The paper presents a structured lighting system for generating high-resolution stereo datasets of static indoor scenes with subpixel-accurate ground-truth disparities. The system includes novel techniques for efficient 2D subpixel correspondence search and self-calibration of cameras and projectors, including modeling of lens distortion. By combining disparity estimates from multiple projector positions, the system achieves a disparity accuracy of 0.2 pixels on most observed surfaces, including in half-occluded regions. The authors contribute 33 new 6-megapixel datasets and demonstrate that these datasets present new challenges for the next generation of stereo algorithms. The datasets are available at http://vision.middlebury.edu/stereo/data/2014/. The paper also discusses the processing pipeline, including image acquisition, decoding and interpolation, fast 2D correspondence search, filtering and merging, calibration refinement, and self-calibration of projectors. The authors test their new datasets using three state-of-the-art stereo methods and show that the datasets provide a range of challenges that significantly exceed those of existing datasets.The paper presents a structured lighting system for generating high-resolution stereo datasets of static indoor scenes with subpixel-accurate ground-truth disparities. The system includes novel techniques for efficient 2D subpixel correspondence search and self-calibration of cameras and projectors, including modeling of lens distortion. By combining disparity estimates from multiple projector positions, the system achieves a disparity accuracy of 0.2 pixels on most observed surfaces, including in half-occluded regions. The authors contribute 33 new 6-megapixel datasets and demonstrate that these datasets present new challenges for the next generation of stereo algorithms. The datasets are available at http://vision.middlebury.edu/stereo/data/2014/. The paper also discusses the processing pipeline, including image acquisition, decoding and interpolation, fast 2D correspondence search, filtering and merging, calibration refinement, and self-calibration of projectors. The authors test their new datasets using three state-of-the-art stereo methods and show that the datasets provide a range of challenges that significantly exceed those of existing datasets.