Watch a video and ask yourself a simple question: How good is the quality of the image?
Like most people, you’ll no doubt easily and quickly form a simple answer. You might say, “it’s good,” or “excellent,” or perhaps “just ok,” or maybe even “dreadful.” (Hopefully not dreadful!)
But how would a computer answer the same question? How can a machine perceive the way a person perceives? Where would it even begin?
The answers are important. In today’s connected, exponential, at-scale world, the ability to predict quality, monitor quality, ensure quality, as thousands of image streams are fed through the internet simultaneously to millions of viewers’ devices and TV screens, is vital if a company is to ensure that its viewers receive a product that makes them say, “This is good.”
The following collection of stories and posts aims to explore this very problem. They are for the benefit of the video engineers, architects, artists and business professionals who are responsible for delivering motion pictures and television to the screens of audiences.
These stories will explore all aspects of the video delivery pipeline — from post-production, encoding, transcoding, packaging, CDNs — right through to the playout on devices.
The posts feature the views of two of the world’s leading researchers and practitioners in the field, SSIMWAVE’s Dr. Abdul Rehman and University of Waterloo Prof. Zhou Wang. Their work currently affects the television viewing experiences of tens of millions of people every day.
Collectively, we hope the series will begin to answer some of the urgent questions surrounding the business, the complexities and the realities of video quality.