It took only a few years for high-definition televisions to make the transition from high-priced novelty to ubiquitous commodity — and they now seem to be heading for obsolescence just as quickly.
At the Consumer Electronics Show (CES) in January, several manufacturers debuted new ultrahigh-definition, or UHD, models (also known as 4K or Quad HD) with four times the resolution of today’s HD TVs.
In addition to screens with four times the pixels, however, UHD also requires a new video-coding standard, known as high-efficiency video coding, or HEVC. Also at CES, Broadcom announced the first commercial HEVC chip, which it said will go into volume production in mid-2014.
At the International Solid-State Circuits Conference this week, MIT researchers unveiled their own HEVC chip. The researchers’ design was executed by the Taiwan Semiconductor Manufacturing Company, through its University Shuttle Program, and Texas Instruments (TI) funded the chip's development.
Although the MIT chip isn’t intended for commercial release, its developers believe that the challenge of implementing HEVC algorithms in silicon helps illustrate design principles that could be broadly useful. Moreover, “because now we have the chip with us, it is now possible for us to figure out ways in which different types of video data actually interact with hardware,” says Mehul Tikekar, an MIT graduate student in electrical engineering and computer science and one of the paper's co-authors. “People don’t really know, ‘What is the hardware complexity of doing, say, different types of video streams?’”
In the pipeline
Like older coding standards, the HEVC standard exploits the fact that in successive frames of video, most of the pixels stay the same. Rather than transmitting entire frames, it’s usually enough for broadcasters to transmit just the moving pixels, saving a great deal of bandwidth. The first step in the encoding process is thus to calculate “motion vectors” — mathematical descriptions of the motion of objects in the frame.
On the receiving, end, however, that description will not yield a perfectly faithful image, as the orientation of a moving object and the way it’s illuminated can change as it moves. So the next step is to add a little extra information to correct motion estimates that are based solely on the vectors. Finally, to save even more bandwidth, the motion vectors and the corrective information are run through a standard data-compression algorithm, and the results are sent to the receiver.
The new chip performs this process in reverse. It was designed by researchers in the lab of Anantha Chandrakasan, the Joseph F. and Nancy P. Keithley Professor of Electrical Engineering and head of the MIT Department of Electrical Engineering and Computer Science. In addition to Chandrakasan and Tikekar, these include Chiraag Juvekar, another graduate student in Chandrakasan’s group; former postdoc Chao-Tsung Huang; and former graduate student Vivienne Sze, now at TI.
The chip’s first trick for increasing efficiency is to “pipeline” the decoding process: A chunk of data is decompressed and passed to a motion-compensation circuit, but as soon as the motion compensation begins, the decompression circuit takes in the next chunk of data. After motion compensation is complete, the data passes to a circuit that applies the corrective data and, finally, to a filtering circuit that smooths out whatever rough edges remain.
Pipelining is fairly standard in most video chips, but the MIT researchers developed a couple of other tricks to further improve efficiency. The application of the corrective data, for instance, is a single calculation known as matrix multiplication. A matrix is just a big grid of numbers; in matrix multiplication, numbers in the rows of one matrix are multiplied by numbers in the columns of another, and the results are added together to produce entries in a new matrix.
“We observed that the matrix has some patterns in it,” Tikekar explains. In the new standard, a 32-by-32 matrix, representing a 32-by-32 block of pixels, is multiplied by another 32-by-32 matrix, containing corrective information. In principle, the corrective matrix could contain 1,024 different values. But the MIT researchers observed that, in practice, “there are only 32 unique numbers,” Tikekar says. “So we can efficiently implement one of these [multiplications] and then use the same hardware to do the rest.”
Similarly, Juvekar developed a more efficient way to store video data in memory. The “naive way,” he explains, would be to store the values of each row of pixels at successive memory addresses. In that scheme, the values of pixels that are next to each other in a row would also be adjacent in memory, but the value of the pixels below them would be far away.
In video decoding, however, “it is highly likely that if you need the pixel on top, you also need the pixel right below it,” Juvekar says. “So we optimize the data into small square blocks that are stored together. When you access something from memory, you not only get the pixels on the right and left, but you also get the pixels on the top and bottom in the same request.”
Chandrakasan’s group specializes in low-power devices, and in ongoing work, the researchers are trying to reduce the power consumption of the chip even further, to prolong the battery life of quad-HD cell phones or tablet computers. One design modification they plan to investigate, Tikekar says, is the use of several smaller decoding pipelines that work in parallel. Reducing the computational demands on each group of circuits would also reduce the chip’s operating voltage.
Sarah McDonnell | EurekAlert!
Producing electricity during flight
20.09.2017 | Albert-Ludwigs-Universität Freiburg im Breisgau
Solar-to-fuel system recycles CO2 to make ethanol and ethylene
19.09.2017 | DOE/Lawrence Berkeley National Laboratory
Whispering gallery mode (WGM) resonators are used to make tiny micro-lasers, sensors, switches, routers and other devices. These tiny structures rely on a...
Using ultrafast flashes of laser and x-ray radiation, scientists at the Max Planck Institute of Quantum Optics (Garching, Germany) took snapshots of the briefest electron motion inside a solid material to date. The electron motion lasted only 750 billionths of the billionth of a second before it fainted, setting a new record of human capability to capture ultrafast processes inside solids!
When x-rays shine onto solid materials or large molecules, an electron is pushed away from its original place near the nucleus of the atom, leaving a hole...
For the first time, physicists have successfully imaged spiral magnetic ordering in a multiferroic material. These materials are considered highly promising candidates for future data storage media. The researchers were able to prove their findings using unique quantum sensors that were developed at Basel University and that can analyze electromagnetic fields on the nanometer scale. The results – obtained by scientists from the University of Basel’s Department of Physics, the Swiss Nanoscience Institute, the University of Montpellier and several laboratories from University Paris-Saclay – were recently published in the journal Nature.
Multiferroics are materials that simultaneously react to electric and magnetic fields. These two properties are rarely found together, and their combined...
MBM ScienceBridge GmbH successfully negotiated a license agreement between University Medical Center Göttingen (UMG) and the biotech company Tissue Systems Holding GmbH about commercial use of a multi-well tissue plate for automated and reliable tissue engineering & drug testing.
MBM ScienceBridge GmbH successfully negotiated a license agreement between University Medical Center Göttingen (UMG) and the biotech company Tissue Systems...
Pathogenic bacteria are becoming resistant to common antibiotics to an ever increasing degree. One of the most difficult germs is Pseudomonas aeruginosa, a...
19.09.2017 | Event News
12.09.2017 | Event News
06.09.2017 | Event News
20.09.2017 | Life Sciences
20.09.2017 | Power and Electrical Engineering
20.09.2017 | Physics and Astronomy