GlacierFilm, LLC. · YouTube · 7:03
This 120-second animated lesson introduces uncertainty notation and one-variable error propagation. It first defines absolute uncertainty using a velocity example, , with , then defines relative uncertainty as , emphasizing that the former has units while the latter is unitless. After a brief terminology note that “error” and “uncertainty” are being used almost interchangeably, the video develops the general rule for a function by drawing a tangent line and stating . It then applies this rule to kinetic energy, , with exact mass and measured velocity , obtaining . The segment ends during the numerical setup; no final computed value of or is shown within these 120 seconds. This 120-second clip teaches uncertainty propagation through two linked cases. It first computes kinetic energy from a velocity measurement with uncertainty using δK ≈ (dK/dv)δv, obtaining . It then generalizes to two uncertain inputs, introducing the quadrature formula δz = sqrt((∂)^2 + (∂z/∂y δy)^2). The middle section explains partial derivatives both geometrically, as slopes along x and y on a tangent plane to a surface , and procedurally, by differentiating one variable while holding the others constant. The final section motivates quadrature with a right-triangle picture: opposing errors can cancel, aligned errors give the worst-case sum , and independent random errors are modeled at 90 degrees, giving sqrt(), which is less than the fully additive worst case. This 120-second segment teaches error propagation through a geometric introduction, one fully worked kinetic-energy example, and three reusable shortcut rules. It first shows that combining two uncertainty components in quadrature gives a value between the largest single component and their direct sum. It then applies the two-variable formula to with +/- 0.11 kg and +/- , obtaining . The remainder of the clip catalogs common patterns: multiplication by a constant scales absolute uncertainty, addition and subtraction combine absolute uncertainties in quadrature, and multiplication and division combine relative uncertainties in quadrature. This 63-second clip teaches basic error propagation through the rule that independent uncertainties add in quadrature. It first shows the product and quotient cases, especially the relative-uncertainty formula . A numerical animation with 0.03% and 5.4% illustrates when the smaller relative uncertainty may be neglected, leading to the simplified example \, for z=xy. The clip then uses 2D and 3D right-triangle vector pictures to explain quadrature geometrically and extends the formulas to three variables and to general differentiable functions via partial derivatives. The last portion is a joke pie chart and credits.
Use the learning inspector for key ideas and moments, or open the reading tabs for the complete notes.
Generated from the video's visuals and explanation; not verbatim speech.
The clip opens with the title “Propagation of errors” and immediately foregrounds the symbol , signaling that the lesson will use uncertainty notation rather than only plain English description.
A radar-gun example gives a measured velocity with a plus-or-minus amount . The narration defines this plus-or-minus amount, when written in the same units as the measurement, as an absolute uncertainty, denoted .
The same numbers are then rearranged into a ratio: . This is identified as the relative uncertainty , and the key distinction is that carries units while is dimensionless.
A short terminology interlude personifies “error” and “uncertainty.” Mathematically, nothing new is added here; the point is that the clip will not dwell on a strict lexical distinction between the two words.
The lesson then generalizes from the concrete example. For a measured input with uncertainty , and a computed output , the graph shows a point on the curve and a tangent line at that point. The visual argument is that a small horizontal shift produces a vertical shift proportional to the local slope, leading to the rule .
Finally, the rule is applied to kinetic energy. The problem states a mass with velocity and asks for the kinetic energy. With treated as exact, the function is , so the propagated uncertainty follows as . The clip ends while substituting numbers into the setup, before any final numerical result is displayed.
The clip opens with a concrete physics-calculus example. The screen states that a 10-kg mass has measured velocity , and it computes the kinetic energy from mv^2. Substitution gives 27.378 J, while the graph of K versus v with a red tangent line signals that the local slope will control the uncertainty.
The next step propagates the velocity uncertainty. The displayed relation is δK = dK/dv δv = (mv)δv. Plugging in kg, , and yields 0.94 J, and the final boxed answer rounds this to . This establishes the one-variable rule: multiply the input uncertainty by the derivative of the output with respect to that input.
The video then raises the difficulty by showing two separately measured quantities, velocity and mass kg, both feeding into . The key conceptual shift is that the output now depends on two uncertain inputs, so the earlier single-derivative formula is no longer sufficient by itself.
To handle that case, the screen rewrites the quantity as a generic two-variable function, , and introduces the quadrature formula δz = sqrt((∂)^2 + (∂z/∂y δy)^2). The narrator names this procedure adding the uncertainties in quadrature. The formula says that each input contributes a first-order term given by its partial derivative times its uncertainty, and those contributions are combined in root-sum-square form.
Before using the formula, the clip pauses to explain the notation. The curly d symbol ∂ is identified as a partial derivative, meaning differentiation with respect to one variable at a time. The visual isolation of ∂/∂x, ∂/∂y, and ∂/∂z prepares the viewer to read the quadrature expression correctly rather than treating the symbols as decorative.
The geometric meaning is then shown on a 3D surface . Whereas an ordinary derivative is the slope of a tangent line to a curve, a two-variable function has a tangent plane. On that plane, ∂f/∂x is the slope in the x-direction and ∂f/∂y is the slope in the y-direction. This connects the algebraic partial derivatives in the uncertainty formula to visible directional steepness on a surface.
The video next turns from interpretation to computation. Using xy^2z+√, it demonstrates the practical rule: choose one variable, treat all others as constants, and differentiate only the terms that contain the chosen variable. The displayed results are ∂f/∂, ∂xyz+1/(2√y), and ∂xy^. The message is that partial differentiation uses familiar single-variable rules once the other variables are frozen.
The final section answers why independent uncertainties combine by quadrature. The animation compares three situations for two error contributions and : opposing directions, which allow cancellation; same direction, which gives the worst-case total ; and perpendicular directions, which produce a right triangle with hypotenuse sqrt(). The narrator frames the perpendicular case as the average relationship for random independent effects.
The clip closes by emphasizing the inequality shown on screen: hypotenuse < . Thus the quadrature sum is smaller than the fully additive worst case, while still exceeding either single contribution. This geometric argument motivates the earlier formula and links probability-style independence assumptions to the root-sum-square rule used in uncertainty propagation.
The clip opens with a right triangle whose legs are labeled and and whose hypotenuse is labeled . The accompanying inequality states that this hypotenuse is less than but greater than either or alone, giving a geometric picture of quadrature combination.
The lesson then returns to a concrete application: kinetic energy with two measured inputs. The formula shown is , and the problem gives and .
To propagate uncertainty, the video writes . The two partial derivatives are evaluated as and .
Substituting the numerical data gives , and the displayed final result is .
After this worked example, the video shifts to reusable patterns under the heading Tricks and Shortcuts (for common examples).
The first shortcut is multiplication by a constant. Starting from , the slide derives . A numerical illustration states that if , then .
The second shortcut covers addition and subtraction. For , the general quadrature formula is specialized using and . Because those signs are squared, the result simplifies to .
The third shortcut treats multiplication and division together. For , the raw propagated form is ; for , it is . Dividing by and simplifying both cases yields the same relative-uncertainty rule: .
The clip opens on a white summary slide headed MULTIPLICATION AND DIVISION:. Two cases are displayed side by side: =xy on the left and on the right. Under each, the absolute uncertainty is written as , and below that the relative form is shown as . The sentence at the bottom states the governing idea: the relative uncertainties add in quadrature.
The view then isolates the relative-uncertainty formula. Two annotations appear above the terms under the radical: : very accurate measurement points to , and : sloppy measurement points to . The narrator explains that when one measurement is much more accurate, relatively speaking, the smaller relative uncertainty can sometimes be ignored.
As that explanation continues, the animation makes the approximation visible: the term shrinks away, while the term remains and is emphasized. Mathematically, this corresponds to replacing by approximately , hence .
A new line appears: Let's say... . Starting from the simplified relation , the denominator is rewritten as , giving . The common factor cancels, and the screen ends at . A speech bubble comments, Wow, it really does behave as if "x" were constant!, while the narration warns that this shortcut saves math but should be used carefully.
The lesson then shifts to a geometric picture. In a 2D coordinate-style drawing, a blue vector and a red vector meet at right angles, and the diagonal is labeled . This visualizes quadrature addition as a Pythagorean construction for two independent uncertainty components.
The figure rotates into three dimensions. A green vector is added perpendicular to the first two, and the new space diagonal is labeled . The narrator asks what happens when combining three or more measurements and answers that the same idea extends: the hypotenuse is the square root of all three uncertainties squared.
A summary board gathers the rules into three families. For Addition and subtraction: . For Multiplication and division: . For Everything else! (Squared things, square roots, logarithms, etc...): . Each formula is then extended by a third term involving , matching the narration’s instruction to add three things in quadrature rather than just two.
The mathematical content ends with a humorous pie chart titled Time arguing with reserach collaborators:. A small green slice is labeled About the results and a much larger yellow slice is labeled About the uncertainties!. After that, an animated door labeled Physics Lab appears, followed by credits naming Katherine Rawlins and Paul Andrew Lawrence.
Absolute uncertainty is the plus-or-minus amount attached to a measurement and written in the same units as the measured quantity. In the opening example, the measured velocity is and the absolute uncertainty is , denoted . The symbol is introduced specifically to mean “absolute uncertainty in …”.
Relative uncertainty expresses the size of the uncertainty as a fraction or percentage of the measured value. It is computed by dividing the absolute uncertainty by the measurement itself. For the example, , and this is written as . Unlike , the relative uncertainty is unitless.
The clip briefly notes that people often say “error” even though “uncertainty” is more descriptive. The narrator explicitly tells viewers not to worry about the difference. This card records terminology only; no separate mathematical definition of “error” is developed in the segment.
For a quantity computed from a measurement through a function , the video estimates the output uncertainty from the local slope. The graph draws a tangent line at the point corresponding to the measured , and the rule shown is . Conceptually, a small input uncertainty is converted into an output uncertainty by multiplying by the derivative at that point.
The worked example takes for a mass with velocity . Because the mass is treated as exact, only contributes uncertainty. Applying the general rule gives , and since , the propagated uncertainty becomes . The segment ends during numerical substitution, before a final value is shown.
The starting physical relation is mv^2. In the first example, kg is treated as exact and is the measured velocity, giving before rounding.
When only velocity is uncertain, the uncertainty in kinetic energy is estimated by multiplying the velocity uncertainty by the derivative dK/dv. Since dK/dv=mv, the rule becomes δK≈mv δv.
Substituting kg, , and gives . The final reported kinetic energy is rounded to .
The more complicated case has both velocity and mass measured with uncertainty. Because K depends on two inputs, the single-derivative rule must be replaced by a two-variable propagation formula.
For a two-variable quantity , the video combines the separate first-order contributions using the square root of the sum of squares. This is called adding the uncertainties in quadrature.
The symbol ∂ denotes a partial derivative: the derivative with respect to one variable at a time, while the other variables are held fixed.
For a surface , the ordinary tangent-line slope is replaced by a tangent plane. The partial derivative ∂f/∂x is the slope along the x-direction, and ∂f/∂y is the slope along the y-direction.
To calculate a partial derivative, differentiate with respect to one variable and treat all other variables as constants. Terms not containing the chosen variable drop out.
For xy^2z+√, the displayed derivatives are ∂, ∂f/∂xyz+1/(2√y), and ∂f/∂xy^.
If two error effects are random and independent, they may partly cancel or partly reinforce. The video models the average relationship by placing the contributions at right angles, producing a root-sum-square combination rather than the worst-case arithmetic sum.
With perpendicular contributions and , the combined size is the hypotenuse sqrt(). The screen explicitly notes that this is less than , matching the idea that quadrature is smaller than the worst-case additive error.
For two positive uncertainty components and , the quadrature combination is larger than either component alone but smaller than the arithmetic sum . The opening right-triangle picture makes this ordering visual.
When a derived quantity depends on two measured variables, the video uses a quadrature formula built from partial derivatives. Each input contributes its uncertainty multiplied by the corresponding sensitivity coefficient.
Using with and , the partial derivatives are and . Substitution gives the displayed propagated uncertainty . The clip does not compute the central value of itself.
If with an exact constant, then , so the absolute uncertainty scales in the same way: . The example shows becoming .
For , the partial derivatives are . Since the propagation formula squares those factors, the sign choice disappears and both addition and subtraction give the same rule: absolute uncertainties combine in quadrature.
For both and , the raw derivative-based expressions simplify after division by into a common rule for relative uncertainties. This is the main contrast with addition and subtraction: multiplication and division are handled most cleanly in relative form.
Independent uncertainty contributions are combined by taking the square root of the sum of their squares. The clip states this explicitly for relative uncertainties in multiplication and division, then reuses the same pattern for three components and for general differentiable functions.
For both and , the video presents the same relative-uncertainty propagation rule: the relative uncertainty in equals the quadrature sum of the relative uncertainties in and .
Explore conditions, steps and evidence. Supplementary explanations are labeled separately from content shown in the video.
The narrator says, "Propagation of errors."
Large title text reads "PROPAGATION of ERRORS".
Title of the lesson topic.
The narrator explains that a measurement should have an uncertainty plus or minus something.
The screen shows "measured velocity" above "v =" and the value "" circled.
v
Measured velocity in the introductory example.
The expression " +/- " is shown, then "absolute uncertainty in v" labels " =" with "" circled.
The narrator says the plus-or-minus quantity in the same units is called an absolute uncertainty and uses lowercase Greek letter delta.
Absolute uncertainty in the measured velocity example.
The narrator says, "we'll use the symbol lowercase Greek letter delta to write it mathematically."
The screen displays '"" = "absolute uncertainty in..."'.
Notation meaning absolute uncertainty in the following quantity.
Text reads "(Other symbols you may see: , )".
The video does not define or in this clip.
,
Alternative uncertainty-related symbols that may be seen elsewhere.
The narrator says an alternative way of expressing uncertainty is as a percentage or fraction of the thing measured, equal to delta something over the something.
The displayed calculation is " % = ".
Red label reads "relative uncertainty in v =" and circles the percentage/fraction expression.
Relative uncertainty in the velocity example.
The narrator says, "So you've measured a quantity, let's call it x, and its uncertainty in the lab."
A horizontal axis is labeled "x" with a marked point "x" and nearby interval labels "".
x
Generic measured input quantity.
The narrator says, "If x might be off by an amount delta x..."
Small intervals around the point x are labeled "".
Absolute uncertainty or offset in the generic input x.
The narrator says one wants to calculate something else, "let's call it z, using some formula z of x."
A vertical axis is labeled "" and a curve is drawn; a point on the curve corresponds to "z".
Generic output quantity computed from x.
The narrator asks, "then by how much might z be off? Delta z."
Vertical uncertainty intervals around z are labeled "".
Resulting absolute uncertainty or offset in z.
The narrator says the movement in z can be figured out based on the slope of the function.
The displayed equation is "".
A red tangent line is drawn at the point on the curve.
Derivative or slope of at the point x, used to relate small offsets.
The narrator says, "you want to calculate the kinetic energy K, which is one half m v squared."
The screen shows "".
Kinetic energy as a function of velocity in the worked example.
Narrator: every measurement should have an uncertainty plus or minus something; if specified in the same units, it is called an absolute uncertainty.
Example shows " +/- " and labels "absolute uncertainty in v" as "".
An absolute uncertainty is the plus-or-minus amount attached to a measured quantity and written in the same units as that quantity. In the clip’s velocity example, the measured velocity is and the absolute uncertainty is , denoted .
Applies to a measured quantity.
The uncertainty is expressed in the same units as the measured quantity.
The symbol is introduced as notation for absolute uncertainty in a quantity.
Narrator: an alternative way of expressing uncertainty is as a percentage or fraction of the thing measured; this is the same as delta something over the something and is called relative uncertainty.
Displayed calculation: " % = ".
A relative uncertainty expresses uncertainty as a fraction or percentage of the measured value. It is obtained by dividing the absolute uncertainty by the measured quantity, so it has no units. The clip computes , identified as .
Requires a nonzero measured quantity in the denominator.
Produces a dimensionless fraction or percentage.
Equivalent to absolute uncertainty divided by the measured value.
Narrator: "Some say error when they're in a hurry and uncertainty has too many syllables. Don't sweat the difference."
Speech bubbles labeled "error" and "uncertainty" appear with thought bubbles about ease of saying versus descriptiveness.
In this clip, “error” and “uncertainty” are presented as closely related informal terminology for the same measurement idea. The narrator explicitly downplays the distinction, while the visual thought bubbles suggest “error” is easier to say and “uncertainty” is more descriptive of what is meant.
Terminological remark only.
No separate mathematical definition of “error” is given in this clip.
Narrator explains that if x might be off by delta x, the amount z might be off depends on the function and can be found from the slope.
Displayed equation: "".
Graph of with tangent line and uncertainty intervals.
For a computed quantity , a small uncertainty in the input is estimated by multiplying the input uncertainty by the slope of the function at that point. Visually, the curve is replaced near the point by its tangent line, giving ; the clip writes this as an equality for the linear estimate.
Used when the uncertainty in is small enough for a local linear approximation.
The derivative is evaluated at the measured value of .
The video presents the relation through the tangent-line picture.
Narrator: kinetic energy K is one half m v squared.
Screen shows "".
The example uses the standard kinetic-energy formula , treating as a function of velocity with mass held fixed. This provides the specific function to which the general propagation rule is applied.
is treated as constant in this example.
The clip states is an exact number.
is the measured quantity carrying uncertainty.
Narrator: there is an uncertainty on v, delta v; the uncertainty on K is the derivative of K with respect to v times the uncertainty on v.
Displayed equations: "" and then "= (mv)".
Applying the general rule to gives . Since , the clip rewrites the propagated uncertainty as . This is the specific one-variable propagation formula used in the example.
Only has uncertainty in the stated example.
is exact/constant.
Uses the derivative of with respect to .
The screen displays .
It substitutes kg and to obtain 27.378 J.
The video uses the standard kinetic-energy expression mv^2 to compute the nominal kinetic energy from a measured mass and velocity. In the first example, the mass is treated as exact at 10 kg while the velocity carries the uncertainty.
m is the mass
v is the speed or velocity magnitude used in the kinetic-energy formula
The displayed relation is δK = dK/dv δv = (mv)δv.
The narrator says the goal is to compute not only K but also the uncertainty on K.
A tangent line is drawn on the K-versus-v curve to motivate the derivative as a local slope.
When only one input variable is uncertain, the video approximates the resulting uncertainty by multiplying the input uncertainty by the derivative of the output with respect to that input. For kinetic energy, dK/dv=mv, so δK≈mv δv.
Only one input variable carries uncertainty in this step
The function is locally differentiable
The uncertainty is small enough for a linear approximation
The screen shows δz = sqrt((∂z/∂x δx)^2 + (∂z/∂y δy)^2).
The narrator names this procedure adding the uncertainties in quadrature.
For a quantity z that depends on two independently uncertain inputs x and y, the video combines the separate first-order contributions by taking the square root of the sum of their squares. Each contribution is the corresponding partial derivative times that input's uncertainty.
z is treated as a function of two variables x and y
The two uncertainty contributions are combined by quadrature
The formula shown is a first-order propagation rule
The narrator says the curly d symbol is called a partial derivative.
The narrator defines it as a derivative with respect to one variable at a time.
The screen isolates ∂ and then shows ∂/∂x, ∂/∂y, and ∂/∂z.
A partial derivative measures how a multivariable function changes when one chosen variable changes while the other variables are held fixed. The video introduces the notation ∂ for this one-variable-at-a-time derivative.
The function has more than one independent variable
All variables except the differentiation variable are treated as constants during the partial derivative
The narrator contrasts the ordinary derivative as the slope of a tangent line with the two-dimensional case having a tangent plane.
A 3D surface is shown with a highlighted planar slice.
The screen labels ∂f/∂x as slope here along x and ∂f/∂y as slope here along y.
In two dimensions, the video interprets partial derivatives as directional slopes on a tangent plane to the surface . The partial with respect to x is the slope along the x-direction, and the partial with respect to y is the slope along the y-direction.
The surface is locally smooth enough to have a tangent plane
The slopes are taken along coordinate directions
The narrator says to look at one variable at a time while treating all the others as if they were constant.
The example crosses out terms not containing the chosen variable and differentiates the remaining terms.
To compute a partial derivative, differentiate only with respect to the selected variable and treat every other variable as a constant. Terms that do not contain the chosen variable disappear because their derivative with respect to that variable is zero.
The function is expressed in terms of several variables
Standard single-variable differentiation rules apply to the remaining factors
Narrator: "So delta something is an absolute uncertainty, it has units. The relative uncertainty is a fraction or a percentage, unitless."
Blue circle marks with units; red circle marks as a percentage.
An absolute uncertainty has the same units as the measured quantity, whereas the relative uncertainty is a fraction or percentage and is unitless.
is a measured quantity with nonzero value when forming .
is expressed in the same units as .
For the measured quantity and uncertainty shown in the clip’s examples.
Narrator says the amount moved up in z is based on the slope of the function, then states the offset in z is the derivative times the offset in x.
Displayed equation: "".
Red tangent line drawn at the point on the curve.
The clip visually motivates a local linear approximation but writes the result as an equality without explicitly stating higher-order neglected terms.
If and is offset by , then the corresponding offset in is given in the clip by , interpreted from the tangent-line slope at the point .
There is a differentiable function .
The discussion concerns a small offset around the point .
The slope is taken at that point.
For the generic example variable and computed quantity in the clip.
Screen shows " = (mv)".
Narrator states the uncertainty on K is the derivative of K with respect to v times the uncertainty on v.
The intermediate differentiation step is not spoken aloud; it is inferred from the displayed simplification.
For with constant, the clip uses , so .
.
is treated as exact/constant.
carries the uncertainty .
For the worked kinetic-energy example.
The narrator says, 'If the regular derivative is the slope of a tangent line...'
Earlier, a red tangent line is drawn on the K-versus-v graph.
For a one-variable function, the derivative is interpreted as the slope of the tangent line to the graph at a point.
The function is differentiable at the point under consideration
At a point where the ordinary derivative exists.
The narrator says that in two dimensions there is a tangent plane, and that ∂/∂x is the slope along x while ∂/∂y is the slope along y.
The screen labels ∂f/∂x = slope here along x and ∂f/∂y = slope here along y.
For a two-variable surface, the partial derivatives give the slopes of the tangent plane in the coordinate directions x and y.
The surface is locally smooth enough to have a tangent plane
The variables are distinguished as x and y
At a point on the surface where the relevant partial derivatives exist.
The narrator says the quadrature sum is less than the worst case but greater than either individual uncertainty.
The screen shows hypotenuse < .
A right triangle with legs and and hypotenuse sqrt() is displayed.
The spoken comparison to the individual uncertainties is clear, but the final sentence is cut off before the narrator finishes the thought.
For two positive uncertainty contributions, the quadrature combination sqrt() is smaller than the fully additive worst case and larger than either single contribution.
and are treated as positive uncertainty magnitudes
The combination is modeled by the perpendicular geometric construction
For the two displayed positive contributions and .
A handwritten note points to the substituted factors and reads: Notice: the +/- signs don't matter because they get squared!
The narrator says both derivatives are really easy and equal to 1 in magnitude.
For , the propagated uncertainty depends on ()^2 and ()^2, so the plus/minus signs do not change the final quadrature formula.
z is a sum or difference of two measured variables x and y.
The propagation formula squares each derivative term.
For the displayed choices of signs in .
Both product and quotient derivations simplify to the same displayed expression for .
The narrator says there is a pattern with the relative uncertainties and that they add in quadrature.
The verbal definition of relative uncertainty is cut off before completion.
After dividing the propagated absolute uncertainty by z, both z=xy and yield the same rule: the relative uncertainties of x and y add in quadrature.
z is either xy or .
x and y are nonzero so the ratios are defined.
The quadrature propagation formula is being used.
For the two displayed functional forms z=xy and .
The narrator says that if one measurement is much more accurate, relatively speaking, than the other, you can sometimes ignore the smaller relative uncertainty and treat that measurement as though it were exact.
The term ()^2 is visually shrunk away while ()^2 remains dominant.
The video does not state a precise numerical threshold for when “much more accurate” justifies neglecting a term.
If one relative uncertainty is much larger than another, the smaller relative uncertainty can sometimes be neglected, so the combined relative uncertainty is approximately the larger one.
The comparison is between relative uncertainties.
One measurement is much more accurate relative to the other.
The approximation is acceptable for the intended level of precision.
sometimes
After simplifying to , the denominator z is replaced by xy and canceled to give \,.
The narrator warns: This will save you a lot of math, but be careful.
A speech bubble says: Wow, it really does behave as if "x" were constant!
The video presents this as an illustrative consequence of the approximation, not as a fully justified theorem.
Under the approximation that x has negligible relative uncertainty, the propagated absolute uncertainty in z=xy reduces to δz≈x δy, making x behave like a constant multiplier.
z=xy.
The relative uncertainty in x is neglected compared with that in y.
under the stated approximation
The narrator says one can construct a right triangle in three dimensions using the three different sources of uncertainty, and the hypotenuse is the square root of all three uncertainties squared.
The 3D vector diagram labels the diagonal as .
The video does not explicitly prove the geometric statement; it presents it as a visual model.
For three independent uncertainty components, the combined uncertainty is represented by the hypotenuse of a right triangular construction in three dimensions, namely the square root of the sum of the three squared components.
There are three uncertainty components.
They are arranged as mutually perpendicular legs in the visual model.
for the displayed three-component construction
Displayed sequence: " % = ".
Narrator explains expressing uncertainty as a percentage or fraction of the measured quantity.
Divide the absolute uncertainty by the measured velocity.
Definition of relative uncertainty as stated in the clip.
Evaluate the quotient numerically.
Arithmetic shown on screen.
Convert the decimal fraction to a percentage.
Equivalent percentage form displayed on screen.
Identify the result as the relative uncertainty in .
Labeling shown on screen and narration defining relative uncertainty.
The relative uncertainty in the example is , equivalently . The exact decimal is rounded to on screen.
Curve , point , uncertainty bars and , and a red tangent line.
Displayed equation: "".
Narrator says the movement in is based on the slope of the function.
The clip presents the tangent-line argument visually and states the formula, but does not explicitly discuss the approximation sign or neglected higher-order terms.
Start with a computed quantity depending on a measured input .
Stated by the narrator and shown by the graph label .
Allow the input to vary by an absolute uncertainty .
Shown by horizontal uncertainty markers around .
Use the local slope of the curve at the point .
Red tangent line drawn at the point on the graph.
Multiply the slope by the input offset to get the output offset.
Formula displayed on screen and stated in narration.
The clip’s visual derivation relates a small input uncertainty to an output uncertainty through the derivative at the operating point.
Screen shows "", then " = (mv)".
Narrator says is exact and the uncertainty on is the derivative of with respect to times the uncertainty on .
The differentiation step from to is not verbally spelled out; it is visible only through the final displayed simplification.
Begin with the kinetic-energy formula for the example.
Displayed formula and spoken by narrator.
Treat mass as exact and velocity as the uncertain measured quantity.
Narrator explicitly says was an exact number and identifies as the uncertainty on .
Apply the one-variable propagation rule from the previous section.
General method established earlier in the clip and restated in narration.
Differentiate with respect to holding constant.
Implied by the displayed simplification to ; standard derivative rule.
Substitute the derivative into the propagation formula.
Displayed on screen.
For the kinetic-energy example with exact mass, the uncertainty in kinetic energy is .
The screen shows mv^2, then δK=dK/dv δv=(mv)δv, then numeric substitution, then 0.94 J, then the boxed answer +/- 0.9 J.
The narrator says the purpose is to compute both K and the uncertainty on K.
Start from the kinetic-energy formula.
Displayed formula on screen.
Substitute the given mass and velocity to get the nominal kinetic energy.
Direct numerical substitution shown on screen.
Differentiate K with respect to v while treating m as constant.
Standard derivative of with respect to v.
Use the linear propagation rule for one uncertain variable.
Displayed relation δK = dK/dv δv.
Insert the numerical values for m, v, and δv.
Substitution shown on screen.
Round the nominal value and uncertainty for the final reported result.
Boxed answer shown on screen.
The kinetic energy is reported as .
The screen writes xy^2z+sqrt(y)+ and then displays the three partial derivatives.
Terms not containing the chosen variable are crossed out, and the word constant appears above ignored factors.
Begin with the displayed three-variable function.
Formula shown on screen.
Differentiate only the terms containing x, treating y and z as constants; the sqrt(y) term disappears.
Partial-derivative rule stated by the narrator and shown by crossing out non-x terms.
Differentiate only the terms containing y, treating x and z as constants; the term disappears.
Partial-derivative rule stated by the narrator and shown by crossing out non-y terms.
Differentiate only the terms containing z, treating x and y as constants; the sqrt(y) term disappears.
Partial-derivative rule stated by the narrator and shown by crossing out non-z terms.
The displayed partial derivatives are ∂, ∂f/∂xyz+1/(2√y), and ∂f/∂xy^.
Arrows and are shown in opposing, aligned, and perpendicular configurations.
The right triangle labels the hypotenuse as sqrt() and the screen writes hypotenuse < .
The narrator explains best case cancellation, worst case addition, and the perpendicular case as an average relationship for random independent effects.
The clip gives a geometric motivation rather than a formal probabilistic derivation.
The animation first shows the two contributions pointing in opposite directions, representing partial cancellation.
Visual arrow configuration plus narration about best case.
The animation then shows the two contributions aligned, representing the worst-case additive total.
Visual collinear arrows plus narration about maximum error.
The perpendicular arrangement forms a right triangle whose hypotenuse is the quadrature sum.
Displayed right-triangle label and Pythagorean form.
Because the hypotenuse of a right triangle is shorter than the sum of the two legs, the quadrature result is less than the fully additive worst case.
On-screen inequality hypotenuse < and elementary triangle geometry.
The perpendicular geometric model motivates why independent uncertainties are combined by quadrature rather than by simple addition.
The screen shows , then the quadrature formula for , then the partial derivatives, then numerical substitution, then = 0.99 J.
The narrator explains that each source of uncertainty involves a derivative with respect to a different variable and then says to plug in all the numbers.
Start from the kinetic-energy formula with both m and v treated as measured quantities.
Given in the problem statement on screen.
Write the two-variable propagation formula for the uncertainty in K.
Stated by the narrator as adding the two uncertainty sources in quadrature.
Differentiate K with respect to m while holding v fixed.
Direct partial differentiation of mv^2.
Differentiate K with respect to v while holding m fixed.
Direct partial differentiation of mv^2.
Substitute the two derivative results back into the propagation formula.
Algebraic substitution of the evaluated partial derivatives.
Insert the numerical values from the problem statement for m, v, , and .
Numerical substitution from the given measurements.
Evaluate the expression to obtain the propagated uncertainty.
Arithmetic evaluation shown on screen.
For the given kinetic-energy example, the propagated absolute uncertainty in K is 0.99 J.
The slide writes , then , then = .
The narrator says the derivative of z with respect to x is simply C, therefore the uncertainty on z is C times the uncertainty on x.
Begin with a derived quantity that is a constant multiple of one measured variable.
Displayed as the heading case on screen.
Apply the single-variable propagation template to z.
This is the one-variable form of the quadrature rule shown earlier.
Differentiate z=Cx with respect to x.
Direct differentiation.
Substitute the derivative into the propagation formula.
Algebraic substitution.
When a measurement is multiplied by an exact constant, its absolute uncertainty is multiplied by the same constant.
The slide shows , then the two-variable formula, then substitutes for both derivatives, then simplifies to .
The narrator says both derivatives are 1 and the formula reduces to the two original uncertainties added in quadrature.
Start with a sum or difference of two measured variables.
Displayed as the case heading on screen.
Write the general two-variable propagation formula.
Same template used in the kinetic-energy example.
Compute the partial derivatives of a linear sum or difference.
Direct differentiation of .
Substitute the derivative values into the propagation formula.
Algebraic substitution.
Remove the signs because squaring makes become 1.
The on-screen note explicitly states that the +/- signs do not matter because they get squared.
For addition or subtraction of two measured quantities, the absolute uncertainties combine in quadrature.
The left column derives from =xy using and ; the right column derives from using and .
Both columns are then divided by z and simplified to the same displayed expression for .
The narrator says the derivatives are more complicated at first glance, but dividing both sides by z reveals a pattern in which the relative uncertainties add in quadrature.
The clip ends before the narrator completes the verbal definition of relative uncertainty.
Begin with the product case.
Displayed as the left-hand case on screen.
Use the two-variable propagation formula after computing and .
Direct partial differentiation and substitution into the quadrature template.
Divide both sides by z=xy to rewrite the expression in relative form.
Algebraic division by z, as stated by the narrator.
Cancel common factors inside each squared term.
Simplification of fractions.
Repeat the process for the quotient case.
Displayed as the right-hand case on screen.
Use the two-variable propagation formula after computing and .
Direct partial differentiation and substitution into the quadrature template.
Divide by and simplify; the minus sign disappears after squaring.
Algebraic simplification shown on screen.
For both multiplication and division, the relative uncertainty of the result is the quadrature sum of the relative uncertainties of the inputs.
The full relative-uncertainty formula is shown before simplification.
The term ()^2 shrinks and disappears while ()^2 remains enlarged.
The narrator explains ignoring the smaller relative uncertainty when one measurement is much more accurate.
The exact numerical criterion for “much more accurate” is not specified.
Start from the relative-uncertainty formula for multiplication or division.
Displayed on screen as the standard rule for products and quotients.
Assume the relative uncertainty in x is much smaller than that in y.
Stated verbally as the condition under which one measurement is much more accurate, relatively speaking.
Drop the smaller squared term from the sum under the square root.
Approximation justified by dominance of one relative uncertainty over the other.
Take the square root of the remaining squared term.
Algebraic simplification of =|a|, with uncertainty ratios treated as nonnegative magnitudes.
When one relative uncertainty dominates, the combined relative uncertainty is approximately equal to the larger one.
The screen shows Let's say... =xy together with .
The denominator z is rewritten as xy and the x factors cancel, leaving .
The narrator says this saves math but warns to be careful.
The cancellation step is shown visually rather than spoken in full detail.
Introduce the specific product relation for z.
Written on screen as Let's say... =xy.
Use the approximation obtained when the x-uncertainty term is neglected.
Carried over from the previous dominant-term simplification.
Substitute z=xy into the denominator on the left-hand side.
Direct substitution from the given functional relation.
Cancel the common factor y and multiply through by x.
Algebraic simplification shown by the on-screen cancellation animation.
Under the approximation that x is effectively exact, the absolute uncertainty in z=xy becomes δz≈x δy.
Radar gun aimed at car; display shows "" and then "+/- ".
Labels identify "measured velocity" and "absolute uncertainty in v" .
Relative uncertainty calculation shown as .
A measured velocity is given as . Identify the absolute and relative uncertainties.
Measured velocity .
Absolute uncertainty .
Express the uncertainty both as an absolute quantity and as a relative fraction/percentage.
Read the plus-or-minus amount directly as the absolute uncertainty.
Definition of absolute uncertainty given in the clip.
Form the ratio of absolute uncertainty to measured value.
Definition of relative uncertainty stated in the clip.
Evaluate and convert to percentage.
Numerical result displayed on screen.
Absolute uncertainty: . Relative uncertainty: .
The units cancel in , matching the clip’s statement that relative uncertainty is unitless.
Problem text: "The measured velocity of a 10-kg mass is 2.34 +/- . Find the kinetic energy."
Screen shows , then substitution .
Screen shows .
The clip stops after setting up the substitution and propagation formula; no final numerical value for or is shown within this 120-second segment.
The measured velocity of a mass is . Find the kinetic energy, propagating the uncertainty from velocity.
Mass , treated as exact.
Measured velocity .
Velocity uncertainty .
Formula .
Set up the computation of and the propagated uncertainty .
Write the kinetic-energy formula for the example.
Displayed formula and spoken by narrator.
Substitute the given mass and velocity into the formula.
Substitution shown on screen.
Apply the one-variable propagation rule.
General method established earlier and restated here.
Replace by for .
Displayed simplification on screen.
Insert the numerical values into the propagated-uncertainty formula.
Follows directly from the displayed formula and given data; the clip reaches the setup stage by the end of the segment.
Within this clip, the example is set up as and . No final numerical answer is displayed before 120 seconds.
The setup matches the general rule applied to , with held exact as stated in the narration.
The problem statement reads: The measured velocity of a 10-kg mass is 2.34 +/- . Find the kinetic energy.
The screen computes and , then boxes +/- 0.9 J.
The measured velocity of a 10-kg mass is . Find the kinetic energy.
kg
mv^2
Compute the kinetic energy and its uncertainty.
Substitute the given mass and velocity into the kinetic-energy formula.
Direct evaluation shown on screen.
Propagate the velocity uncertainty using the derivative of K with respect to v.
Displayed linear propagation formula.
Insert the numerical values and compute the uncertainty.
Substitution shown on screen.
Round the result and uncertainty for the final answer.
Boxed answer shown on screen.
The final boxed answer matches the displayed nominal value 27.378 J rounded to 27.4 J and the displayed uncertainty 0.94 J rounded to 0.9 J.
The screen lists velocity = +/- and mass = 11.6 kg +/- 0.1 kg.
It then writes , function two variables, and δz = sqrt((∂z/∂x δx)^2 + (∂z/∂y δy)^2).
The narrator says this is a more complicated case where two measured things each have uncertainty and are combined into kinetic energy.
The clip sets up the two-variable example and the quadrature rule but does not carry out the full numerical computation within this segment.
Suppose velocity and mass are both measured with uncertainties, and you want the kinetic energy from the combination.
velocity =
mass = 11.6 kg ± 0.1 kg
Introduce how to combine two uncertain measurements into one derived quantity.
Identify the derived quantity as kinetic energy depending on both mass and velocity.
Formula shown on screen.
Re-express the quantity as a generic function of two variables.
Screen text function two variables.
Apply the quadrature rule for combining two independent first-order uncertainty contributions.
Displayed formula and narrated term 'adding the uncertainties in quadrature'.
The setup leads to the quadrature formula for δz; no final numerical value for this two-measurement example is completed in the clip.
The displayed formulas confirm the intended method, but the segment ends before a numeric answer is produced.
The screen writes An example: xy^2z+sqrt(y)+.
It then displays ∂, ∂f/∂y, and ∂f/∂z.
Terms without the chosen variable are crossed out and labeled constant.
Compute the partial derivatives of xy^2z+√.
xy^
Find ∂f/∂x, ∂f/∂y, and ∂f/∂z.
Treat y and z as constants and differentiate the x-dependent terms.
Shown by crossing out √y and differentiating 3xy^2z and with respect to x.
Treat x and z as constants and differentiate the y-dependent terms.
Shown by crossing out and differentiating 3xy^2z and √y with respect to y.
Treat x and y as constants and differentiate the z-dependent terms.
Shown by crossing out √y and differentiating 3xy^2z and with respect to z.
∂f/∂, ∂f/∂xyz + 1/(2√y), ∂xy^
Each displayed derivative matches standard term-by-term differentiation with the other variables held constant.
The problem text reads: A mass of 10.12 +/- 0.11 kg has a measured velocity of 2.34 +/- . Find the kinetic energy.
The worked solution displays and the full propagation calculation ending in = 0.99 J.
The narrator says to revisit the kinetic energy example with two variables and then plug in all the numbers.
The clip computes and displays only the propagated uncertainty , not the central value of K itself.
A mass of 10.12 +/- 0.11 kg has a measured velocity of 2.34 +/- . Find the kinetic energy uncertainty using .
kg
kg
mv^2
Propagate the measurement uncertainties to obtain .
Set up the two-variable uncertainty propagation formula for K.
The narrator states that the two uncertainty sources are added in quadrature.
Compute the needed partial derivatives.
Differentiation of mv^2 with respect to m and v.
Substitute the numerical values from the problem statement.
Direct numerical substitution shown on screen.
Evaluate the expression.
Final arithmetic result displayed on screen.
The displayed answer matches the final line of the on-screen calculation. The clip does not show a separate check against the central value of K.
The slide states: If measured +/- 0.006, then 10 times +/- 0.06.
The narrator says that when you multiply a measurement by a constant, you multiply its uncertainty by the same constant.
If a measured quantity is +/- 0.006, what is 10 times x?
multiplier
Apply the constant-scaling uncertainty rule.
Identify the new quantity as a constant multiple of x.
The example explicitly asks for 10 times x.
Multiply the uncertainty by the same constant.
Uses the rule shown immediately above.
Compute the scaled value and scaled uncertainty.
Direct arithmetic from the given numbers.
10 times +/- 0.06
The result matches the example line shown on screen.
Text above the formula reads 0.03%: very accurate measurement pointing to ()^2 and 5.4%: sloppy measurement pointing to ()^2.
The 0.03% term shrinks away while the 5.4% term remains and enlarges.
The example is presented qualitatively; no explicit numeric computation of the final percentage is shown.
Compare two relative uncertainties, 0.03% and 5.4%, in the quadrature formula and see which term controls the result.
One relative uncertainty is 0.03%.
The other relative uncertainty is 5.4%.
The combination rule is .
Show why the smaller relative uncertainty can be ignored in this approximation.
Insert the two displayed relative uncertainties into the quadrature formula.
Direct substitution into the shown rule.
Observe that the squared small term is negligible compared with the squared large term.
Numerical comparison of the displayed percentages.
Approximate the combined relative uncertainty by the larger one.
Matches the visual removal of the smaller term and the narrator’s explanation.
The combined relative uncertainty is approximately 5.4%.
The animation removes the 0.03% contribution and leaves only the 5.4% term under the radical.
The screen writes Let's say... =xy and then transforms into .
A speech bubble comments: Wow, it really does behave as if "x" were constant!
The example is symbolic rather than numeric.
Given z=xy and the approximation that the relative uncertainty in x is negligible, find the resulting absolute uncertainty in z.
=xy.
.
Rewrite the approximation as an absolute-uncertainty relation.
Substitute z=xy into the approximate relative-uncertainty equation.
Direct substitution from the stated function.
Cancel y and solve for δz.
Algebraic simplification shown on screen.
\,
The on-screen cancellation animation ends exactly at , and the speech bubble interprets x as behaving like a constant.
A door labeled "Physics Lab" opens to reveal a large yellow symbol; title text "PROPAGATION of ERRORS" appears.
Door labeled "Physics Lab"
Yellow symbol
Title text "PROPAGATION of ERRORS"
The door opens.
The symbol is revealed.
The title text appears.
The setting remains a simple cartoon laboratory doorway.
The opening visually introduces the uncertainty symbol that will denote absolute uncertainty later.
A hand holds a radar gun aimed at a teal car; the display changes from blank to "" and then "+/- ".
Teal car
Radar gun
Display values "" and "+/- "
The radar display first shows the measured speed.
Then the plus-or-minus uncertainty is added.
The car and radar-gun setup remain the same.
This animation introduces a concrete measured quantity and its associated absolute uncertainty.
The expression is rewritten with labels: black oval around "" under "measured velocity v="; blue oval around "" under "absolute uncertainty in v"; red oval around the percentage/fraction under "relative uncertainty in v".
A yellow arrow points to the blue-circled absolute uncertainty with text "How to find this...? (That's another video.)".
Circled measured value
Circled absolute uncertainty
Circled relative uncertainty
Yellow arrow annotation
The single measurement expression is separated into labeled parts.
Blue emphasis marks the absolute uncertainty.
Red emphasis marks the relative uncertainty.
A note indicates how to find absolute uncertainty is deferred to another video.
The underlying numerical example remains .
The visual coding distinguishes the measured quantity, its absolute uncertainty, and its relative uncertainty.
Two speech bubbles labeled "error" and "uncertainty" appear with thought bubbles: "I'm easier to say." and "I'm more descriptive of what you actually mean."
Speech bubble "error"
Speech bubble "uncertainty"
Thought bubbles
The two terms are personified as competing labels.
No new formula is introduced in this scene.
The animation comments on terminology rather than adding a distinct mathematical rule.
A coordinate graph builds from axes to a curve labeled , with a marked point , corresponding , horizontal uncertainty bars , vertical uncertainty bars , and a red tangent line.
Axes labeled and
Curve
Point
Horizontal markers
Vertical markers
Red tangent line
Axes and point are drawn first.
The curve appears.
Uncertainty intervals are added on both axes.
A tangent line is drawn at the point.
The formula is written.
The graph represents one input variable mapped to one output .
The picture shows that a small horizontal uncertainty is translated into a vertical uncertainty using the local slope.
A new graph labeled shows a rising curve, a point , blue vertical uncertainty markers at , green horizontal uncertainty markers at , and a red tangent line.
Text and formulas include the problem statement, , and .
Problem statement text
Graph axes and
Curve
Point
Blue markers
Green markers
Red tangent line
Formulas
The kinetic-energy problem statement appears.
The graph of is drawn.
Velocity uncertainty is marked in blue on the horizontal axis.
Energy uncertainty is marked in green on the vertical axis.
A tangent line is drawn and the propagation formula is written.
Numerical substitution for begins near the end.
Mass is treated as fixed while velocity varies.
This scene applies the generic tangent-line rule to the specific function .
A black curve rises on axes labeled K vertically and v horizontally.
A red tangent line touches the curve at a marked point, and dashed guides connect the point to the axes.
A yellow arrow points first to 27.378 J and later to 0.94 J.
K-versus-v curve
red tangent line
point marking v and K
numeric labels 27.378 J and 0.94 J
The tangent line is emphasized to connect dK/dv with local slope.
The yellow arrow shifts attention from the computed kinetic energy to the computed uncertainty.
The axes remain labeled K and v.
The underlying formula mv^2 stays on screen.
The picture links the algebraic derivative dK/dv to the geometric slope of the tangent line, supporting the linear estimate δK≈(dK/dv)δv.
A hand holds a device labeled SPEED LADAR showing with +/- .
A balance scale shows a car on one side and a weight labeled 11.6 kg +/- 0.1 kg on the other.
speed gun display
toy car
balance scale
mass label 11.6 kg +/- 0.1 kg
The scene moves from a single velocity measurement to a paired velocity-and-mass measurement scenario.
Both examples concern kinetic energy.
Each measured quantity is accompanied by an explicit uncertainty.
The visuals motivate the need for a rule that combines two independently uncertain inputs into one derived quantity.
The painting The Scream appears with a speech bubble containing ∂.
The isolated ∂ is then shown in ∂/∂x, /∂y, and /∂z with the label Partial Derivative.
speech bubble with ∂
notation ∂/∂x
notation ∂/∂y
notation ∂/∂z
The curly symbol is isolated from the surrounding formula and identified by name.
The symbol remains ∂ throughout the introduction.
This visual sequence separates notation from computation, preparing the viewer for the geometric and procedural explanations that follow.
A 3D surface labeled is shown with axes X, Y, and Z.
Text labels ∂f/∂x = slope here along x and ∂f/∂y = slope here along y.
A highlighted planar slice or tangent plane is shown on the surface.
The exact algebraic formula for the displayed surface is not given in the clip.
3D surface
highlighted tangent plane
axis labels X, Y, Z
slope annotations
The view emphasizes first the x-direction slope and then the y-direction slope on the same surface.
The surface remains a two-variable graph in three-dimensional space.
The animation turns the abstract notation ∂/∂x and ∂/∂y into coordinate-direction slopes on a tangent plane.
Arrows labeled and rotate through opposing, aligned, and perpendicular arrangements.
The right triangle is labeled with hypotenuse sqrt().
The screen writes hypotenuse < .
arrow
arrow
right triangle
hypotenuse label sqrt()
The relative angle between and changes from opposition to alignment to perpendicularity.
The final static triangle isolates the perpendicular case used for quadrature.
The two contributions are represented by the same labels and throughout.
The animation contrasts extreme error combinations and then uses the perpendicular case to motivate why independent uncertainties are combined by the square-root-of-sum-of-squares rule.
A right triangle appears with blue horizontal leg , red vertical leg , and magenta hypotenuse labeled .
Text below the triangle reads: hypotenuse < , but > alone, and > alone.
horizontal leg labeled
vertical leg labeled
hypotenuse labeled
inequality text below the triangle
The clip opens on the completed triangle and inequality statement, then transitions away to a road-sign graphic.
The triangle remains a right triangle throughout its brief appearance.
The labels identify the two components and their quadrature sum.
The picture encodes the Pythagorean relationship behind adding two uncertainty components in quadrature and visually compares that result with simple addition.
Narrator says some people say error because uncertainty has too many syllables and tells viewers not to sweat the difference.
Cartoon bubbles contrast "error" and "uncertainty".
Learners may think “error” and “uncertainty” must be treated as sharply different concepts before proceeding.
In this clip, the narrator explicitly tells viewers not to worry about the terminology difference; the mathematics being taught is the same plus-or-minus uncertainty idea.
Narrator contrasts absolute uncertainty having units with relative uncertainty being a fraction or percentage, unitless.
is shown with units , while is shown as .
One might incorrectly attach the original measurement units to the relative uncertainty.
The clip shows that keeps the units of , but is dimensionless because the units cancel.
Yellow arrow points to the absolute uncertainty with text "How to find this...? (That's another video.)".
Viewers may expect the lesson to teach how the experimental uncertainty itself is determined.
The clip explicitly defers that question to another video and focuses instead on how a given uncertainty propagates through a formula.
The narrator says, 'You already know how to do it. The special symbol just makes it look impressive.'
The ∂ notation may seem to require a new kind of calculus beyond ordinary differentiation.
The video states that partial derivatives are calculated like regular derivatives once you choose one variable and treat the others as constants.
The narrator distinguishes best case cancellation, worst case addition, and the random independent case represented at 90 degrees.
The screen shows hypotenuse < .
One might think the total uncertainty is always the arithmetic sum of the individual uncertainties.
The video explains that simple addition corresponds to the worst case, while independent random effects are modeled by quadrature, which gives a smaller combined value than full addition.
A handwritten annotation explicitly says: Notice: the +/- signs don't matter because they get squared!
The derivation substitutes into squared terms and then removes the signs.
One might expect to produce a different uncertainty formula from because of the minus sign.
In the quadrature formula shown, the derivative signs are squared, so addition and subtraction both reduce to .
The narrator says that at first glance the result looks messy, but dividing both sides by z reveals a pattern.
The raw derivative expressions for xy and are rewritten as a common relative-uncertainty formula.
The unsimplified expressions involving , , (), and (-x/y^2) can make multiplication and division seem unrelated or overly complicated.
After dividing by z and simplifying, both cases become the same rule for relative uncertainties: .
The narrator says: This will save you a lot of math, but be careful. Can you think of an example when this would not be a good idea?
The speech bubble emphasizes that x behaves as if it were constant under the approximation.
The video asks for a counterexample but does not supply one within this clip.
Because dropping a much smaller relative uncertainty simplifies the algebra, one might think it is always valid to treat the more accurate measurement as exact.
The video explicitly warns that this shortcut must be used carefully and invites the viewer to think of cases where it would not be a good idea.
The opening slide writes =xy and even though two independent variables are present.
Later the clip writes =xy.
This is an analyst note about presentation consistency, not a misconception stated by the video.
A viewer may infer that z depends only on x because the header writes .
The later slide and the formulas themselves make clear that z depends on both x and y; the initial notation is shorthand or imprecise.
Relative uncertainty is written as after absolute uncertainty has been defined.
Relative uncertainty is defined by dividing the absolute uncertainty by the measured quantity, so the absolute-uncertainty notion comes first.
The displayed identity links the numeric example to the general ratio form.
The clip contrasts absolute uncertainty, which has units, with relative uncertainty, which is the corresponding unitless ratio.
The propagation rule uses and , the absolute-uncertainty notation introduced earlier.
The propagation method operates on absolute uncertainties and .
The example writes , directly matching the general pattern .
The kinetic-energy uncertainty formula is a direct application of the general one-variable propagation rule.
From , the clip obtains .
The propagated uncertainty formula depends on differentiating the specific kinetic-energy expression.
The generic graph is followed by the specific graph with the same tangent-line structure.
The kinetic-energy example is a special case of the generic propagation method with and .
The same screen moves from mv^2 to δK=dK/dv δv=(mv)δv.
The kinetic-energy formula supplies the function whose derivative is used to propagate the velocity uncertainty.
The narrator says, 'Let's make it a little more complicated' after the one-variable example.
The screen replaces the one-variable setup with and the two-variable quadrature formula.
The two-variable quadrature rule extends the one-variable linear propagation idea to functions depending on two uncertain inputs.
The quadrature formula explicitly contains ∂z/∂x and ∂z/∂y.
Immediately afterward the narrator explains what the curly d symbols mean.
Understanding partial derivatives is required before using the displayed quadrature formula for two-variable uncertainty propagation.
The narrator first defines partial derivative and then says that in two dimensions there is a tangent plane.
The surface animation labels the coordinate-direction slopes.
The geometric tangent-plane interpretation is one way of understanding the meaning of the partial-derivative definition.
After the surface explanation, the narrator says partial derivatives are easy to calculate just like regular derivatives.
The example then computes ∂, ∂, and ∂f/∂z algebraically.
The video separates the geometric meaning of partial derivatives from the practical rule for computing them.
The narrator ties random independent effects to the 90-degree representation.
The right-triangle hypotenuse is labeled sqrt().
The independence and randomness assumption is the conceptual basis used in the clip to motivate the quadrature combination rule.
Text defines '"" = "absolute uncertainty in..."'.
Narrator contrasts units of absolute uncertainty with unitless relative uncertainty.
Both and are displayed side by side.
The calculation is shown.
Narrator says not to sweat the difference between error and uncertainty.
A red tangent line is drawn on the graph of .
Equation appears.
Displayed formula: .
Displayed chain: .
Narrator says, "Suppose that m was an exact number."
The substitution for is shown, but the clip ends before a final numerical result appears.
Whether the next segment completes the arithmetic is outside this clip.
The screen shows δK = dK/dv δv = (mv)δv.
The narrator says, 'Let's make it a little more complicated' and introduces two measured things each with uncertainty.
The narrator asks, 'What's with those curly d symbols?' and answers that this is called a partial derivative.
Covered · Title card and opening reveal of the symbol introduce the topic and notation context.
Covered · Radar-gun example defines absolute uncertainty with and .
Covered · The clip computes and contrasts units of absolute versus relative uncertainty.
Covered · Terminology aside comparing “error” and “uncertainty.”
Covered · Generic graph of with tangent line leads to the propagation rule .
Covered · Worked kinetic-energy example sets up , derives , and begins numerical substitution; no final numeric answer appears within this segment.
Covered · First worked example computes kinetic energy and its uncertainty from uncertain velocity.
Covered · Transition to a more complicated case with two measured quantities, each having uncertainty.
Covered · The two-variable quadrature formula is displayed and named.
Covered · The partial-derivative symbol is introduced and defined verbally.
Covered · Partial derivatives are interpreted as coordinate-direction slopes on a tangent plane.
Covered · A three-variable example demonstrates how to compute partial derivatives term by term.
Covered · Geometric motivation explains why independent uncertainties combine by quadrature rather than by worst-case addition.
Covered · Opening triangle and inequality summarize quadrature combination of two uncertainty components.
Covered · Transition road-sign graphic returns the lesson to a practical example; no new mathematical content is introduced.
Covered · Full worked kinetic-energy example with two measured variables and final propagated uncertainty 0.99 J.
Covered · Title card introduces the tricks-and-shortcuts section.
Covered · Multiplication by a constant rule and its numerical example.
Covered · Addition and subtraction shortcut derived from the general quadrature formula.
Covered · Product and quotient derivations simplified to a common relative-uncertainty rule; the spoken definition of relative uncertainty is cut off at the end, but the displayed mathematics is complete for this segment.
Covered · Opening slide presents product and quotient cases with absolute and relative uncertainty formulas and the quadrature statement.
Covered · Animated comparison of 0.03% and 5.4% relative uncertainties shows neglecting the smaller term.
Covered · Symbolic example for z=xy converts the simplified relative relation into δz≈xδy.
Covered · 2D and 3D vector diagrams give a geometric interpretation of quadrature addition.
Covered · Summary board lists addition/subtraction, multiplication/division, and general partial-derivative rules, then extends each to three variables.
Covered · Humorous pie chart about arguing with research collaborators; no new mathematical content.
Covered · End credits and Physics Lab door animation; no mathematical content.
Reviewed subject paths
For addition and subtraction, the absolute uncertainties of the inputs combine in quadrature. The plus or minus signs in the formula do not matter because they are squared in the propagation equation, resulting in the same rule for both operations.
Conditions: The derived quantity is a sum or difference of two measured variables ().; The variables and are independent measured quantities.
You can sometimes ignore the smaller relative uncertainty when one measurement is much more accurate, relatively speaking, than the other. In this case, the combined relative uncertainty is approximately equal to the larger one, and the more accurate measurement behaves as if it were an exact constant.
Conditions: The comparison is between relative uncertainties.; One measurement is much more accurate relative to the other.; The approximation is acceptable for the intended level of precision.
When both mass and velocity are measured, the uncertainty in kinetic energy is found by adding the two uncertainty sources in quadrature. You compute the partial derivative of kinetic energy with respect to each variable, multiply each derivative by the corresponding absolute uncertainty, square those terms, sum them, and take the square root.
Conditions: The kinetic energy formula is .; Both mass and velocity carry independent absolute uncertainties and .; The uncertainties are small enough for a first-order linear approximation.
On a two-variable surface, partial derivatives represent the slopes of the tangent plane in the coordinate directions. The partial derivative with respect to is the slope along the -direction, and the partial derivative with respect to is the slope along the -direction.
Conditions: The surface is locally smooth enough to have a tangent plane.; The variables are distinguished as and .
The curly d symbol () represents a partial derivative. It means taking the derivative of a multivariable function with respect to one specific variable at a time, while treating all other variables as constants.
Conditions: The function has more than one independent variable.; All variables except the differentiation variable are treated as constants during the partial derivative.
For products and quotients, the relative uncertainties of the inputs add in quadrature. You calculate the relative uncertainty of the result by taking the square root of the sum of the squared relative uncertainties of the individual variables.
Conditions: The derived quantity is a product () or a quotient ().; The variables and are nonzero so the relative uncertainty ratios are defined.; The uncertainties are treated as independent.
Independent uncertainties are added in quadrature because they represent random effects that can push a measurement up or down. If the effects are random and independent, their combined influence is neither always cancelling nor always additive.
Conditions: The two uncertainty effects are random.; The two effects are independent from one another.; The combination is modeled by a perpendicular geometric construction.
Absolute uncertainty is the plus-or-minus amount attached to a measured quantity and is written in the same units as that quantity. Relative uncertainty expresses the uncertainty as a fraction or percentage of the measured value, obtained by dividing the absolute uncertainty by the measured quantity, making it dimensionless.
Conditions: Applies to a measured quantity with a nonzero value when forming the relative ratio.; The absolute uncertainty is expressed in the same units as the measured quantity.
Continue with related videos in the same subject.
YouTube7:15
YouTube3:33
YouTube4:48