Category: The Front Line

True Facts

A recent post on LinkedIn led me to write this, and it has to do with 4K video and imaging. Or, at least how marketing types have redefined it as “True” 4K or “Faux” 4K.

The post in question had to do with a projector manufacturer’s 4K offerings and how other manufacturers may not be offering a “true” 4K product by comparison, calling those other products “faux” 4K (or “faux” K to be clever). That prompted more than a few comments about what “true” 4K is in the first place.

One comment pointed out that the projector brand behind the original post doesn’t even have a “true” 4K imaging device in its projector, as it uses Texas Instruments’ .66” DMD with 2716×1528 micromirrors and requires image shifting to create an image with full 4K resolution. (Some irony in that?)

Now, I know more than a few marketing folks in the AV industry, and they work very hard and diligently to promote their company’s products. However, sometimes they step out of bounds and create more confusion, particularly with new technologies. Which, by the way, was one reason I started teaching technology classes at InfoComm and other trade shows two decades ago – as a way to counter marketing hype with facts.

What, exactly, is “true” 4K? If you use spatial resolution as your benchmark, then your imager must have at least 4000 horizontal or vertical pixels. The fact is, very few displays today have that much resolution, save for a limited number of digital cinema projectors, a handful of home theater projectors, and a small selection of reference and color grading monitors. All of which will set you back quite a few $$$.

Most displays that are lumped into the 4K category are really Ultra HD displays, having a fixed resolution of 3840 horizontal and 2160 vertical pixels. This would include every so-called 4K consumer TV, many digital signage displays, and production monitors. Are they “true” 4K? Going by spatial resolution, no.

What make things even more confusing is projection specsmanship. Sony’s original SXRD projectors had Ultra HD resolution. Although Epson has shown a prototype HTPS LCD chip with UHD resolution, they’ve never brought it to market. And the only DMD that Texas Instruments makes with 4K resolution is the 1.38” dark chip they sell into the digital cinema marketplace.

What projector manufacturers do instead to get to 4K is to use lower-resolution chips and shift the image with very fast refresh rates to effectively create 4K images. I’ve seen demos of the .66” DMD creating 4K images vs. a native UHD imager and you can see the difference between native and shifted images, particularly with fine text and detail. But it represents a low-cost way to get something approaching UHD resolution.

Panasonic also did this with their PT-RQ32U 4K DLP projector, using devices with 2550×1536 resolution and mapping quadrants to get to 5120×3200 total pixels. Presumably, they’ve retained this trick on their newer 4K models shown at InfoComm 2019.

Is that “true 4K?” Not when it comes to spatial resolution. But what if you base your claims on each finished frame of video, after all sub-fields are created? In that case, you could have an argument that your device is actually creating 4K video. Since our eyes can’t keep up with refresh rates much past 60 Hz, we’re not likely to see any flicker from this technique (also known as “wobbulation” and used by such luminaries as JVC and Hewlett-Packard on their display products in the past).

In fact, Digital Projections’ Insight Laser 8K projector employs three 1.38” dark chip DMDs and some clever image shifting to get from native 4096 x 2160 resolution to get to 8K (presumably 8192 x 4320 pixels in the finished images). Native 8K DMDs don’t exist, and like 8K camera sensors, wouldn’t come cheap if they did. Scaling down, it would make no sense financially to try and ship single-chip 4K DLP projectors with the 1.38” 4K DMD, not to mention the optical engine would have to be a lot larger, resulting in a bigger and heavier projector.

At this point, we should stop using the nomenclature “4K” altogether and switch to the more accurate CTA designation for Ultra HD (3840 x 2160) when we talk about the next generation of displays past Full HD (1920 x 1080) and 2K (2048 x 1080). Also, SMPTE designates two sets of resolutions that go beyond Full HD – UHD-1, or anything up to and including 3840 x 2160, and UHD-2, anything beyond UHD-1 up to 8K (7680 x 4320) and beyond.

From my perspective; if your imaging device can show me a complete frame of video with at least 3840 x 2160 pixels, refreshed at 60 Hz, then I’m okay with calling it UHD (NOT 4K). But there’s a catch: High frame rate video is going to be a big thing with UHD-1 and UHD-2 and will require refresh rates of 90, 100, 120 Hz, and even 240 Hz. Can your current projector show me a complete video frame with at least 3840 x 2160 pixels of spatial resolution when refreshed at 240 Hz? 120 Hz?

Boy, I can hardly wait for 8K projector marketing campaigns to start…

(This article originally appeared on 9/19/2019 in Display Daily.)

Selling Wisconsin By the Panel

Politics and technology intersect more often than you might think. And no better example of this can be found than the announcement in 2018 by Terry Gou of Hon Hai Precision (Foxconn) that he would build a state-of-the-art, Gen 10 LCD panel manufacturing plant in Wisconsin, presumably to crank out large panels for Ultra HDTVs. The plant would be the center of a $10 billion high-tech campus near Racine and would have earned over $3 billion in state and local tax credits, and Gou predicted over 13,000 jobs would have been created.

To say that many of us analysts and members of the press greeted this announcement with surprise is an understatement. There’s a pretty good reason why no one manufactures LCD panels in the United States, and that would be the high finished cost of each panel. Given the collapsing retail price trends for 4K televisions, no one could ever hope to recover such an investment. Was Gou trying to pass off a bill of goods? What would those 13,000 people actually do at this facility?

Now, it appears he’s wavering in his commitment. After the village of Mount Pleasant issued bonds and borrowed $300 million dollars to acquire and clear land for the project, Gou was quoted in January of this year as saying that the television factory was a dead idea. “In terms of TV, we have no place in the U.S.,” he told Reuters. “We can’t compete.”

Apparently that comment set off a series of phone calls between Foxconn and the White House, because two days later, Gou reversed course. “After productive discussions between the White House and the company…Foxconn is moving forward with our planned construction of a Gen 6 fab facility,” the company said in a statement. “This campus will serve both as an advanced manufacturing facility, as well as a hub of high technology innovation for the region.”

This Gen 6 fab (motherglass of 1850mm x 1500mm) will take up over a million square feet and is scheduled to hire 1,500 positions. That seems a bit high to me for a facility that is largely automated – less than 100 people run one entire LCD line for LG Display in Paju, Korea. But Gou’s original deal with the state also promised that 5,000 people would be employed at the Gen 6 fab.

From an NBC News report this past Sunday, it appears quite a bit of land has been cleared for the project, some roads have been laid down, and some buildings have been put up. But little if any activity is taking place at the site, and residents are getting a bit irritated about the whole thing. Several sold their homes as part of eminent domain acquisitions by the town and now feel they were misled. A former gubernatorial candidate, Matt Flynn, exclaimed “This state got snookered. We really got taken to the cleaners on this!” in an interview with WMAQ-TV in Chicago.

The Gen 6 facility was described as one that would make smaller glass cuts or things like smartphones, tablets, computer screens, and small TVs. But the smartphone market is shifting to OLEDs, tablet sales are in steady decline, and small televisions are an extremely low-profit business. So why make that kind of investment?

Gou may also be changing his manufacturing plans again, according to a story in the Nikkei Asian Review from June 11. According to that story, “The products to be manufactured in the Wisconsin factory would be extended beyond the planned display panels to servers, networking products and automotive central controls. The factory is expected to be operational from the end of next year…while Foxconn had originally intended the plant to manufacture new generation liquid crystal displays, the company has scaled back its ambitions several times, citing the overcrowded display market.”

Gou could also be playing a long game, figuring any investments in Wisconsin will spare him the tariff fights currently going on between Washington and Beijing. Perhaps his thinking is that bringing any jobs at all to this part of Wisconsin will certainly curry favor. Indeed, Tim Sheehy, president of the Metropolitan Milwaukee Association of Commerce, was quoted in the NBC story as saying, “Even if the project stopped where it is now, 1,500 jobs are nothing to sneeze at, and no one is out any money. This will still be the largest foreign investment that we’ve ever had in this region.”

In any case, the current governor of Wisconsin, Tony Evers (D), is all over Gou to get this project completed. What it will actually look like when completed is anyone’s guess, but from my perspective, it appears to be a classic “play for play” deal. And regardless of your political leanings, it still doesn’t make any sense economically to build an LCD panel fab in the United States, unless you’re happy to pay an arm and a leg for a new TV…

(This story originally appeared on Display Daily.)

All Good Things Must Come To An End…

It’s been a great 16-year ride, but I’m shutting down HDTVexpert.com at the end of this year to pursue other interests.

I’ll still be contributing to Sound & Communications magazine (https://www.soundandcommunications.com/) and you can find my monthly AVent Horizon columns there.

In addition, look for articles here and there on the Display Daily Web site (https://displaydaily.com/.

I hope you’ve enjoyed perusing these pages over the years. When I started the site, HDTV was just coming into its own. Now, we’re accelerating past 4K / UHDTV and 8K appears to be the next big thing. And who knows what lies beyond that? 16K? 32K? (Probably a new form of display technology that hasn’t been invented yet…)

Always ask better questions!

(In remembrance of Chris Campbell, 1959 – 2009)

AV-over-IT, Unplanned Obsolescence, and Unintended Consequences

The audiovisual industry, like many others, is in the midst of a big paradigm shift. That term is often beaten to death, but in this case, it’s apropos as we are redefining the process of switching and distributing audio and video signals (with control and metadata thrown in). And we’re doing it by migrating to a packet structure, using standard TCP/IP headers and streaming protocols.

And we’re not alone. Broadcast, cable, and satellite video services are adopting IT-centric models for signal management, a trend that began year ago with the first streaming video services. We laughed at those early attempts, as a combination of slow networks and inefficient video compression imposed limits on image resolution and refresh rates. Remember the old jokes about “dancing postage stamps” heard at the NAB show in 1999, in the Streaming Media pavilion?

Well, no one’s laughing these days. We can send 4K video with multichannel surround over fast broadband networks to multiple homes, reliably and repeatedly. Streaming media has become so popular that it is putting a noticeable dent in subscriptions to traditional multichannel video services (like cable and satellite TV). Some smaller cable MSOs have even encouraged viewers to drop pay TV packages and sign up for streaming services – those systems are actually losing money on pay TV subscriptions!

Streaming companies like Netflix and Amazon pose a real threat to established content producers. Indeed, the industry consolidation we’re seeing in Hollywood is largely a response to this trend, and media conglomerates now have plays in movie theaters, broadcast television, and streaming services. Think Disney, with its recent acquisition of Fox and its ownership of Marvel Studios, Lucasfilm, and Pixar, plus the ABC network and ESPN.

But that’s all old news. And while people endlessly debate the future of traditional linear video services, video-on-demand, and pay-per-view, there is another debate that isn’t happening. And that’s the future of skilled technical staff, particularly engineers.

As the move to IT structures for signal management and program distribution picks up speed, so too is a move to shed engineering staff. Yes, there are redundancies any time one company buys another, which is why so many staff positions were eliminated when Disney completed its acquisition of 21st century Fox studios. But Disney’s been cutting back for some time on the broadcast side, both at the O&O station level and at the ABC network.

A similar shift is taking place at CBS, possibly to prepare for an acquisition by Viacom. Many engineering positions are being eliminated as a result. And this downsizing isn’t limited to broadcast networks – it’s happened recently and will happen again at cable MSOs, not to mention satellite MVPDs that are seeing their days coming to an end in favor of lower-cost streaming.

The question is, why? And the answer appears to be along the lines of “since everything is moving to an IT network, and so much of the signal management process is automated, a few IT specialists can handle the job.” This thinking tends to oversimplify the process, but there’s no question that engineers don’t carry nearly as much weight at the TV station and network level as they used to, even a decade ago.

There is a second undercurrent, and that’s the commoditization of hardware. To save even more money, the focus has shifted from capital expenses (CapEx) to operating expenses (OpEx) at a myriad of facilities from colleges and universities to TV stations, corporate offices, and post-production facilities. Here, the thinking is that hardware has become so powerful and so inexpensive that it’s foolish to spend lots of money on products that might be obsolete in a year.

And with all signal distribution moving to an IT structure, it’s simply easier to buy two or more of a given product so that spares are on hand when the original purchase fails. This is particularly true of video cameras – even UHD models have gone cheap – but we’re also hearing about similar purchases of things like routing switchers, monitors (even TVs are used), and audio gear. If it lasts a year and blows/burns up, simply toss it and pull another one out of the box.

There’s ample evidence of the first trend taking root across a wide variety of institutions, whereby dedicated AV departments are being replaced by IT specialists or having their job functions moved over to in-house IT administrators. There may be real bottom-line savings as a result, but how do you replace all of that technical expertise? Is an in-depth technical education in all aspects of video, audio, lighting, and control going to be worth as much down the road to employers and customers?

Similarly, are we moving to a model of planned obsolescence and commoditized hardware with short-term ROI goals? Do we get rid of old hardware like we throw out used paper plates and cups? Would it make sense for AV integrators to shift their business models away from conventional hardware installation and move to a leasing model, i.e. build out a complete room (or rooms) with all the requisite gear and lease everything to the client for a year or two, then cart it all away for recycling or disposal?

Carrying this model to the extreme, we now see the rise of “rent by the hour/day/week” flex office space in major cities. Who needs to own a building and deal with the substantial CapEx when a company can exist on paper with no real permanent office, except when you need it to impress/meet/inform a client? It’s not much of a stretch to argue in a similar manner against retaining full-time technical and engineering staff in a given facility where most of the hardware and software in use is automated.

Food for thought…

NAB 2019: Where Does It Go From Here?

This year’s NAB Show marked a milestone for me as it was my 25th consecutive April visit to the halls of the Las Vegas Convention Center. Back in 1995, my first impression was of enormous booths full of expensive hardware (lots of cameras with five-figure price tags), tape-based recording and editing systems, huge audio consoles, and the costly first iterations of non-linear video editing systems.

The World Wide Web was just becoming a thing, and NAB had set up an area in the then-Hilton ballroom so that a bunch of small companies (none of whose names I remember and none of who are around any more) could amaze us with stories of video streaming (basically tiny stamp-sized clips of low-resolution talking heads) and how someday, “all of this will be available to anyone with an Internet connection.” Of course, it would have to be a lot faster than the dial-up speeds available then.

I just remember shaking my head, thinking none of this would ever fly, and moving onto the big, expensive hardware. SDI, VGA, BNC connectors – stuff I could understand. “Reference” video monitors were bulky, heavy chassis with cathode-ray tubes and sold for tens of thousands of dollars.

Video itself was largely standard definition back then – HDTV was still in its infancy, and the so-called “Grand Alliance” of companies like Zenith/LG, AT&T, General Instrument, and others were pushing for an all-digital broadcast television system to replace NTSC. (Grand Alliance banners were prominently featured on the outside facades of the North Hall and Central Hall.)

Sharp’s 8C-30A is the first 8K DSLR to market. It only shoots 30p, but that’s a bus speed and interfacing limitation.

 

NHK’s been broadcasting 8K video content via satellite since last December and will cover the 2020 Olympics extensively in 8K.

And the “broadcast” part of NAB was BIG. Lots of companies showing exciters, transmitters, antennas, monitoring equipment, tubes and solid-state rigs, and hardline and waveguide dominated the North Hall and part of the Central Hall. Industry giants like Panasonic, JVC, Hitachi, Sony, Ikegami, Canon, and Toshiba constructed booths larger than the average house. Life was good, sales were brisk, and there was plenty of profit for everyone.

At the time, I was writing columns and feature articles for Video Systems, and our annual NAB issue was so stuffed with ads that it ran well over 300 pages. Other publications were jockeying for ad sales and editorial coverage, filling NAB press conferences to the rafters. You may remember some of them – Videography, Post, Millimeter, AV Video, Broadcast Engineering, TV Technology, and Television Broadcast, to name a few.

We all know what happened in the intervening years. The “broadcast” part of NAB is a tiny portion of the exhibit floor now. Those streaming video guys with their advanced codecs now rule the roost. High-quality video cameras capable of shooting 4K video sell for one-tenth of those 1995 models and do it all to solid-state memory cards. Non-linear editing is so ubiquitous that you can buy the software for about $100 and run it on your everyday laptop. And trade publications are largely dinosaurs.

We made the transition to digital TV broadcasting in 2009 and are about to move again to a newer, more IP-centric system, ATSC 3.0. CRTs are a distant memory, replaced by large high-resolution flat screen LCD and OLED displays. High dynamic range with its associated wider color gamut is muscling its way into our homes and theaters. High frame rate video, once considered a major obstacle, is becoming reality, even at Ultra HD resolution.

Chris Chinnock of the 8K Association ran a full day of 8K seminars on Wednesday.

 

Socionext is the only chip manufacturer able to deliver HDMI 2.1 TX and RX chipsets at present.

And those once-enormous booths are steadily shrinking each year as profit continues to evaporate from hardware sales. Indeed; the focus at this year’s NAB Show seemed to be shifting (as one company VP put it) away from capital expenditures (CapEx) to operating expenditures (OpEx). Hardware is increasingly becoming generic – the ATSC 3.0 single-frequency network broadcasting demos in the North Hall ran off a pair of compact Dell servers – and software is the new king of the hill. (How much booth space do you really need to demo software?)

The shift away from hardware to software may also be impacting attendance. The official head count for 2017 was 103,443. Two years later, attendance was pegged at 91,460; representing a decline of 11.6%, or 12,000 visitors. That’s quite a dip – not as pronounced as the drop from 2008 to 2009 during the Great Recession, but still enough to give show organizers some reason for concern.

The general feeling among many of my colleagues and friends was that the show was trying to find an identify. The theme of the show was “Every Story Starts Here,” which is about as generic and vague as you can get. Was that a reference to the fact that that high-quality tools to capture, edit, and produce compelling video are a relative bargain these days, and that we should concentrate more on how we use them and less on how they work?

Astro exhibited many things 8K, including this compact 8K/120 camera with built-in CCU.

 

Leyard Planar brought back their 8K LED fine pitch videowall from a year ago.

I don’t have the answers. What I can tell you is that there were still some interesting products to be found in Las Vegas, and from my perspective, a great deal of them pertained to UHD video capture, editing, and distribution. My focus has always been forward-looking (aside from the nostalgia piece I opened this feature with), so the transition of HDTV to Ultra HDTV and the attendant transitions to a new digital TV broadcast system and advanced video codecs held the greatest interest for me.

Yes, 8K has arrived, Yes, it’s not immediately obvious why we need it. But like it or not, the cameras are here, the displays are coming to store shelves, and at least one broadcaster (NHK) is operating a 24/7 8K video channel. Companies that showed 8K cameras at NAB included Sony, Ikegami, RED, Hitachi, and Sharp. The latter brand made 8K the focus of their entire booth, a bold and impressive statement considering Sharp was nearing bankruptcy not that long ago.

Panasonic’s take on 8K was something called an “area of interest” camera, allowing anyone to dynamically select and switch between any of four 2K video slices of the overall image. This technique was described by NHK a few years back at the annual SMPTE Conference as a way to achieve a multi-camera shoot with just one or two cameras. Astro (who has commercialized many of NHK’s 8K innovations) had several flavors of 8K on display including a stereo VR camera, a fisheye lens fitted to an 8K camera, and a compact 8K/120p camera with CCU built-in that weighed less than 10 pounds.

Panasonic’s 8K Area Of Interest (AOI) camera lets you create a virtual four-camera 2K shoot and switcher – all from one static view.

 

Sony’s UHC-8300 8K portable camcorder will get a workout at the 2020 Olympics.

Astro also addressed an on-going challenge for adoption of 8K video and display, and that was upconverting 1080p/2K content in an acceptable manner. Their demo of an AI-based up-scaling system was quite impressive, particularly given the challenging test patterns and fine text used in the demo. Sharp’s big announcement was the availability of the world’s first DLSR with 8K native resolution. The 8C-30A can shoot 8K/30p video using a Micro Four Thirds sensor and a variety of lenses, but will be pricey at around $4,000.

Notable by their absence in this market: Canon. I was told by a company official that Canon does sell an 8K camera in Japan for broadcast and production, but has no plans to offer it stateside unless there is sufficient demand. Of course, Japan remains the focus for every player in 8K, what with the summer Olympics coming up in 2020 and brands like Hitachi and Sony jockeying to provide cameras and hardware to cover the event.

Given the enormous volume of data that 4K and 8K cameras generate, I was also on the lookout for advanced codecs. HEVC H.265 has been around for a few years and could be suitable for the job. The only question is latency, particularly for contribution. NTT seems to have a handle on the problem, as they showed a prototype encoder/decoder combination for 4K/120 video that has an end-to-end latency of just 37 milliseconds.

If you’re having trouble keeping track of all the players in UHD HDR, this chart may help. Or not.

 

Ikegami also has an 8K production camera ready for the 2020 Olympics.

Another approach is to improve codec efficiency. At the Wednesday 8K Association seminars, a more detailed explanation of the new Versatile Video Codec (VVC, perhaps to become H.266?) was offered. The key to improving efficiency is increasing the maximum encoding block size from 64×64 used in H.265 to 128×128, with a targeted reduction of 50% in bit rate. Presently, the actual improvement is about 34% as software evolves.

Astro and Sharp also demonstrated a practical 8K/60p non-linear editing and color correction platform known as Tamazone. It imports 8K 4:2:2 10-bit YUV from Sharp’s 8C-60A camcorder using four 12G SDI connections through BlackMagic Design’s DeckLink 8K Pro interface, with DaVinci Resolve 15 Studio and Resolve Live software. To view what you’re working on, an nVIDIA Quadro Pro 4000 graphics card drives a pair of DisplayPort 1.2 interfaces on an associated 8K monitor, stitching together two 4K images.

Yes, it’s a bit of a “Scotch Tape and paper clips” solution at the display end…but then, so was 4K when it first got off the ground more than a decade ago. The real challenge now is at the monitor – there really aren’t any true 8K reference monitors out there. Sharp showed a 32-inch prototype using their IGZO TFT backplane technology, but for now, the few 8K monitor products being talked about are largely based on consumer television designs. (LG Display has also shown an 8K 31.5” monitor for several years now.)

Advantech was streaming 8K/60 video across their booth over a fast IP network at 200 Mb/s using HEVC H.265.

 

NTT claims they can encode and decode 4K / UHD video with a 120 Hz frame rate using H.265, but with just 37 milliseconds of latency end-to-end.

It’s possible that the new 31.5” 4K LCD monitors shown by Panasonic, Sony, and others might fit the bill eventually. This HDR monitor uses two panels. The first is a 4096×2160 IPS-Alpha LCD with full array backlight, while the second panel is identical in construction but free of color filters. It is precisely aligned with the first panel and works as a monochromatic light shutter to provide really deep black levels. Of course, it requires a lot more horsepower in the backlight as a consequence.

Another popular discussion on the show floor was the concept of shooting, editing, and archiving at higher resolution and using lower-resolution (4K, Full HD) for distribution. In this way, the goals of improving current HD and 4K video quality can be improved significantly, but the finished product is a bit more friendly to bandwidth-constricted distribution systems like broadcast and streaming. Hitachi has argued for Full HD with HDR as a practical broadcast format (it is) and a 4K program derived from an 8K master would look pretty darn good on Netflix and Amazon Prime.

So many questions and not a lot of answers, just possible solutions. I expect NAB to get smaller over time as the emphasis shifts from hardware to software (ironic, given the LVCC is in the middle of another expansion to almost 4 million square feet). No doubt 8K will be a part of it, as will IP-based distribution of media. No wonder there was an air of “where do we go from here?” during the show…