LG Electronics Announces U.S. Availability Of First 84-inch Class \'Ultra-Definition\' TV

[Shane]

LG Electronics today announced U.S. availability for its enormous 84-inch (84.04 inch diagonal) LED display, boasting a native \"ultra definition\" (UD) screen resolution of 3840 x 2160 to match its impressive size*.

The gigantic 84-inch screen quadruples the level of detail from full 1080p HD resolution to a massive 8 million pixels. Even before so-called \"4K\" content is available, LG\'s proprietary Resolution Upscaler Plus delivers higher detail from current HD/SD external sources. A major step forward for the display industry, ultra-definition TV makes use of the screen real estate available with today\'s largest-size flat panels, much like the jump from 720p to 1080p years ago. The LG UD 84LM9600 is expected to be available in October at an MSRP of...

Read Bulletin

[johnsidle]

"LG's proprietary Resolution Upscaler Plus delivers higher detail from current HD/SD external sources."

How is this possible? It seems to me that HDTV signals control individual pixels within a 1080 panel; a pixel being the smallest area of detail. If the number of pixels is increased by a factor of four, the signal that set one pixel now will control four pixels which will be identical. Sitting at the same distance from the screen, the group of four pixels will be the same size as one pixel in a comparably sized 1080 display yielding the same definition.

Marketing hype?

[Rodolfo]

"LG's proprietary Resolution Upscaler Plus delivers higher detail from current HD/SD external sources."

How is this possible? It seems to me that HDTV signals control individual pixels within a 1080 panel; a pixel being the smallest area of detail. If the number of pixels is increased by a factor of four, the signal that set one pixel now will control four pixels which will be identical. Sitting at the same distance from the screen, the group of four pixels will be the same size as one pixel in a comparably sized 1080 display yielding the same definition.

Marketing hype?

Size wise you may assume that, if the 4K panel is of the same size, the 4 small pixels would occupy the space of one in 1080p, but the 3 interpolated pixels added to total 4 are not the exact copy of the pixel they replace, but rather contain detail that is video processing calculated to blend well with the surrounding pixels, like motion calculation does when adding video frames to smooth out 60fps video displayed in a 120fps cadence.

The pixels are still a man-made creation to the best a mathematical model in the video processor engine can do though.

It is done in a similar way 480p DVD was upscaled to 1080p (6 times the pixel count per video frame). How accurate is that? it is getting better as newer video processing technology is introduced every year.

When I activate the "Reality creation" feature in my 4K projector (for a similar objective) the image is stunning, not because I can not see the small pixels, but because of its impact of "perceived reality", as studied by the UHDTV TV engineers in Japan when analyzing the effects of UHDTV and the human reaction to it.

Best Regards,

Rodolfo La Maestra

[johnsidle]

That makes sense that the pixel edges are processed so that they smoothly transition to the adjacent pixel.
Thanks for the explanation.

[Roger Halstead]

"Up Scaling" can be done in a number of ways. The most basic, crudest, easiest, fastest and cheapest is to treat 4 as one. Trying to keep it simple and using a failing memory, another and much better way is to compare the pixel against the 8 adjoining ones, then see if any of them are an edge. If not the 4 (basic) or 8 (mathematically better) "up scaled" pixels are interpolated, often using FFT (Fast Fourier Transforms) to form a smooth gradient between the 4 or 8 up scaled pixels and the adjoining ones. There are 8 adjoining pixels for a 4 sided pixel because you need to look at those off the corners as well as the sides. This adds up to a lot of processing, but can result in a seamless and clear 4X to 8X up scaling. The hardware and software combination capable of doing this much processing has not been around for long. IIRC (It's been a long time since grad school!) The FFT typically involves the Derivative of a series of signs and cosigns, performing some manipulation and then doing the integral of the series to get back where you started with the results. In effect they can compare the specific pixel to adjoining ones or much farther out. To maintain a smooth transition they also have to do edge detection and we get into some very complicated algorithms and hardware intensive operations.

This is the same kind of processing done on the Voyager images which were completely black before processing. At that time it took a top end 386 with a math coprocessor about 10 to 12 hours to process a portion of one of these images that was only about 3/8ths of an inch per side. Although not cheap, now days a multi-core video card can pretty much do this in real time. Whether this is similar to LG's approach, I don't know.