• A case of several scans: up to 600 frames per session (the iPhone computes a preview from the best 200 — Apple's limit, the Mac all of them), library grouped by case, a detail scan assigned to a finding, the report lists the other scans of the case • Guide “Crime-scene examination: body and surroundings” (About, web) • Targets T3.1/T3.2: the QR code drawn as vectors — crisp in print • New measurement kind Polyline: the sum of straight segments (body length with flexed limbs: crown–hip–knee–heel); “along the surface” is shown for Distance only up to 0.5 m

• Accurate measuring: measurement and finding points always lie on the model surface (up to version 3.0 they could land on the model hull on iPhone). We recommend re-checking older iPhone measurements. • Fine-tune: 3× loupe and precise nudging; press-and-hold placement fixed. • Lasso: trace an outline along the skin (path, area). Undo/Redo, custom finding colour, marker size. • Capture in four guided steps: overview, detail + target, fill gaps, surroundings. • Large scans: quick preview from a frame subset, Mac textures downsized, diagnostics after a system kill; package export for the Mac. • Target T3 in two print versions (T3.1 clean, T3.2 print); the reference check reports a mismatch.

• Metrologically substantiated scale: the measurement scale is always derived from the camera path (LiDAR + IMU) and from the target in the model — a photogrammetry model is not in metres by itself; the viewer and the report state the correction and its uncertainty • Uncertainty with every measurement: value ± U (k = 2) from the scale, frame resolution and input method (finger, Apple Pencil) • New calibration target T3 (A4 landscape): circular fiducials the app finds even from 2 m, triangulation in the model independent of LiDAR, print verification with a caliper, cut-out for the ColorChecker Classic Mini; the older T2 target still works • More precise point placement: markers with a fixed on-screen size, a loupe on press-and-drag, Apple Pencil including hover • Capture as a measuring process: sharpness of every frame, rejection of blurred frames, photo mode waits for the phone to be still, depth only when synchronous with the image, coach hints “Target too far” and “Frames are blurred” • Target control network: with several targets, tape distances between their centres as a long scale baseline; validation CSV export • Certificate: class A–D by the new method (scale from the camera path, independent check by the target in the model), verification of findings and surroundings hashes on import • Package format 2 (backwards compatible); unit tests for the maths and the format

• Findings transfer after a recompute — pins and measurements are carried over to the new model by the camera positions (error usually a few millimetres; stated in the viewer and in the report). The original file is kept aside, measurement values do not change • Works for packages recomputed in Fotograme3ka 3D Calc on the Mac too • The .f3d package carries the camera positions (model/poses.json) — a recompute anywhere keeps the findings • Support development (Donor, Pro) directly on the main screen