1 Kyran O'Brien Original investigation was submitted on April 5th, 2022 This report was submitted on May 29th, 2022 Report On Transformations Of Sine Functions And Their Effect On These Functions' Amplitude And Period Introduction We were examining the function π(π₯) = π΄(π ππ(π΅(π₯ + πΆ)) + π· with regards to the effect of the constants on the graph of π relative to the graph of π¦ = π ππ π₯ and the amplitude and period of π. For a mathematical function, amplitude is πππ₯−πππ 2 , while period is the magnitude of π₯-values over which the π¦-values of a function repeat themselves. In the case of π¦ = π ππ π₯, its amplitude is 1−(−1) 2 = 1 and its period is 2π radians, because that is how many unique angles there are (one full revolution of a circle) and the sine function's input is always an angle. Figure 1: This amplitude and period is shown in the hand-drawn graph of π¦ = π ππ π₯ I made above to fully anticipate the big picture of where all the ordered pairs on its graph will transfer to with the application of π΄, π΅, πΆ, and π·. As such, points that make the scale unclear and excessively cluttered - having π₯-values with multiples of π¦-values of 2 2 π 3 and π 6 and similarly off-center - are not included. In this graph, both the π₯- and π¦- axes have the same, consistent scale. Likewise, the π¦-values for multiples of explicitly labeled. π 3 and π 6 are roughly accurate, just not 2 The role of the constant π΄ in π(π₯) = π΄(π ππ(π΅(π₯ + πΆ)) + π· Figure 2: I used Desmos to graph π for different values of π΄, ignoring the other constants. I notice π΄ multiplies every π¦-value of π¦ = π ππ π₯ by a factor of π΄, stretching the graph vertically and reflecting it in the π₯-axis for π΄ < 0 as the π¦-values switch signs in this case due to the multiplication by − 1. π΄ affects the amplitude of π¦ = π ππ π₯ by factor |π΄| (as amplitude is positive) as well since π΄(1)−π΄(−1) 2 = π΄(1−(−1)) 2 = π΄( 1−(−1) 2 ) however because the amplitude of π¦ = π ππ π₯ is 1, |π΄| is the amplitude of π: π΄πππππ‘π’ππ = |π΄| π΄ does not affect the period of π because it does not affect π₯-values (and obviously does not affect π¦-values in different ways over magnitudes greater than the period). 3 The role of the constant π΅ in π(π₯) = π΄(π ππ(π΅(π₯ + πΆ)) + π· Figure 3: I graphed π for different values of π΅, ignoring the other constants. I notice π΅ multiplies every π₯-value of π¦ = π ππ π₯ by a factor of 1 π΅ because π₯-values of 1 π΅ times the π₯-values of π¦ = π ππ π₯ are needed in π¦ = π ππ(π΅π₯) to achieve the same π¦-values as 1 π¦ = π ππ π₯: π¦ = π ππ(π΅( π΅ π₯)) = π ππ π₯. This stretches the graph horizontally and reflects it in the π¦-axis for π΅ < 0 because now π₯-values of opposite signs are required to produce the same π¦-values as π¦ = π ππ π₯. π΅ does not affect the amplitude because it does not affect π¦-values. π΅ does affect the period. Because π₯-values of 1 π΅ times the π₯-values of π¦ = π ππ π₯ are needed in π¦ = π ππ(π΅π₯) to achieve the same π¦-values, π¦-values on π¦ = π ππ(π΅π₯) repeat themselves by times the magnitude of π₯-values over which they repeat themselves on π¦ = π ππ π₯; in other words, the period of π¦ = π ππ(π΅π₯) is 1 π΅ times the period of π¦ = π ππ π₯. Precisely, ππππππ = 2π |π΅| (since period is positive (a magnitude)). This appears to fly in the face of my claim that the period of π¦ = π ππ π₯ is 2π radians since that is how many unique angles there are, because 1 π΅ 4 there are still only [0, 2π[ unique angles for π¦ = π ππ(π΅π₯). The subtlety comes from the fact that the input angle into the sine function is π΅π₯, not π₯. For this input angle, it is true that π ππ(π΅π₯) = π ππ(π΅π₯ + 2π) but since the relationship between π₯ and π ππ(π΅π₯) is being plotted rather than the relationship 2π |π΅| 2π )) π΅ between π΅π₯ and π ππ(π΅π₯), we have a period of π ππ(π΅π₯ + 2π) = π ππ(π΅π₯ + π΅( because = π ππ(π΅(π₯ + 2π π΅ )) = π ππ(π΅π₯) The varying period of π₯ arises as a result of the fact that 2π is how many unique angles there are, not in opposition to it. The role of the constant πΆ in π(π₯) = π΄(π ππ(π΅(π₯ + πΆ)) + π· Figure 4: I graphed π for different values of πΆ, ignoring the other constants. I notice πΆ shifts every π₯-value on π¦ = π ππ π₯ by − πΆ units because π₯-values − πΆ units the π₯-values of π¦ = π ππ π₯ must be substituted into π¦ = π ππ(π₯ + πΆ) to achieve the same π¦-values as π¦ = π ππ π₯: π¦ = π ππ((π₯ − πΆ) + πΆ) = π ππ π₯. πΆ does not affect the amplitude because it does not affect π¦-values. πΆ does not affect the period because all π₯-values are shifted by the same number of units, so the magnitude of π₯-values over which the π¦-values of the graph repeat themselves does not change. 5 The role of the constant π· in π(π₯) = π΄(π ππ(π΅(π₯ + πΆ)) + π· Figure 5: I graphed π for different values of π·, ignoring the other constants. I notice π· shifts every π¦-value of π¦ = π ππ π₯ by π· units due to it being added or subtracted from π ππ π₯. π· does not affect the amplitude because all π¦-values are shifted the same number of units, so the difference between any two π¦-values, including the difference between the maximum and minimum that works to define the amplitude, stays constant. π· does not affect the period because it does not affect π₯-values. Putting everything together Even though I examined the effect of each of the constants in isolation, leaving the other constants in their trivial cases (π΄ = 1, π΅ = 1, πΆ = 0, π· = 0), my reasoning holds even when all of them are in play non-trivially at the same time in the function π(π₯) = π΄(π ππ(π΅(π₯ + πΆ)) + π·. This is because they all affect π in different ways and, as laid out in this form, do not act upon each other. Based on this form, I can also deduce that stretches, signified by π΄ and π΅, are always applied before shifts, signified by πΆ and π·. In the case of vertical transformations, this is clear; however, it is less obvious for horizontal transformations as the parentheses (π΅(π₯ + πΆ)) seem to suggest the horizontal shift is to be done first. The reason this is not the case is because a horizontal 6 1 shift followed by a horizontal stretch - ( π΅ (π₯ − πΆ)) - substituted into π¦ = π ππ(π΅(π₯ + πΆ)) does not achieve the same π¦-values as π¦ = π ππ π₯: 1 π₯ π¦ = π ππ(π΅(( π΅ (π₯ − πΆ)) + πΆ)) = π ππ(π₯ − πΆ + πΆπ΅) ≠ π ππ π₯ = π ππ(π΅(( π΅ − πΆ) + πΆ) π₯ By contrast, the substitution of ( π΅ − πΆ) - stretch then shift - does. Therefore in summary the constants π΄, π΅, πΆ, and π· transform the function π(π₯) = π΄(π ππ(π΅(π₯ + πΆ)) + π· from π¦ = π ππ π₯ in the following way, in the following order: Horizontal transformations: 1) A horizontal stretch by factor 1 π΅ 2) A horizontal shift by − πΆ units Vertical transformations: 1) A vertical stretch by factor π΄ 2) A vertical shift by π· units I checked these predictions for a variety of functions and their graphs matched, suggesting they are correct. Figure 6: Graph of π(π₯) = 1 2 π ππ(π₯ − π 4 ) + 2 alongside π¦ = π ππ π₯. I used the graph of π¦ = π ππ π₯ to inform my sketch of π based on the aforementioned transformations as well as checking exact points as a result of the transformations. The transformations were 7 1) A horizontal shift by + π 4 units 2) A vertical stretch by factor 1 2 3) A vertical shift by + 2 units Figure 7: My graph agreed with what I graphed on my graphing calculator. Figure 8: Graph of π(π₯) = 2π ππ(3(π₯ + π 3 )) + 1 alongside π¦ = π ππ π₯ and π¦ = 2π ππ(3π₯) + 1. Here, due to the number of different transformations, I graphed π¦ = 2π ππ(3π₯) + 1 as a dashed line before applying the horizontal shift as an intermediary step. The transformations were 1) A horizontal stretch by factor 1 3 8 2) A horizontal shift by − π 3 units 3) A vertical stretch by factor 2 4) A vertical shift by + 1 units Figure 9: My graph also matched that on my calculator. 1 Figure 10: Graph of β(π₯) =− π ππ( 2 π₯) − 3 alongside π¦ = π ππ π₯. The transformations were 1) A horizontal stretch by factor 2 2) A vertical stretch by factor − 1 3) A vertical shift by − 3 units 9 Figure 11: Again, my drawing agreed with what I graphed on my calculator.