice as far before the old machine receives the same number.
- Formula: f(bx) with 0 < |b| < 1.
- Width factor: .
- For b = , divide each old input by , which multiplies it by 2.
- The old input zero remains zero when scaling without a shift.
- A horizontal stretch by 4 uses f(x), because the inside coefficient is the reciprocal of the width factor 4.
Counterexamples
What it is not, and the tempting wrong moves.
Tips and tricks
- Use the displayed landmark or known column as a check before drawing any additional points.
Part 3: Negative horizontal scale
First think about the amount of squeezing or stretching. Then switch the address to the opposite side of the vertical axis because a negative input multiplier reverses left and right.
- The sign of b supplies a Horizontal reflection.
- The magnitude |b| supplies a reciprocal width factor.
- For g(x) = , the domain is x ≤ 0 and the range is y ≥ 0.
- Changing inputs does not by itself make positive square root outputs negative.
Worked example
Reflect and compress a square root
Move old point (4, 2) of to g(x) = .
This asks you to change the coordinate named by each mirror move and keep the other coordinate.
Show the steps and the answer
- Set .Inside the root must equal old input 4.
- Divide by to get .The division halves the distance and reverses its sign.
- Keep the old height 2.There is no minus sign outside the square root.
Counterexamples
What it is not, and the tempting wrong moves.
Tips and tricks
- Use the displayed landmark or known column as a check before drawing any additional points.
5 How
- Find b, the multiplier of the input before f receives it.
- Set bx = u for each known old input u, because this recreates its old height.
- Divide the old input by b, including the sign. A fractional b calls for division by a fraction.
- Keep the old height unless an outside change is also present.
- Move endpoints, corners, and several other points before drawing the curve.
- Substitute the new input into bx to check that it returns u.
Convert an inside coefficient into a width change
- Identify b and verify b ≠ 0 before using the coordinate division rule. This prevents dividing by zero.
- Set bx = u for a known old input u. This equation finds the new address that reproduces its old height.
- Divide both sides by b, including its sign. For u = 4 and b = −2, x = −2, and substituting back gives −2 × (−2) = 4.
- Keep the matching old output. For old point (4, 2) of , this gives new point (−2, 2) on .
- Move the other landmarks using the same division, and rebuild the Domain from the allowed old inputs. In , require −2x ≥ 0 to find valid inputs, then divide by −2 and reverse the comparison: x ≤ 0.
- Check a new input inside the formula and use only known table entries. For g(x) = f(x), g(12) = f(6) = 7 is known, while g(2) needs the unavailable f(1).
6 Worked example
Worked example
Stretching and reflecting a parabola: g(x) = f(−x)
The graph of f(x) = (x + 2 − 3, shown in the figure, is an upward-opening parabola with vertex (−2, −3). It passes through the five points (−6, 1), (−4, −2), (−2, −3), (0, −2) and (2, 1). Let g(x) = f(−x). (a) Find b and describe how the graph of f changes. Say whether it is stretched or compressed horizontally, by what factor, and whether it is reflected. (b) Find the five corresponding points on the graph of g. (c) Give the vertex and the axis of symmetry of g, and explain why one of the five points does not move.
Show the steps and the answer
- Find b. In , the input x is multiplied by before f receives it, so .b is the multiplier of the input inside f. Nothing is added inside f and nothing is done to its output, so b is the only change to the graph.
- Classify the change. lies between 0 and 1, so the graph is stretched horizontally away from the y-axis by the factor . Since b < 0, the graph is also reflected across the y-axis.For −x to equal an old input u, x must be times as far from 0 as u, which spreads the points away from the y-axis. The minus sign forces x and u to have opposite signs, so each point off the y-axis lands on the opposite side of it.
- Set for each old input. For the vertex , solve .Since g(x) = f(−x), g gives the old height −3 exactly when −x hands f the old input −2.
- Divide the old input by b, including the sign: . For every point, .Dividing by a fraction means multiplying by its reciprocal. The reciprocal of − is −, and the minus sign is kept. Multiplying u by − instead would pull the points toward the y-axis. That is a compression, the opposite of what f(−x) does.
- Keep each old height and move all five points: ; ; ; ; .Only the input changed. There is no multiplier or shift outside f, so each output v carries over unchanged.
- Read off the features of g. The vertex moves from to , so the axis of symmetry moves from to . The y-intercept stays where it is.The vertex is a point of the graph, so it follows the same rule as the others. For u = 0, x = 0 ÷ b = 0 whatever b is, so a point on the y-axis is fixed by any horizontal stretch, compression or reflection.
- Plot and (9, 1), then draw a smooth upward-opening parabola through them.Moving the vertex and points on both sides of it before drawing fixes the position and the width. The result is wider than f. The reflection has also reversed the left-to-right order: (−6, 1), the leftmost of the five points on f, became (9, 1), the rightmost on g.
- Check each new input in bx: .Each new input gives back its old input u, so g(x) = f(u) = v and every point carries the correct height.
Work to write
- |b| = < 1: horizontal stretch by a factor of ; b < 0: reflection across the y-axis
- new x = u ÷ (−) = −u, heights unchanged
- vertex of g: (3, −3); axis of symmetry: x = 3
- (0, −2) is fixed because 0 ÷ b = 0
- check: −·3 = −2, so g(3) = f(−2) = −3
Ladder: from easy to exam-hard
Try each rung before you open its steps.
Rung 1 of 4
Stretch a table without inventing values
Use the original f table shown beside this question. Give known values of g(x) = f(x). Can this table determine g(2)?
This asks you to attach each known output to its transformed input and calculate any requested new height.
Show the steps and the answer
- Read the displayed original table column by column. Each top input belongs to the output directly below it.The picture supplies the known values used in the following calculations; keep each column paired.
- Double the old input coordinates: 4, 8, 12, 16.Solve x = u by multiplying u by 2.
- Keep outputs 1, 3, 7, 11.Each new input supplies the corresponding known old input to f.
- , which is not supplied.The table contains no output at old input 1; its gaps do not imply a formula.
- Use the displayed stretched g table.
- g(2) cannot be determined from the given table.
Rung 2 of 4
A horizontal stretch by a factor of four
For this separate original point example, suppose F(3) = 7. Write a horizontal stretch by 4 and find where that known point appears. This asks you to spread each input address to four times its old value while keeping its matched height.
Show the steps and the answer
- Use inside coefficient , giving .The horizontal distance factor is the reciprocal of the inside coefficient: 1 ÷ = 4.
- Solve by multiplying both sides by 4: .This finds the new input that sends old input 3 into F. At X = 12, the inside × 12 is 3.
- Keep the known old height 7, so (3, 7) becomes (12, 7).No outside operation changes the output. The separate function name F distinguishes this illustration from the earlier supplied table.
- Horizontal stretch by 4.
- (3, 7) becomes (12, 7).
Rung 3 of 4
Recognize horizontal compression from a graph
An original graph comparison is known to have g(x) = f(bx) with b > 0 and no other changes. The old right endpoint (9, 0) corresponds to the new endpoint (3, 0). Find b and the new point corresponding to old point (6, 4).
This asks you to find the new address or height of the given graph information.
b = 3.
g(x) = f(3x).
Horizontal compression by .
(6, 4) becomes (2, 4).
Show the steps and the answer
- Set .At new input 3, the inside must reach old endpoint input 9.
- Divide by 3: .The coefficient is old input divided by its corresponding new input.
- Write , a horizontal compression by .The width factor is the reciprocal of the inside coefficient.
- Move old point (6, 4) to (2, 4).Divide its input 6 by 3 and keep its output 4.
- b = 3.
- g(x) = f(3x).
- Horizontal compression by .
- (6, 4) becomes (2, 4).
Rung 4 of 4
Build a horizontal scale from words
Write the Square root function stretched horizontally by a factor of 4. Find three graph points and its domain and range.
This asks you to write the new rule from the specified moves and check its graph landmarks.
Show the steps and the answer
- Choose .The width factor 4 is the reciprocal of the inside coefficient.
- Write .Replace the old input by one-fourth of the new input.
- Move old points (0, 0), (1, 1), (4, 2) to (0, 0), (4, 1), (16, 2).Divide each old input by , multiplying its coordinate by 4.
- Require , so . Outputs are nonnegative.The positive horizontal scale does not change the direction of the square root's allowed inputs.
- Points: (0, 0), (4, 1), (16, 2).
- Domain: [0, ∞).
- Range: [0, ∞).
Counterexamples
What it is not, and the tempting wrong moves.
Tips and tricks
- Inside multiply, coordinates divide. Put bx = old input on the exam when a memorized direction feels uncertain.
- A width factor 4 needs inside coefficient . The coefficient and the width factor multiply to 1.
- Divide by the signed b. For old input 4 and b = −2, new input is −2, which handles both the compression and the reflection.
- Keep the original heights attached to their matching inputs. A table's gaps remain unknown after the transformation.
7 The trap
Calling f(2x) a horizontal stretch because 2 is bigger than 1. The multiplier is bigger, but the coordinate must be smaller: 2x = u gives x = u ÷ 2. Also, f(2x) and 2f(x) generally give different graphs.
Problems marked OpenStax come from Precalculus 2e by Jay Abramson, published by OpenStax, under the CC BY-NC-SA 4.0 licence. Access for free at https://openstax.org/books/precalculus-2e/pages/1-introduction-to-functions.