A year ago, we demonstrated a major breakthrough that enabled our image processing algorithms to do the fine-tuning themselves, and completely automatically. In most cases, those advanced reduced complicated digital film restoration processes to the click of a single button. But behind this click there were four or more successive passes of processing. The restorations were automatic, but required a sequence of processing routines to be run one after the other. We knew that our mission to make restoration simple enough and affordable enough for the masses required more work. The restorations had to require less computer power and less computer storage space.
During the last year of our research, we have found a way to combine all the sequential image processing steps into a single combo-pass. Applying and extending state-of-the-art algorithms being used in the field of “big data”, we achieved not only powerful processing acceleration and storage saving, but we achieved a surprising additional result. The image quality of the final results was even better than before.
We also added the ability to re-grain the restored footage using a fraction of the original natural grain after the unwanted imperfections and artifacts had been removed automatically, and all in this same combo-pass.
This new technology is very important to motion picture archives because up until now, digital restoration has only been practical for a relatively small number of films. There are massive quantities of content in archives around the world, and without high levels of automation and efficient processing; restoration is beyond the practical timescales and budgets of many.
The bulk of our tests to date show that this new technology makes highly automated, high-quality image improvement possible for a very wide spectrum of motion picture footage – from the earliest cinematographic paper prints to footage shot with the latest digital professional and consumer cameras. These new algorithms are fast too- with processing speeds approaching real-time at 2K when running on powerful NVidia GPUs.