IAI CFVE study materials
IAI CFVE study resources

IAI CFVE: Forensic Video

IAI CFVE · Work through your exam, one topic at a time.

All guide sections

Forensic Video

Forensic Video is the single domain tested by the IAI CFVE examination. It covers the technical and procedural knowledge needed to acquire, preserve, analyze, and present video evidence: analog and digital signal fundamentals, codecs and compression artifacts, DVR/NVR export, hashing and chain of custody, frame timing, measurement, photogrammetry, comparison, and courtroom reporting. Expect questions that test why a method is defensible, not just what button to press. A recurring theme is least-destructive acquisition followed by repeatable analysis.

Core rules

  • Preserve the original bitstream before any playback or transcoding; hash at acquisition and again after transfer. Work from a write-blocked copy.
  • Distinguish capture rate from playback rate. Interlaced 1080i29.97 carries 59.94 fields/s, not 59.94 frames/s; PAL is 25 frames/s (50 fields/s); NTSC 29.97 timecode may be drop-frame or non-drop; confirm which.
  • Record the DVR clock offset against a trusted reference and state the uncertainty; never treat an on-screen timestamp as ground truth.
  • Calibrate measurements with in-scene references and report units, method, and error. Use same make/model/firmware/lens/lighting for comparison when possible.
  • Follow SWGDE best practices and Daubert/Frye admissibility principles: testability, peer review, known error rates, general acceptance.

Worked example

A 25 fps PAL camera records a vehicle crossing two painted marks 18.0 m apart. The vehicle’s rear bumper is at mark A on frame 120 and mark B on frame 165. Time = (165-120)/25 = 45/25 = 1.80 s. Speed = 18.0/1.80 = 10.0 m/s = 36.0 km/h. The examiner documents frame numbering, measurement endpoints, and uncertainty. If the DVR clock is 2 min fast, that affects timeline correlation, not the calculated speed.

Traps

Treating a player’s “30 fps” display as the capture rate; deinterlacing and reporting doubled frame counts; measuring from re-encoded frames; assuming square pixels; ignoring GOP artifacts; failing to document hash values, tool versions, and DVR export settings; and asserting a positive identification from a few pixels.