Framing a Shifting Image

TechKinetic offers aspect conversion


Since the digital wave first began to splash against television, there's been no shortage of debate and speculation regarding its ultimate impact on the video screen. Somewhere at the top of the list came the challenges regarding aspect conversion. Letterboxing was a short-term answer and has been adopted to broad if mixed response; other techniques, such as stretching and zooming SD images to conform to an HD world, fall short to nearly all but the most myopic.

As the digital transition continued to develop around these techniques, Anthony Beavers, president of TechKinetic LLC, says he experienced a moment he did not immediately liken to the inventor's "a-ha!" of revelation, or a light bulb incandescing. To the contrary, looking at the solutions being used to address aspect conversion issues, he says that his light bulb dimmed. "The degraded images of the older techniques are in direct contrast to the spirit of HD and DTV, which includes superior quality video images," he said.

Perceiving that flaw in the process and observing no other solutions being introduced - Beavers - a Shreveport, La.-based entrepreneur with several U.S. patents pending for products and systems in the casino industry - began to reconsider a concept he'd previously envisioned. As a result, TechKinetic began turning its focus toward developing its trademark TVa Variable Aspect technology with the goal of addressing this broadcast shortcoming.


Beavers says that a fully deployed TVa technology can serve as a "universal translator" between various aspects and enables the optimal display of media of any aspect on display devices of any differing aspect.

"There are a number of disadvantages with the current approaches to multiformat, multidelivery issues," says Beavers. Despite letterboxing's status as the short-term jab at the 16:9/4:3 dragon, the approach is less than ideal when considering the black bands that represent wasted space to many viewers who still haven't paid the bill on their 35-inch screens.

Beavers also faults auto-zoom and image-stretching techniques currently used by manufacturers to provide aspect ratio conversions. "Auto-zoom clips an image without considering its composition, and stretching results in image distortion," he says. "They [manufacturers] try to enhance stretched images by minimizing the distortion in the center portion of a screen, while increasing distortion in the outer regions."

To Beavers, the key flaw surrounding these earlier conversion methods is their passive nature. "The problem is that manufacturers are responsible for the conversion process as opposed to content creators," he says.


A better solution, according to Beavers, is to bring content creators more directly into the ball game by enabling them to describe an absolute framing position that will translate from any original aspect to any desirable aspect with predictable results. Beavers' variable-aspect display system - which he has dubbed TVa - does this by providing a relative pan description for each frame (or GoP, depending on the application) of an encoded video stream.

(click thumbnail)TVa Variable Aspect Technology finds the optimal display.
Behind Techkinetic's TVa proposal is the proprietary core technology called the Relative Pan Descriptor Method (RPDM), which provides a relative pan description that is used to extract a desirable portion of an original image frame. Beavers says that the RPDM is based on a system of percentages that takes advantage of a common relationship between differing aspects.

"TVa's core algorithms essentially generate an offset value in the form of a percentage, which in an MPEG program or transport stream is provided as private data," he explains. An encoder enables information for three cardinal axes (x, y and z) to be linked to an original program. "At the receiver end, a decoder can determine what portion of the image should be extracted from the original frame, regardless of the original aspect or desirable aspect of conversion. A viewer sees what the director or editor intends for them to see."

Zoom-ratio data, along with x and y data, can be used to enlarge a particular area of interest within an image. This feature can further enhance media presentations on small-screen devices, where screen space is particularly valuable, says Beavers.


MPEG anticipated the use of encoded pan-and-scan information for aspect conversions, but the industry has not implemented any form of this capability into encoders and decoders. "There is no standard," says Beavers. "Simply providing discrete pan-and-scan information doesn't guarantee interoperability between systems, or backward and forward compatibility."

He points out that while MPEG-4 goes a long way toward addressing many issues surrounding converging technologies, MPEG-2 is deeply entrenched and is likely to remain that way, at least for a while.

"TVa technology can significantly enhance MPEG-2-based products now, and become an integral component of MPEG-4 systems as they gain market share. TVa will eliminate the undesirable effects of prior systems and is positioned for rapid development within MPEG environments."

Beavers is also hopeful that future TVa deployments may include "fuzzy filters" that will provide relative pan values for nonencoded programs to a TVa decoder core.


Beavers lists several potential values behind TVa. Citing the fact that content providers often distribute programming on DVDs in both wide and narrow formats, Beavers believes that benefits could come in the form of savings in both bandwidth and bankfold.

"It's more efficient for a studio to distribute a program in a single format encoded with a TVa elementary stream," he says. "It could save them money in terms of bandwidth and archival storage requirements, and generate revenue for application vendors who can provide higher-quality products."

Beavers allows that successful deployment will ultimately rely on support from the martketplace. TechKinetic is seeking developer relationships with several electronics manufacturers.