QuantCinemaCinema, decoded.
FilmsRankingsCompare
QuantCinema

AI-powered cinema analytics.

DISPATCH

Suggest a film · get notified when new data drops

© 2026 QuantCinema
DEEP DIVE · DUNKIRK

The full breakdown

Detection, audio and scene analysis across every frame of Dunkirk — derived from object recognition, transcript alignment and shot segmentation.

← Back to Dunkirk
2017War, Action, Drama106.6 min● analyzed May 29, 2026
Total Detections
75,247
objects + faces
Unique Objects
94
distinct types
Detections / min
705.6
avg rate
Night Scenes
54.2%
low-light frames
Words Spoken
2,685
transcribed
Named Actors
15
face-id matched
SIGNAL CHART

Motion · Brightness · People — across 106.6 minutes

All three series normalized to 0–100. Hover to scrub the runtime.

Motion · spikes = action
Brightness · dips = night
People · crowd density
0:00 → 106.6m · all series normalized 0–100
02550751000m27m53m80m107m