HDR doesn't have to be hard
summary
Claim: HDR is overpowered for storytelling and that 10 stops (1000:1) is the practical instantaneous range viewers use, not the trillion‑to‑one theoretical peak. Evidence: demos with a pro‑display XDR and camera auto‑exposure show a ten‑times brightness sweep barely changes perceived image, while old SDR TVs crush shadows and blow highlights. Proposal and tool: expose a default SDR LUT during HDR grading, embed it with the file so devices stop fighting creators with aggressive dynamic contrast. It’s simple, interoperable, and should reduce arbitrary program‑to‑program variation, though adoption takes years.
AI-written summary · about
chapters (7)
- 0:00Start
- 1:51Dark Adaptation
- 2:44The Predictive Processing Model
- 3:29The Brain Has a Really Strong Opinion of the Way Light and Shadow Works
- 5:48Transition from a Bright Shot to a Really Dark
- 12:01Dynamic Contrast
- 14:51Closing Thoughts
related talks
20mUnderstanding HDR
26mDelivering HDR
30mFocus on Imaging: Lenses, Cameras, and Slo-Mo Replay
13mPractical HDR: Not 101, but Flavors, Pros, Cons, Dos and Don’ts
47mHDR10+ Contents Generation Tools and Workflow
41mHandling HDR Media on Android Without Breaking It
31mLights, Colors, Artwork
12mApproach to HDR and Tonemap on Android