Computational Photography, an AI-powered Slopendium

▸Part 17 HIGH DYNAMIC RANGE CAPTURE⧉

▸17.1 Dynamic range and multiple-exposure capture⧉

17.1.1 The gap between the scene, the sensor and the screen⧉

17.1.2 Bracketing, and which parameter to vary⧉

17.1.3 Doing it in the sensor instead⧉

17.1.4 Two caveats: single-shot range, and where this is used⧉

▸17.2 Camera response and radiance calibration⧉

17.2.1 The linear case: recovering the exposure ratios⧉

17.2.2 The non-linear case: recovering the curve⧉

17.2.3 The pipeline end to end⧉

17.2.4 From exposure to photometry⧉

▸17.3 Merging exposures⧉

17.3.1 The weighted merge⧉

17.3.2 When every exposure is bad at a pixel⧉

17.3.3 What comes out is not a picture⧉

17.3.4 Optimizing the capture and merge⧉

▸17.4 Ghosts and motion⧉

17.4.1 Align first, and know what alignment cannot do⧉

17.4.2 Rejection: pick a reference and disagree with it⧉

17.4.3 The other approaches, and why rejection is the usual choice⧉

17.4.4 The bracket you did not have to take⧉

▸17.5 Single-shot range and inverse tone mapping⧉

17.5.1 Range without a bracket, in the sensor⧉

17.5.2 Inverse tone mapping: inventing the range⧉

17.5.3 Choosing among the three routes⧉

▸17.6 HDR+ and burst imaging on phones⧉

17.6.1 Why a phone shoots a burst, and why it underexposes⧉

17.6.2 The HDR+ pipeline: align and robust-merge in raw⧉

17.6.3 From burst HDR to burst super-resolution⧉

▸17.7 Encoding HDR images⧉

17.7.1 How many bits does a range need?⧉

17.7.2 Floating point: half floats and OpenEXR⧉

17.7.3 Shared exponents and logarithms: Radiance RGBE and LogLuv⧉

17.7.4 Editable HDR: floating-point DNG and TIFF⧉

17.7.5 Backward-compatible HDR: from JPEG-HDR to gain maps⧉

17.7.6 Display-referred encodings: PQ, HLG and video⧉

17.7.7 Choosing an encoding⧉