WODAnalysis – Teacher handout

Complete overview of the features, shown for Science mode, together with the limits of the app in classroom use. It complements the student handout – it does not replace it. As of program version 95. · Deutsche Fassung

What WODAnalysis is. Video analysis for physics class that runs entirely in the browser: no installation, no account, no upload. Video and measured values stay on the device. After the first visit the app works without internet as well, and on a phone it can be saved like an app via “Add to Home Screen”. It is the browser-native counterpart to Viana (iOS) and Tracker (Java desktop).

Contents

1 Before the lesson: checklist

Check the versionAt the bottom of the start screen: Version 95. If a device shows a lower number, an old version is still active there – reload the page once. If the line adds “version … ready – reload the page”, a newer version has already been downloaded but is not running yet. Different versions within one class lead to “on my screen it looks different”.
Device testOn the actual school computer, load a video once and run a short tracking pass in Science mode. That loads the ~5 MB template-matching module (and on a PC the ~19 MB AI guard) once; after that they are available offline.
PC linkNever test it for the first time during the lesson. On managed school computers it often fails – the reasons and the alternative are in section 9.
Recording rulesTell the class: 1080p, 30 frames per second, no more than 30 seconds, camera steady and perpendicular to the plane of motion, a reference object (meter stick) in exactly that plane and in shot. iPhone: Camera → Formats → Most Compatible.
Reference lengthsDecide them in advance and write them on the board – the app offers no list of standard lengths to choose from; you type the number in.
Handing in workDecide how results are collected (learning platform, network drive, USB stick). The app itself cannot send anything.

2 The two modes

Choosing the mode on the start screen is the only decision you have to make. It can be changed at any time via ⚙ → Mode; all data is kept. Switching Science → Student only hides objects 2 and 3 and the advanced tools, it deletes nothing.

Student modeScience mode (this handout)
Load video, frame by frame, limit clip range, track one object, manual point correction, scale, move the origin, s(t) and v(t), path points, CSV export. Axes: only t, s, x, y, v.
After the mode has been chosen, a short guide is offered: a speech bubble walks through loading the video, the playback buttons, clip range, axes and marking up to tracking, axis choice and scale, with “Next” and “Back”; completed steps advance on their own.
In addition: a(t), all quantities as x/y components, momentum p, force F = m·a, energies, mass per object, rotating the axes, adjustable g, vector arrows, stacked view, area calculation, model fits, strobe image, center-of-mass curve, up to 3 objects, JSON export.
No guide – hence this handout.

3 Structure of the interface

Header bar (always visible)

Science mode t = 1.20 / 4.30 s · Frame 36 / 129 · 30 fps · 412 px/m 📁⤓→ Pair⚙

Mode badge · status bar · Open project / Save project / Export · Pair · Settings. The status bar is your most important monitoring tool: it shows the time, the frame number, the frame rate actually in use and – only once you have calibrated – the scale in px/m. If the px/m figure is missing, the app is calculating in pixels.

Toolbar (three drop-down menus)

Analysis ▾ Display ▾ Evaluation ▾ 🔓

Video, chart, bottom bar

The default is stacked (video on top, chart below). Two columns side by side appear only from about 900 × 480 pixels – that is on a PC and on a tablet, not on a phone. On a PC the boundary can be dragged freely between 20 % and 80 % with the splitter. The bottom bar holds the clip handles, the timeline, ⏮ ◀ ⏵ ▶ (⏮ jumps to the start of the chosen clip – which is also where t = 0 sits) and Load video.

4 The workflow of an analysis

1
Load the video. Load video, or on a PC drag the file into the window. The app then measures the frame rate (details in section 7) and sets the timeline and the clip range to the whole video.
A new video discards all existing measurements without asking. Use ⤓ Save project beforehand.
2
Limit the clip range. Drag the two handles above the timeline to the start and the end of the motion. Only this part is tracked and analyzed – and t = 0 is at the first data point, not at the start of the video. Move them later and every time value changes.
3
Place the coordinate system. Drag the blue dot where the axes cross to the position you want; for an inclined plane, rotate it at the rotation handle on the x axis (an angle readout appears). Then lock it straight away with 🔒 – the whole x axis acts as a rotation handle and is easily knocked askew while you work in the image.
The origin also sets the zero height for potential energy.
4
Mark the object. Zoom in first (mouse wheel or two fingers), then Analysis → Mark object and drag a tight box. The data point is the center of the box, not the object's center of mass – place it centered and close. Then reset the zoom to 1×.
5
Start tracking. Analysis → Start tracking. The first time, the template matching engine is loaded (~5 MB). While the run is in progress, a centered panel shows the progress (“Analysis running. … % done.”); all other functions are locked until it finishes – only Cancel analysis inside the panel remains available.
If the marking sits on a different frame from the start of the clip, the app asks first (“Marking and clip start do not match”): the tracker always takes its reference pattern from the first frame of the clip, where the box might hold nothing but background. As a rule, choose Start at the marked frame – the clip then begins there. Escape cancels without tracking. With project files from older versions and over the PC link the frame number of the marking is missing; there the question does not appear.
The data points are color-coded by confidence: green from 0.60 · yellow 0.45–0.60 · red below that; under 0.40 the run stops and reports the loss.
After the run, the closing message says how the frames were obtained – “(read while playing …, sought …, color tracking …)”; this addition only appears when frames had to be sought or the color tracking stepped in. If the warning about frames measured twice comes as well, the same video frame was evaluated twice – v and a are worthless at those points. Remedy: bring the window to the front, close other programs, track again; ⚙ → Accuracy over speed avoids the case but takes longer.
6
Set the scale. Analysis → Set scale, tap two points whose distance you know, enter the length in meters. Only after that do the charts carry real units; before that the app honestly calculates in px.
7
Enter the mass. Analysis → Set mass. Only then are momentum p, force F and the energies available. With two carts: for both of them.
8
Analyse the measurements. Choose the axis quantities at the selectors, then area, curve fit, table, strobe image or export (section 6).

5 Reference: every control

5.1 The “Analysis” menu – producing measured values

ControlWhat it doesRequires
Object 1 (24) ▾Opens the list of all objects; a click makes one of them the active object. The number in brackets is the number of data points. Almost every tool refers to the active object.from 2 objects or data points onwards
🗑 in the object listDeletes the data points of this object after a confirmation prompt. The region is kept – you can re-track right away.object has points
Mark objectMarking mode; dragging a box in the video defines the search region.video, no tracking running
Start trackingFrame-by-frame run over the limited clip range. Deletes this object's existing points without asking.region set
Set scaleTwo-point calibration, then you enter the real length.video
Set mass
(reads “Mass: 145 g”)
Mass of the active object. The basis for p, F, Eₖ, Eₚ, Eₘ and the center-of-mass curve.Science mode
Track 2nd / 3rd objectCreates another object (turquoise, raspberry) and makes it active. Three at most; objects once created cannot be removed again (only their points).Science mode

5.2 The “Display” menu – what is drawn on the video

ControlWhat it doesRequires
Coordinate systemAxes on/off. While hidden, they are also protected against being moved.video
PathAll tracked positions as an evenly spaced trace of dots – shows the shape of the trajectory (projectile parabola, circle).data points
TrailOnly the last 8 positions, fading out – emphasises the motion at this instant.data points
Show vectorsDialog with checkboxes for s (position vector), v, a, p, F – each with its own length slider (0.1× to 10×).
Below that, three settings that apply to all arrows at once: Symbols on the arrows labels every arrow with s, v, a, p or F (on by default; switch it off when many arrows on several objects clutter the image – the size slider is then greyed out), Size of the symbols 14–48 px (default 20) and Arrow thickness 1–5 as a factor on the automatic base width. All three are remembered – for the projector, setting them once is enough.
Science mode; p and F only with a mass
Center of mass⊕ in the video with a connecting line, plus a dashed curve in the chart. In a collision this is the didactic core: the center of mass moves on undisturbed.≥ 2 objects with mass and points
MirrorSwaps left and right in the image so that a motion running to the left gives positive x values. The rotation of the axes is not mirrored along with it – on an inclined plane, realign them afterwards.video, no tracking

5.3 The “Evaluation” menu, “Chart” group

ControlWhat it doesRequires
LinesConnects the data points. Off by default: measured values are points, the trend comes as a model curve.–
Time lineVertical cursor line through the chart. Off by default – useful only when t is on the x axis.–
Stack (+ count)1–3 compact charts one above the other, preset to x, vₓ, aₓ over t, with derivative arrows in between.Science mode
Chart rangeType in the axis limits by hand (min/max per axis) – for axes that stay comparable across several groups and for clean diagrams for the board. Automatic resets it.Science mode
AreaArea mode: dragging in the chart selects a range (section 6.1).Science mode
ModelChoice of no model / linear / quadratic / oscillation / damped oscillation (section 6.2).Science mode

5.4 The chart area itself

ControlWhat it does
Axis selectors
v in m/s t in s
Choice of quantity for y (top left) and x (bottom right). The two axes can be combined freely – including y over x (the trajectory) or Fₓ over x. The label also shows the unit: if it reads px, the scale is missing.
Selectable quantitiest · s · x · y · v · vₓ · vᵧ · a · aₓ · aᵧ · F · Fₓ · Fᵧ · p · pₓ · pᵧ · E (three curves) · Eₖ · Eₚ · Eₘ
LegendAppears from two curves on. A click hides a series – that state is kept (even after the clip range is changed). Important: area and fit apply to the first visible series.
Red markerMarks the current frame on every visible curve. A click on a data point jumps the video there.
Readout boxTop right: the numerical values of the quantities currently plotted, for the current frame, each object in its own color.
Zoom / panMouse wheel or two fingers; dragging pans. A double click restores the automatic fit. In area mode, dragging means “select a range” instead.

5.5 Settings ⚙

Only Language, Smoothing and Mode are visible; everything else sits collapsed under Advanced settings – expanding it reveals the remaining rows of this table, including Measure playback (a diagnostic card that counts drawn, expected and presented frames during playback – useful when a device stutters).

LanguageDE / EN, takes effect immediately; units and symbols stay unchanged.
Display size (projector)Presentation mode: 100 % / 125 % / 150 % enlarge all text – controls, chart axes, data table. The table then shows correspondingly less, but with larger rows. The choice is remembered and applies again at the next start.
Why by hand? The browser knows the resolution of the output device, but neither its physical size nor the viewing distance – a projector reports the same as a monitor. So the level has to be set; on a student's device 100 % remains sensible.
Frame rateautomatic / 24 / 25 / 30 / 50 / 60 / 120. The most important lever for corrections – see section 7.1.
ModeStudent / Science, the data are kept.
Gravity gDefault 9.81 m/s². g is not measured, it is set – the value enters Eₚ = m·g·h only (and thereby Emech). Distance, velocity, acceleration, momentum, force and Ekin come entirely from the video and do not change with it.
So why change it at all? First, for videos that do not take place on Earth (Moon 1.62 m/s², Mars 3.71 m/s²) – the hammer-and-feather demonstration of Apollo 15, for example. Second, as a check exercise: the class reads a off a free-fall measurement or fits it, enters that measured value here and tests whether Emech then stays more nearly constant than with 9.81 – which brings home that what sits here is an assumption, not a measurement. Third, for location-specific work (g varies between 9.78 and 9.83 m/s²), though in school measurements that difference vanishes into the noise.
Accuracy over speedMakes the tracker seek exactly to every frame during the run instead of letting the video run through. That takes longer, but delivers the same data points on every device – without the checkbox, the speed of the computer has a say in how precisely the points sit. Off by default; the choice is remembered. Whether it took effect is shown by the balance after the run (“read while playing 0”).
SmoothingFour steps 3 / 5 / 7 / 9 (default 5): over how many data points the values are averaged before differentiating. Applies to all objects at once, and the analysis is recalculated immediately. The hint below states the corresponding time span in ms for the current frame rate – at 30 frames per second that is 100 / 167 / 233 / 300 ms. Larger windows calm v and a, but flatten short events (section 7.2).
Version, Legal noticeProgram version and the legally required publisher details together with the privacy statement.

6 The analysis tools in detail

6.1 Area calculation

Evaluation → Area, then drag across the range you want in the chart. The result box gives the value, the unit, the object concerned, the physical meaning and the range. The meaning depends on both axes:

ChartArea meansUnit
v(t), vₓ(t), vᵧ(t)distance traveled Δs – with a direct cross-check from s(t)m
a(t), aₓ(t), aᵧ(t)change in velocity Δvm/s
Fₓ(t), Fᵧ(t)impulse = momentum change Δpₓ or Δpᵧkg·m/s
Fₓ(x), Fᵧ(y)work Wₓ = ∫Fₓ dx (stretches traveled back count as negative)J
F(s)the work done, but only if the force acts along the path – otherwise a warningJ
anything elseno value, but a note on which combinations are meaningful–
Physics to watch: The areas under the magnitudes |a| and |F| are traps. Over one full period of an oscillation Δp = 0, while the area under |F| is clearly greater than zero; in circular motion the work done is zero, yet the area under |F|(s) is positive. The app warns about this in plain language – in class, work consistently with the components.
And: “F” is always the net force derived from the measured acceleration, never a single individual force. The area under Fₓ(x) is therefore the total work and, by the work–energy theorem, equal to ΔEₖ.

6.2 Fitting a model

Once a model has been chosen, sliders appear below the chart with rough starting values taken from the data – deliberately not the finished fit. The class brings the curve into line by hand, watching R² update live as they go. Auto-fit then computes the best solution and shows how close they got. Back to video only folds the sliders away; the curve and its parameters remain.

Reading R² correctly: R² measures only the scatter about the curve, not whether the model is right – a high value can go with a wrong model. Always judge it together with the picture of the curve; systematic deviations at the ends weigh more heavily than a value close to 1. The oscillation fit needs at least one full period within the measured range, otherwise it returns no result at all.

6.3 The stacked view

Replaces the single chart with up to three charts on top of each other, with derivative arrows in between (vₓ = dx/dt, aₓ = dvₓ/dt, v = ds/dt). The arrows deliberately appear only where the relation holds exactly: for the components and for the arc length. There is none between v and a – that is no oversight but the pointer to |a| ≠ d|v|/dt. The view shows only one object and offers neither area calculation nor fitting. On a phone the stacked view hides the video and uses the whole display area – otherwise each chart would be only a few centimeters tall; Mark object or Set scale bring the image back. On a PC the stack sits next to the video as usual.

6.4 The data table – two views

All calculated quantities of the active object as a table of numbers, row by row; the row of the current frame is highlighted, and tapping it jumps the video there. Columns in Science mode: t, x, y, s, h, vₓ, vᵧ, v, aₓ, aᵧ, a, F, Fₓ, Fᵧ, p, pₓ, pᵧ, Eₖ, Eₚ, Eₘ – the ones that depend on the mass and on the scale appear only when those prerequisites are met. In Student mode there are five columns (t, x, y, s, v).

In the menu Evaluation → group “Results” there are two buttons for this; both exist in both modes:

ViewLayoutWhat for
Table + video The table takes the place of the chart: on a PC video on the left, table on the right; on a phone video on top, table below. Check measured values against the image – tap an odd-looking row and the video jumps there and shows what happened at that point.
Table + chart The video gives way (as in fit mode): on a PC chart on the left, table on the right; on a phone chart on top, table below. From the chart to the number and back – reading off values, copying them into the exercise book, spotting outliers.

Pressing the same button again closes the table; the other one switches the layout. The stacked view and the fit sliders claim the same space – they therefore switch the table off (and vice versa). The red marker, the highlighted table row and the video frame stay in sync in every view.

Hide columns – so students work the values out themselves. A click on a column header hides the measured values of that quantity; the header stays, is struck through, and the column shows a fine hatching. A second click brings the values back.
Didactically this is the real point of the table: hide v and have the class determine it from x and t; hide a and have them estimate it from two v values; hide Eₖ and have them calculate it from m and v – then check at once with a click. The values only become invisible, they are not deleted: export, chart and calculations stay complete. The column shrinks to the width of its symbol, so that the columns behind it move up – on a phone this makes several quantities visible at the same time.

Set the split yourself. Between chart and table sits a splitter (vertical on a PC, horizontal on a phone). It divides the space freely between 15 % and 85 % – important in Science mode, where up to 19 columns sit side by side and you want to see them at the same time. A double click on the splitter restores the default. In the “Table + video” view this job is taken over by the splitter that already sits between the video and chart columns.

6.5 Strobe image and strip image

Combines 3 to 30 tracked frames into one picture – for the strobe image the cut-outs of the object are copied into the base frame, for the strip image vertical slices in time order are lined up next to each other. Optionally with the measured points drawn in. The PNG export delivers the full video resolution and suits worksheets on which students measure with a ruler.

6.6 Export

CSVOne file per object. Switch between “German (Excel)” (semicolon, decimal comma) and “International”. What is exported are the smoothed analysis values of the selected clip range in the current coordinate system – exactly what the charts show.
JSONA full dump of the app state including the raw points. Science mode only. Cannot be imported back.
.wod projectVideo and analysis in one file – the way to hand work in and to pass on ready-made examples. As large as the video.
On saving, Chrome and Edge open the system dialog: folder and name are yours to choose, straight onto a network drive if you like. Cancelling saves nothing, and no substitute download appears either. In Firefox and on iOS the app only asks for the file name, and the file then goes to the downloads folder. CSV and JSON still download immediately, without asking.

7 Limits I: measurement and physics

7.1 The frame rate is measured, not read from the file

Limit. The browser's video API does not supply the frame rate. WODAnalysis measures it while the video loads, over about 20 frames. If that fails, 30 fps are assumed instead – with a visible warning.
Why it matters: All time values depend on the frame rate, and with them every velocity, acceleration, force and energy. A motion recorded at 60 fps but analyzed as 30 fps yields half the velocities – without anything looking implausible.
What to do: After every video you load, check the fps value in the status bar against the camera setting and, if necessary, set it by hand under ⚙ → Frame rate. Videos with a variable frame rate (some phones) violate the basic assumption of equal time steps; choose a fixed rate in the camera app.

7.2 The second derivative is noisy, and stays noisy

Limit. a, F and the arrows that go with them come from differentiating the position data twice. A position error of one pixel is amplified by 1/dt² in the process. The app does smooth (Savitzky–Golay, window width adapted automatically to the frame rate) – but the noise does not disappear with it, it is only spread out. Under ⚙ → Smoothing the window can be set to 3, 5, 7 or 9 data points (default 5, applies to all objects at once; the analysis is recalculated immediately). The hint below states the corresponding time span in ms – at 30 frames per second that is 100/167/233/300 ms. Larger windows calm v and a, but flatten short events (a collision, a turning point).
What to do: Start with the recording rather than the analysis: steady camera, a high-contrast object that fills the frame, a tight box. When you care about a, 30 fps is often better than 60 – the finer time resolution amplifies the noise more than it adds in information. For quantitative statements about acceleration, determine the slope in the v(t) chart with a curve fit rather than reading off single a values.

7.3 |a| is not the derivative of |v|

In uniform circular motion the magnitude of the velocity is constant (dv/dt = 0), while the magnitude of the acceleration is v²/r > 0. The derivative relation holds exactly only for the components (x → vₓ → aₓ) and for one-dimensional motion. The app shows this indirectly (no derivative arrow between v and a in the stacked view), but does not say it explicitly on the magnitude chart – this point has to come from you, before a circular motion or an oscillation is analyzed.

7.4 The height for Eₚ always follows the vertical in the image

Even with a rotated coordinate system (inclined plane) the app measures the height for Eₚ = m·g·h vertically in the image, not along the rotated y axis. That is the only way mechanical energy stays conserved. The zero line of Eₚ is therefore a horizontal line through the origin – not the x axis as drawn. The app shows a note about this the first time an energy chart appears.
Consequence for recording: a camera held at an angle makes the vertical in the image the wrong vertical – Eₚ and Eₘ are then no longer correct, and nothing on screen gives it away. Rest the device against something or put it on a tripod; as a check, compare a vertical edge in the image (door frame, table leg) with the axes.

7.5 Spatial resolution, perspective, reference plane

8 Limits II: what the app cannot do

Missing / limitedWhat it means for teaching
No chart export as an imageNeither the chart nor the stacked view nor the table can be saved (only the strobe image, as a PNG). Use screenshots for worksheets – set round axis limits via Chart range beforehand.
Fit parameters and area values are not exportedR², model parameters and area results appear on screen only. They belong in the lab write-up – that is pedagogically defensible, but you have to tell the class.
The center-of-mass curve is not exported anywhereFor a collision experiment, export the individual objects and work out the center of mass yourself in a spreadsheet.
No JSON importFor backups and for passing work on, always use Save project (.wod).
Individual points cannot be deletedDrag stray points to the plausible position; if a whole section is unusable, set the clip range more tightly and track again.
Points placed by hand are always greenThe color coding rates the automatic detection only. Corrected points therefore look more reliable than they are – say so in class.
Tracking again deletes the manual correctionsOrder: track first, then correct. Save the project before making bigger corrections.
Only four model functionsExponential growth, power laws or hyperbolas cannot be fitted. Tailor the questions accordingly.
At most three objects, not removableObjects created by accident stay in the list (but are only a visual nuisance).
“Restart” (⚙, since version 95)After a confirmation prompt it deletes the saved state and starts empty; the language choice is kept. On shared devices, introduce this as a ritual at the end of the lesson.
No search-radius or threshold sliderThese values adapt automatically; contrary to older documents there is no way to set them.
One autosave state per deviceThe autosave is crash protection, not a filing system: it stores only the session last worked on, and never the video. For several analyses, create a separate .wod file for each one.
There is no lossless way back from the app to the start screenMode, Language and Legal notice are reachable via the gear icon; the student handout only from the start screen – so open it beforehand. (“Restart” does lead to the start screen, but discards the analysis.)

9 Limits III: devices, network, school computers

9.1 Devices and videos

9.2 The PC link – and why it can fail on a school laptop

The PC link transfers the video and the data points from the phone to a second screen: one device starts the session and shows a six-character code, the other types it in. The data travels directly between the devices; a small signaling server on the network only brokers the connection and never sees video or measurement data.

Case study: a managed school laptop

Observation: two browser windows on the same laptop link up without trouble – laptop and phone never find each other, not even on the phone's hotspot. The app's diagnostics (the browser console) showed:

[WODAnalysis] Verbindung failed: lokaleKandidaten: { host: 1, srflx: 1 } entfernteKandidaten:{ host: 2, srflx: 2 } ice: "checking" → verbindung: "failed"

How to read it: both devices had exchanged complete addresses (host = local address, srflx = public address), yet connection setup failed on every path. Firefox says so explicitly: “ICE failed, add a TURN server”.

Causes on managed machines – usually several at once:

Important: the first two points are down to the device, not the network – which is why switching to a phone hotspot does not help either.

What the app does: if no connection is established 20 seconds after the code was exchanged, it says so in plain language, lists the likely causes and appends the short candidate summary – for example [own addresses: host×2 srflx×1 · other side: host×2]. If srflx is missing there, outbound traffic is blocked; if the other side is missing entirely, the connection request never got through.

What the app deliberately does not do: it uses no relay (TURN). Such a detour would get past even strict firewalls, but it would route all payload data through someone else's server – which contradicts the principle that videos never leave the device. So if the devices cannot find each other directly, there is no fallback route.

What works in practice in the classroom: (1) on private devices and on a home network the PC link usually works. (2) On a school computer use the file route: save the project (.wod) or a CSV and pass it on – that always works and needs no network. (3) If you want the PC link to work permanently on the school network, you need your IT admins: an exception that allows the local addresses for exactly this one web address, and an inbound firewall rule for the browser, limited to the local subnet.

9.3 Other quirks of the PC link

10 Data protection

In normal use nothing leaves the device. No account, no login, no upload, no usage statistics. The video remains a file on the device and is never stored anywhere by the app; measured values and settings live in the browser storage. CSV and JSON exports go to the downloads folder, and for a project file Chrome and Edge let you choose the folder.
Only while pairing is active does the app talk to the signaling service on wodanalysis.info – and all it sends there are connection details (addresses and session code), never video or measurement data. To learn its own public address the browser also queries a STUN service run by sipgate or Deutsche Telekom (servers in Germany) – these too see only addresses, never content. Video and measurement data travel directly between the devices and are encrypted on the way; since version 95 the video only sets off after an explicit confirmation (video + data points, data points only or send nothing). If you want to rule all of that out, do not pair; all other functions remain fully usable.
For recordings showing recognisable people the usual school rules apply. Keep in mind: when pairing, the video moves (after confirmation) to the paired device, and a project file (.wod) contains the video – so before passing either on, check who appears in it.

11 Troubleshooting

SymptomCause and remedy
Values in px instead of mScale missing. Check the status bar (the px/m reading) and run Set scale.
Momentum, force or energy missingMass missing (with two objects, for both of them!). Energies additionally need the scale.
Chart stays emptyThe selected quantity does not exist for this object – or the series has been hidden in the legend. Briefly switch to a quantity that has a legend and show the series again.
All velocities twice/half as largeFrame rate detected incorrectly. ⚙ → set the frame rate by hand.
Curve jumps at the startOutlier in the first frame. Correct the point or let the clip range start later.
Tracking loses the objectDraw the box tighter and around a higher-contrast area; if need be, stick a colored dot on the object. Reduce motion blur with more light.
Video does not loadCodec (usually HEVC). Analyze it on the recording device or record it again in H.264.
A click in the chart no longer jumps to the frameArea mode is active – switch Area off.
The axes keep slippingTap the lock 🔒 to close it. Do it again after every session restore and every time you open a project, because the locked state is not saved.
The devices can't find each otherSection 9.2 – and when in doubt fall back on the file route.

12 Suggested experiments (tried and tested)

ExperimentAnalysis in WODAnalysis
Free fall / vertical throwy against t with a quadratic fit → g from the parameter a (y = a·t² + …, g = −2a). Cross-check: aᵧ in the chart, Eₖ + Eₚ = Eₘ ≈ constant.
Horizontal projectile motionThe trajectory as y against x (axes freely selectable), plus x against t (linear) and y against t (quadratic) – the independence of the components becomes directly visible.
Collision of two cartsTwo objects with masses; pₓ of both carts, sum them by hand; show the center-of-mass curve – it moves on undisturbed through the collision. Area under Fₓ(t) = impulse, area under Fₓ(x) = work.
Mass–spring oscillatorx against t fitted with the oscillation model, or the damped oscillation model where there is damping; read off T, and from it the spring constant D = m·(2π/T)². Film at least two to three full periods.
Simple pendulumAs for the mass–spring oscillator; in addition the energy conversion Eₖ ↔ Eₚ with the lowest point as the origin.
Circular motionShow the path points; the a vector points to the center. The textbook case for |a| ≠ dv/dt – see 7.3.
Inclined planeRotate the axes (x along the path) and lock them; aₓ constant. Note: Eₚ still measures the height vertically in the image – exactly as it should.

WODAnalysis runs entirely in your students' browsers: no login, no upload, and after the first visit even without internet. This handout describes program version 95; the version shown on the start screen should match it. Please send feedback and bug reports with the version number and – for pairing problems – with the short candidate summary from the status bar.