The EuclidIQ team had a great Streaming Media East 2016. From a talk on subjective testing by Dr. Nigel Lee, to a record number of booth visitors and Visual Challenge test takers, the annual show held in the New York Hilton Midtown was a resounding success for EuclidIQ.
On the show floor, visitors to EuclidIQ’s booth could discuss our IQ264 perceptual quality optimization technology that provides improved H.264 encoding or take the EuclidIQ Visual Challenge, evaluating the compression benefit of our IQ264 against reference x264 encoding.
More than 60 attendees took the Visual Challenge. While our custom testing lab at EuclidIQ’s headquarters in Concord, MA, is the ideal viewing environment for subjective testing — with proper lighting and significantly less noise than that of a busy exhibit hall — we are pleased to report that a large majority (about two-thirds) of attendees rated IQ264-augmented content as having higher quality than the reference encoding in the Visual Challenge’s double-blind subjective test (another 22% rated the two the same). Post-test analysis on the subjective test mean opinion scores estimates the compression benefit of IQ264 in this test at 21% (bandwidth reduction for equivalent quality, relative to reference x264). Again, the Visual Challenge was conducted under less-than-ideal conditions; we have seen gains above 30% in internal testing under more controlled conditions for the same videos.
Meanwhile, in the Discovery Track, our Chief Science Officer, Dr. Nigel Lee, gave a presentation on “The How and Why of Using Subjective Testing for Perceptual Quality Optimization.” The presentation slides can be found here and a video of Dr. Lee’s presentation should be online soon here.
Dr. Lee discussed how no current objective testing metric and underlying toolset is capable of discerning quality the way human viewers can, which is why EuclidIQ uses subjective testing to not only measure the gains from IQ264-augmented encodings but also to inform the further development of IQ264 technology.
Commented Dr. Lee, “Accurate evaluation of the benefits of perceptual quality optimization technology requires subjective measurement of the gains, and we have developed a practical way to conduct true mean opinion score [MOS]-based subjective testing for this purpose. There are all kinds of claims about compression gains in the industry, so we feel it is important to be clear how we measure the gains that we claim.”