Video forensics from container structure to individual frames. We measure temporal consistency, compression history and noise behaviour, and report exactly which frames and time ranges raise concern and why.
Current
Teams watch a clip, trust their instincts and argue about artefacts. Findings vary by reviewer and are hard to explain when challenged.
Target
Sensors measure each frame and each segment. The report shows the measurement, the threshold and the time range, so a second examiner can check the same result.
01Preserve
The original file is hashed and sealed before any frame is decoded.
02Extract
Frames, motion and encoding structure are extracted from a working copy.
03Measure
Independent sensors score temporal, noise and compression behaviour.
04Explain
Findings are placed on a timeline with the evidence beside each one.
Video Analysis Chain
Control planeAnalysis plane
01Container Parse
STREAMS · METADATA
Reads container structure, codec parameters and timestamps without decoding more than needed.
02Frame Extraction
KEYFRAMES · GOP
Extracts frames and group-of-pictures structure so compression history can be examined.
03Temporal Sensors
MOTION · NOISE
Measures motion, noise and lighting consistency across frames and across face regions.
04Compression Trace
DOUBLE ENCODING
Looks for signs that segments were encoded differently from their neighbours.
05Frame-Level Report
TIMELINE · EVIDENCE
Marks the frames and time ranges where findings occur, with the measurements behind them.
Control plane / CUSTODY LOG
Case Ledger
Records the original hash and every extraction step.
Real footage changes smoothly from frame to frame. Manipulated regions often do not.
MOTION · LIGHTING · FACE REGION
Check what should stay stable.
Every encoder leaves a pattern. Edited segments can carry a different one.
GOP STRUCTURE · DOUBLE ENCODING
Read how the file was made.
Camera sensors leave a faint noise pattern. Synthetic or pasted regions rarely match it.
NOISE RESIDUAL · CONSISTENCY
Compare the noise, not the picture.
Timestamps, device tags and software fields tell a story that the pixels may not.
STREAMS · TIMESTAMPS · TAGS
Trust nothing, record everything.
INVESTIGATION
A sensor sweep that tells an examiner where to look first and what to document.
PUBLIC SAFETY
Rapid triage of circulating video, with a clear note on confidence and limits.
FINANCIAL
Recorded verification calls checked for face-swap and replay artefacts.
Analysis runs on your hardware. Reports are exported as signed PDF and machine-readable JSON, with the parameters needed to reproduce each result.
Video forensics from container structure to individual frames. We measure temporal consistency, compression history and noise behaviour, and report exactly which frames and time ranges raise concern and why.
Teams watch a clip, trust their instincts and argue about artefacts. Findings vary by reviewer and are hard to explain when challenged. Sensors measure each frame and each segment. The report shows the measurement, the threshold and the time range, so a second examiner can check the same result.
Yes. Analysis runs on your own hardware, and the platform can be installed with no connectivity. Reports are exported as signed PDF and machine-readable JSON.
Request a briefing by form, WhatsApp (+60 17-773 7302) or email (kumar@spearcompute.com). We scope a pilot on samples you provide.
Tell us what media you need to trust and where it has to be analysed. We will scope a briefing or a pilot on your own samples.
Request a briefing