GPP Conditioning

In our contact with the broadcast world we have observed that while the 1990's and early part of this decade were dominated by interest in visual communications infrastructure, the recent trend is towards better picture quality. This has been driven in part by the emergence of HD formats in consumer devices but also because the consumer has become more discerning since display technology for digital pictures has improved. The problem is that broadcasters, especially over the internet are caught in that better picture quality means bigger filesizes which requires more drivespace. While drivespace is getting cheaper, it is still finite and large filesizes always means slower read times. With that in mind we have been developing technologies that allow picture quality to be maintained or improved along a digital broadcast chain particularly over the internet, while reducing filesizes.

We target two requirements: the need to keep visual file sizes low in the face of ever increasing user uploads to large repositories and the need to maintain picture quality in the face of increased compression. Our researchers have recognised that simply by pre-processing content for small amounts of noise and sometimes unwanted camera movement, we can achieve at least 10% reduction in filesizes or bandwidth without affecting picture quality and even improve picture quality. We call this process Visual Conditioning.

Our GPP Conditioning toolkit (called GPP-PilotFish) contains primarily our flicker removal, noise reduction and stabilization algorithms. It can be deployed as a standalone command line "crawler" application that would touch all your visual files, clean the content, and re-encode into the same or a different format. The resulting database will be smaller and the visual quality will be at least as good and most times improved. Our technology can also be deployed inside clients' own crawling software using our GPPEngine manifestation.