No abstract

High‐resolution display bandwidth requirements often now exceed the capacity of display link channels necessitating compression. The goal of visually lossless compression codecs such as VESA DSC 1.2 is that viewers perceive no difference between the compressed and uncompressed images, maintaining long‐standing expectations of a lossless display link. Such low impairment performance is difficult to validate as artifacts are at or below sensory threshold. We have developed a 3D version of the ISO/IEC 29170‐2 flicker paradigm and used it to compare the effects of image compression in flat images presented in the plane of the screen (2D) to compression in flat images with a disparity offset from the screen (3D). We hypothesized that differences in the location and size of the compression errors between the disparate images in the 3D case would affect their visibility. The results showed that artifacts were often less visible in 3D compared to 2D viewing. These findings have practical applications with respect to codec performance targets and algorithm development for 3D movie, animation, and virtual reality content. In particular, higher compression should be attainable in stereoscopic compared to equivalent 2D images because of increased tolerance to artifacts that are binocularly unmatched or have disparity relative to the screen.

Stereoscopic display technology provides immersive experiences in VR/AR/XR, but requires markedly higher bandwidth and is perceived differently than 2D content. Here we adapt the ISO/IEC 29170‐2 flicker paradigm for subjective assessment of low impairment stereoscopic image compression. We compared the performance VESA VDC‐M codec on stereoscopic images with 2D image performance.

In augmented and virtual reality (AR/VR), magnifying display lenses typically introduce pincushion distortion and transverse chromatic aberration. These distortions are corrected by pre‐processing, one side‐effect of which is disruption of the spatial correlation between color channels. As a result, the standard practice of performing color space conversion prior to image compression may introduce undesirable, visible artefacts. To assess this, we evaluated the performance of two low impairment display stream codecs on distortion corrected stereoscopic images based on converting to YCoCg color space or bypassing the color conversion.