The Case Against Sonic Foundry’s Mediasite Lecture Capture System

By Lynn Kenneth Packer January, 2011

It was a chilly early November morning, really early—about 1 am—in 1968, when a bus full of us

draftees and enlistees pulled into the Ford Ord entrance parking lot. A drill sergeant/reception committee-of-one barked out our first orders: “Line up, in a single row, facing me, at attention, next to the bus!”

“Now bend down and grab your ankles!” he yelled. As I reached for my ankles, I wondered if PT—physical training—began right out of the gate. “Grip those ankles tight and pull as hard as you can!” About thirty of us recruits, not yet even in uniform, gritted our teeth and tugged away trying

desperately to yank our feet off the ground. “Keep pulling!” the drill sergeant screamed, “and just maybe you’ll pull your heads out of your

asses!” Point well taken. There was a war on. Most of us would end up in ‘Nam within a year, and just maybe we’d learn

enough in the next few weeks to keep from getting ourselves or our buddies killed. Traditional American “residential” universities are now at war with enemies as formidable as

were the NVA and Viet Cong. A big foe of hallowed halls of learning: For-profit online schools. The battle for students has forced bricks and mortar institutions to expand their own online offerings and counter with lower prices that are often subsidized with tax dollars.

“… demand for online courses might soon jump even more as expanding ranks of traditional ivy-covered universities and Internet entrepreneurs introduce online programs that are just a few hundred dollars per course. (Or, if you don't care about getting credit, they're absolutely free.) Meanwhile, technological improvements, such as easier-to-use video cameras and software, are helping online schools make their courses more rigorous and more engaging.” (U.S. News and World Report, April 2, 2009) But their biggest enemy? Universities’ own change-resistant

bureaucracies. As Pogo said, “We have met the enemy and he is us.” Talk about the need to pull heads out of asses. Traditional universities

may be driven to extinction by the internet because, dinosaur-like, they simply cannot adapt well to the digital age. Just recently Microsoft founder Bill Gates predicted that “five years from now, on the Web for free, you’ll be able to find the best lectures in the world. It will be better than any single university.” He believes the $50,000 a year university education could be obtained via the web for as little as $2,000.

Take lecture capture. By now every key lecture hall on every campus should have had it. Every student should have been able to go online and review every important lecture. But universities, with heads stuck well up their butts, don’t even come close.

2

That November, 1968 bus ride from San Francisco International Airport to Fort Ord took us smack dab through Santa Clara Valley, soon to be dubbed Silicon Valley. Although Bill Gates would settle his company near Seattle, the key concepts behind his Windows operating systems were conceived in Silicon Valley.

But that was later. In 1968 several Santa Clara area manufacturers were busy printing semiconductors on silicon wafers creating integrated circuits (ICs, aka microchips), tiny devices that were making vacuum tubes obsolete.

Just four months earlier a new Palo Alto company called Intel (INTegratedELectronics) had split off from Fairchild Semiconductor and set about turning ICs into microprocessors, devices that could replace vacuum tubes in mainframe computers, eventually making small personal computers possible.

Having just been drafted away from a television news reporting job in Salt Lake City, having left behind a pregnant wife, and with the turmoil of Vietnam swirling about, I was rather oblivious to the high tech tremor rumbling through Santa Clara Valley. Besides, the television news industry was still decidedly low tech, still using film, not yet even videotape. Solid-state minicams that could capture light on microchips and record onto magnetic tape were still another decade away.

The first computer I had actually seen was a few years earlier when a neighbor gave me a tour of Thiokol Chemical Corporation’s main office at the remote Utah site where it static-tested Minuteman rockets. The mainframe computer filled a large, air conditioned room. Its hundreds of vacuum tubes were much larger than the ones in car radios I would test and replace while working at my father’s car dealership. (Back then motorists would have to wait for their radios to warm up while enduring the hum of vibrators that stepped up battery voltage.)

In 1968 I had already heard of HP—Hewlett Packard. At that time HP made various types of electronic equipment but would become best known three or four years later for its pocket calculator replacement for slide rules, like the one I used in my engineering classes in college. But I didn’t know the Palo Alto garage Stanford grads Bill Hewlett and David Packard used as a shop was the birthplace of Silicon Valley. Back then it was just another destined-to-be historic site just off the bus route to Fort Ord.

The holiest of sites along that route to the military base was the building at 3180 Porter Drive in a research park just south of Stanford. In a little over a year—in 1970—it would be the first location for Xerox’s Palo Alto Research Center (PARC). It was there key strands of lecture capture DNA originated: The first practical

PC named the Alto (lecture capture appliances are merely PCs of one sort or another with encoding software); the graphical user interface (GUI; used by many lecture capture products to display content); multiple, on-screen windows (most commonly used inside a viewer interface to show instructors in one window and rich media/visual aids in the other); and the local area network (often used to manage capture devices and the video they generate). The idea that a computer should not only just crunch numbers but also be used to teach, even visually, propelled a lot of PARC’s research.

PARC opened while I was in Vietnam, a war that had everything yet nothing to do with PARC. Nothing in the sense that relatively little Pentagon

money funded PARC. It mostly came from Xerox and its wildly profitable copy machines. But everything in the sense that the spiraling cost of the war was forcing the military to cut back on funding computer-related research projects, including the Advanced Research Projects Agency Network (ARPANET), the precursor to the internet. That made dozens of immensely talented researchers available to work at PARC.

3

Robert Taylor, who Xerox brought in to staff PARC, had been the driving force behind ARPANET when he directed ARPA, an arm of the Department of Defense at the Pentagon. Among reasons Taylor wanted to leave ARPA was the drying up of funds and his disillusionment with the war. Indeed, just prior to going PARC he had already left the Pentagon for the University of Utah where David Evans and later Ivan Sutherland, one of his former ARPA cohorts, ran an ARPA-funded computer science department that specialized in cutting edge graphics and display technology. (In 1968 the University of Utah created the most influential computer graphics programs in the world. Individuals involved in the Utah program went on to establish leading graphics companies like E&S, Silicon Graphics, Adobe, Netscape and Pixar.)

Berkley, MIT, Carnegie Mellon and Stanford were also ARPA darlings, receiving millions in government research grants. (No surprise that the lecture capture products MediaSite and Panopto both have roots at Carnegie-Mellon.) Even then, in 1968, Stanford was helping design ARPANET which, within the year, would be up and running with connections at Stanford, UCLA, UC Santa Barbara and the University of Utah.

Almost 30 years later Stanford became the first to successfully stream live lecture video over the internet. But was Stanford helping sow the seeds of its and other universities’ demise? Can online learning dramatically reduce or even eliminate face-to-face university classrooms? Or will lecture capture, at the very least, eventually show up in every major classroom?

Despite being a lecture capture pioneer Stanford never rolled out the technology campus-wide. Despite creating Standoff Online years ahead of its time and today webcasting lectures on iTunes U and YouTube University, Stanford still makes only a fraction of its lectures available on the internet.

What gives? In a 1997 Forbes Magazine article college professor and business guru Peter Drucker predicted

universities won’t survive. “Thirty years from now the big university campuses will be relics,” he said. “Do you realize that the cost of higher education has risen as fast as the cost of health care?” Drucker asked. “Without any visible improvement in either the content or the quality of education, means that the system is rapidly becoming untenable. Higher education is in deep crisis.”

Okay, let’s do the math. Drucker says 30 years from 1997. That would be 2027. Gates says five years from 2010 the best lectures in the world will be on the web, for free. That’s 2015. I suspect the latter will help the former prediction come true, maybe faster than Drucker predicted.

Drucker died in 2005 but before that, in 2000, Forbes did a cover story headlined “Webucation is the Next Great Growth Opportunity.” Drucker predicted that the Internet would become the most important teaching tool of the 21st Century. “Already we are beginning to deliver more lectures and classes off campus via satellite or two-way video at a fraction of the cost,” Drucker said. “The college won't survive as a residential institution. Today's buildings are hopelessly unsuited and totally unneeded."

Time may be running out for traditional universities just as it has for traditional newspapers.

Low Tech U

By now most key lectures at most universities should be captured on video and available online.

While a few schools are moving toward campus-wide lecture capture, most are not. Not even Stanford. (Hewlett and Packard may be rolling over in their graves. HP paid big bucks for Stanford to supply training videos first via microwave and later the internet; money that jump-started Stanford’s lecture capture pioneering.)

4

A much better case for universal lecture capture needs to be made for what few universities might listen—most probably won’t. But also for state legislatures who fund at least the public institutions of higher learning.

Who is making the case? Lecture capture makers among others are trying: Sonic Foundry (Mediasite), Accordent ( Capture Station), Winnov ( C-Box), VBrick ( Rich Media Studio), nCast (Telepresenter), 323Link (EDUcast), Echo360 (capture appliance), Matterhorn ( open source make-it-yourself encoder), Qumu ( Capture Studio), software products like Panopto and Tegrity, and my own V-Chassis product. None of those marketing campaigns has met with overwhelming success.

By now even K-12 should have been widely adopting lecture capture. But it remains practically nonexistent below the college level.

The path to universal lecture capture in the United States is still strewn with too many mines that need to be cleared. The industry effort to democratize lecture capture is clearly off course. There needs to be a new attack plan. Some new products and ways to promote them. Some serious head-out-of-ass pulling.

For one there needs to be a stronger move toward cross-platform compatibility, proselytizing against the false religion of Proprietarianism and evangelizing for more open more sensible lecture capture bid specifications. There could be joint R&D, lobbying and marketing to the extent allowed by anticompetitive laws like those which prohibit collusion and bid rigging. And maybe even ganging up on self-declared market leader Mediasite whose million dollar marketing campaigns and anticompetitive tactics have muddled the lecture capture landscape from coast to coast.

Origins of Lecture Capture

Just months prior to getting basic at Fort Ord I graduated from Utah State University with a BA

in broadcast news. I was editor of the campus paper, was anti-war and would have dodged the draft had I been able to figure out how to do it short of moving to Canada. I rolled the dice, did not enlist for an extra year in order to stay out of the infantry, and hoped for a Military Occupation Specialty (MOS) in broadcasting. The odds were against it. Most draftees were assigned combat.

During basic at Ord I wrote two or three columns for my campus paper back in Logan. One column described how well the Army was actually able to get most recruits to extract heads from asses and learn quite a lot in eight weeks.

One of the methods they used was teaching by television. Wow! The US Army was experimenting with teaching techniques more advanced than any I had seen at USU. The way it worked was a classroom instructor would teach some stuff live then present another part using video. It was displayed on cathode ray tube (CRT) television sets probably no bigger than 36” diagonal. They didn’t have half-inch VHS videotape cassettes back then so they would have used fairly expensive, two-inch reel playback machines linked to classrooms via closed circuit. The picture was black and white, not color.

It was crude by today’s standards. But the Army was able to use its best instructors, edit the video to keep up the teaching pace and use close-ups to show detail that would have otherwise been even harder to see than the small displays.

Okay, I admit I dozed during some of them, but I was still impressed. The army continued to use some 16 mm film with projectors. Some of it was kinescope recordings of the videotaped lectures. (Near the time the military was developing its teaching via television the Ford Foundation funded the creation

Sand or butt: Little difference

5

of the Educational Television and Radio Center [ETRC] and by 1954 it was operating as a PBS precursor network that mailed kinescope film to affiliated stations. Nicknamed “University of the Air” most of its educational shows were for adults.)

Still it was a bit incongruous. How it is the US Army used more advanced training techniques than a university? Well, universities were dabbling in it. And many were expanding on the ETRC concept by deploying educational television channels for distance learning. But nothing like what the military was doing.

It turns out that well before 1968 the military had even been using live television instruction, where a really good instructor on one base would teach a live course to soldiers at bases in other states. The Army abandoned live training in the early sixties with the advent of videotape and replaced live with on-demand. Some universities also did live instruction. The Utah Education Network (UEN) still does.

Part of the explanation—why the Army was a high tech teaching leader— was right there in Silicon Valley. Most of the funding for the research that did take place on university campuses, research that led to the internet and personal computers came from the military. Apparently military brass figured high tech is important to keep an edge over their enemies whether on the battlefield or in the classroom.

Maybe the main reason the military uses teaching technology better than universities is that a general can issue a top-down order and command, “This is the way we’re going to do it,” and the enlisted and officer instructors promptly salute and execute the order. Top-down directives work in the military.

But not on campus. A university president can’t order diddly squat. Presidents would be lucky to avoid being flipped off if they tried to require across-the-board lecture capture. During the nine years I taught broadcast news at Brigham Young University the president of the school had next to nothing to do with how we taught in our department. Our student labs were computerized while I was there but it was mostly a departmental initiative. While I underwent several student evaluations I never had a department head sit in on a class or require the use of this or that teaching technology.

Meanwhile the American educational system falls further behind. A 2006 National Conference of State Legislatures (NCSL) Blue Ribbon Commission on Higher

Education concluded “The American system is no longer the best in the world." "Other countries are outperforming us. At the same time, tuition and fees are skyrocketing and financial aid and loan programs aren't keeping up. As a result, a postsecondary education is not accessible to many Americans.”

This all makes universal lecture capture a no-brainer. But brains don’t work very well when they’re half way up someone’s intestines.

Kevin Carey, policy director of Washington think tank Education Sector, says to survive and prosper, “universities need to integrate technology and teaching in a way that improves the learning experience while simultaneously passing the savings on to students in the form of lower prices.” “Some people will argue that the best traditional college courses are superior to any online offering, and they're often right. There is no substitute for a live teacher and student, meeting minds.”

“But remember” Carey said, “That’s far from the experience of the lower-division undergraduate sitting in the back row of a lecture hall. All she's getting is a live version of what iTunes University offers free, minus the ability to pause, rewind, and fast forward at a time and place of her choosing.”

Its clear university presidents are not well positioned to initiate—top down— the deployment of lecture capture technology. Or enforce its use once it’s adopted. Another reason time is running on traditional universities.

I’m getting off track. Back the to the origins of lecture capture and back to Silicon Valley where lecture capture was taking root.

6

I figure I was on the Fort Ord rifle range on the beach facing Monterey Bay the very day the world’s greatest PowerPoint presentation was being given just up the peninsula. Okay, so PowerPoint wasn’t invented yet. It was still an elaborate dog and pony show with a giant screen (22’X18’) just like Steve Jobs uses to make big Apple announcements.

On December 9, 1968, Douglas Engelbart and his Stanford Research Institute (SRI) research presented a 90-minute live public demonstration dubbed “The Mother of All Demos.” Engelbart himself was on a stage with the large screen in San Francisco. But down the peninsula, at Stanford, another camera was pointed to the computer that was connected to the hall using video links, telephone lines and maybe bubble gum.

Engelbart used a strange little box about the size of his fist, a wooden shell with two metal wheels, to control his demo. He called it a

mouse. The big screen often showed Engelbart on one side with the live demo on the other, a forerunner of Microsoft’s Windows and mother of the multi-frame viewer interfaces used by most lecture capture concepts. Engelbart’s “PowerPoint” was the world debut of the mouse, hypermedia, and on-screen video teleconferencing.

There were no LCD or DLP projectors back then. Engelbart’s people had borrowed an Eidaphor video projector from NASA, arranged by none other than Bob Taylor before he left ARPA. (Taylor had earlier directed government money to SRI when he was with NASA.) That contraption was housed in a six-foot-high cabinet and projected an 875-line image. It formed those images with an electron beam directed onto a sheet of oil that was repeatedly wiped away by a windshield wiper. At one point the amazing demo showed Engelbart moving a cursor on the screen using the little box with wheels and using a keyboard to create on-screen text. That was the beginning of word processing.

So in 1968 television technology was used with computer technology to create Engelbart’s mind-boggling production. A more complete merging of television and computer display technology would take another 40 years.

The San Andreas Fault runs right through Silicone Valley. Engelbart’s ground-shaking demo was just figurative, but his work at Stanford shook the computer world as if it had hit a 6 on the Richter scale. It set the stage for PARC on the other side of Stanford’s campus that would register a magnitude 8.

In June, 1970 I was clear across the country at Fort Gordon, Georgia. The bad news: I received orders for ‘Nam. The good news: I was assigned to the Armed Forces Vietnam Network (AFVN). Based on my commercial radio and television experience I dodged combat by getting a direct, Broadcast Specialist MOS.

I was assigned to AFVN’s Detachment Five in Quang Tri, near the demilitarized zone between North and South Vietnam. The unit had been relocated there from Hue after the detachment was overrun during the 1968 Tet Offensive. Six members of the unit were taken prisoner. There were still bullet holes in the van where I anchored evening newscasts.

The detachment was made up of a tight-knit collection of members of all branches: Army, Air Force, Navy and Marines. The unit’s engineers had to keep generators and transmitters running under very trying circumstances, far from Saigon and further yet from the “real world.” What I learned there about doing a lot with a little would heavily influence my thinking about how lecture capture technology should be designed, forty years later.

7

For example, I was able to report the June 3, 1969 moon landing live. Ingenious Det 5 engineers routed a live radio audio feed to the television van’s switcher and jerry-rigged a document camera to show still photos ripped off the AP photo wire. (We did not have access the live video feed of the moon landing. News footage was always at least a day late and consisted of kinescoped CBS and NBC videotape flown in daily from Japan via cargo plane and distributed to AFVN stations.)

The live, NASA video of the moon landing was a technical wonder in itself. Westinghouse developed a special, electronic mini camera that ran on only five watts of power—the current needed for just a single Christmas tree light bulb. The camera was too small to achieve broadcast quality and ran at 10 frames a second rather than 30 and captured 320 lines of resolution rather than 525. There was insufficient bandwidth, anyway, for a higher quality signal.

Bandwidth: The capacity of an electronic pathway/the size of the pipe carrying analog or digital signals. That’s exactly the difficulty technicians faced when they first tried to webcast lecture video on the internet a quarter century later. At first there was not enough available on dialup, copper telephone lines to support anything but jerky, blurry video.

Just like with the first Moon shot available bandwidth would have everything to do with how Sonic Foundry’s Mediasite appliance was designed, leading to one its many present-day flaws.

The Origin of Dual Stream, Dual Window Lecture Capture

Failed Attempts to Achieve Full Automation. Why Some Mediasite Design Flaws Are In Its Genes.

The history of distance learning goes back decades. It begins with printed material followed by

film strips supplemented with phonograph recordings, then videotape, laser disks, CDs and distribution channels like postal mail, phone networks, microwave links, and radio and television networks.

Internet delivery of captured lectures only goes back only about 14 years. It is very recent history. It’s because it’s only been for the past decade or so that sufficient internet bandwidth and streaming technologies enabled pumping full motion video to distance learners. Or, for that matter, YouTube video of cats doing funny things. (During this same time period time even videoconference systems which require so-called codecs at each end have been shifting from POTS/ISDN phone lines to the internet.)

It was not until 1992, just 19 years ago, that the World Wide Web (www) debuted as one of the descendants of ARPANET, and a year later became free to the public. The stage was set. But the real turning point was the introduction of the Mosaic Web browser in 1993 which became Netscape Navigator in 1994. (The success of Mosaic helped lead to the dot com bubble and its burst in 2001, another factor affecting Sonic Foundry’s Mediasite product as we’ll soon see.) Microsoft also wanted a piece of the web browser action and quickly followed with Internet Explorer 1.0 as a part of the Windows 95. Suddenly the general public had easy access to the internet. And just as quickly Microsoft

8

was accused of anticompetitive activity by tying its operating system and web browser. (Just like Mediasite ties its box and server software.)

Back again to Silicon Valley. As I said earlier, Stanford was one of the pioneers, one of the first to capture then stream video of a lecture live over the web. Stanford referred to its internet home study as a “virtual classroom” and named the project ADEPT (Asynchronous Distance Education ProjecT.)

It was 1995 and the internet in terms of general public access was a newborn. Despite the very slow 56Kps dial-up bandwidths Stanford was able to run QuickTime video (.mov format— a competitor at the time to Microsoft’s .wmv and .asf formats) at a super slow .4 frames per second (30 fps is normal speed) in a 320X240 frame. (Compare that glacial-speed .4 fps to the still jerky 10 fps of the moon minicam.) Stanford students could download the video/audio files in batches then replay the lectures after downloading. The download could theoretically take place during a live lecture but viewing was delayed a long time due to the clunky transmission process.

In 1996 I was working on a news set project in Berlin and we needed to get an email of a jpg color photo of the proposed set design from San Diego as quickly as possible. It took IT people on both sides of the ocean working hard and several hours of time (overnight) to transmit the still photo to us in Berlin. Today it would take a few seconds to transmit a far better image or even high res video.

About that time Stanford professor Anoop Gupta was developing a commercial product called VXtreme that emerged out of their Virtual Classroom project. By 1997 the university morphed its fledgling ADEPT experiment into its full-fledged Stanford Online and added VXtreme with its new features.

VXtreme bumped up the frame rate to 10 fps in a smaller 160X120 frame, but also added the streaming/buffering concept where a remote viewer could begin watching even before the download was complete. VXtreme offered two more features: A table of contents so students could access the material they needed, and a new frame inside the viewer interface that occupied about two thirds of the screen that could house the presenter’s manually synchronized rich media graphics and text. (It’s that rich media synching process that other lecture capture products, among them Mediasite, would later accomplish automatically, in real time.) By early 1997 Stanford offered three courses that were on demand and two webcast live.

That same year Microsoft bought VXtreme for a reported $75million, Gupta took leave from Stanford, went to work in Microsoft's research division, and stayed on to become a big wheel there including work on lecture capture. At the same time Microsoft acquired Stanford’s VXtreme it announced a new product called NetShow which incorporated the technology developed by Gupta, “to deliver sound and video images, either stored or live, over the Internet” as a “way of telecasting lectures or presentations,” according to the Chronicle of Higher Education. NetShow eventually became Windows Media Player to compete against RealPlayer and QuickTime players.

Certainly Microsoft wanted in the elearning market. “As part of its distance-learning push, Microsoft formed marketing alliances with elearning companies like ActiveClass, Convene, and Real Education which sold on-line education services to colleges like Web sites where students could register for classes, buy books, and get course materials.”

This marked the beginning of many failed attempts by Microsoft to develop commercial lecture capture products. Microsoft, like several other companies, went off in search of the Holy Grail—an automated lecture capture system that could be launched with the push of a button and run without further human intervention.

Gupta was promoted to become a Microsoft vice president in charge of Education Products & Solutions among other departments. He wrote a 2003 paper comparing Microsoft’s virtual classroom to other seekers of the sacred cup/killer app, noting especially how they differ in robotic camera tracking of lecturers:

9

Obtrusive techniques require people to wear infrared, magnetic or ultra-sound based sensors. Unobtrusive or transparent tracking employs computer vision and microphone array techniques; their quality is approaching that of the obtrusive measures, especially in the context of lecture room automation. Our system relies on unobtrusive tracking techniques. Several projects involve lecture room automation, most focusing on different aspects of classroom experience.

Classroom2000 focuses on recording notes in a class. It also captures audio and video, but by using a single fixed camera limits the coverage and avoids the issues addressed in our research.

STREAM discusses effort on cross-media indexing. Gleicher and Masanz deal with off-line lecture video editing.

Stanford’s iRoom aims at high-end meeting rooms with large displays. Mukhopadhyay and Smith present a lecture-capturing system that uses an obtrusive sensor to

track the lecturer and a static camera to capture the podium area. Because their system records multiple multimedia streams independently on separate computers, synchronization of those streams is their key focus. In our system, various software modules cooperatively film the lecture seamlessly, so synchronization is not a concern. Our main focus is on sophisticated camera management strategies.

Bellcore’s AutoAuditorium is a pioneer in lecture room automation. It uses multiple cameras to capture the lecturer, the stage, the screen, and the podium area from the side. A director module selects which idea to show to the remote audience based on heuristics. The AutoAuditorium system concerns overlap ours, but differ substantially in the richness of video production rules, the types of tracking modules used, and the overall system architecture. Furthermore, no user study of AutoAuditorium is available. Our system, in contrast, has been in continuous use for the past 18 months. McGill University in Canada also built what it called the

Classroom of the Future. It used camera tracking with a “presenter-tracking algorithm which follows the instructor's movements, even when in front of a projected video screen, thereby obviating the need for a professional cameraman.” The instructor’s pen on an electronic whiteboard, provided additional tracking cues for the camera.

The lecture capture part of the McGill system was based on Eclass (earlier known as Classroom 2000) from Georgia Tech, thus it used a conventional web browser interface to display the lecturer and content at remote sites.

Microsoft obtained patents and actually used its fully automated system on its Redmond, Washington campus for employee training. But it eventually fell out of use. For the time being Microsoft seems to have come around to my view that it is not worth the time, money and complexity to fully automate the process. But it has not given up on plain Jane lecture capture. More on that in a bit.

IBM, like Microsoft, worked on a one-button lecture capture product that was used internally before failing to make it to market. Big Blue called it the” e-Seminar Lecture Recording and Distribution System.” Launched internally in about 2000 the system enabled IBM-Research employees web access to videos and slides of lectures, talks, seminars, and presentations.

The same thought that spurred the development of my V-Chassis product also occurred to IBM. To ease classroom setup they designed mobile carts that one of the developers said looked a lot like caskets. The carts held “a tripod-mounted video camera, an IBM VideoCharger video server with MPEG encoder, a VCR for backup, wireless microphones, a client machine to control live multicasts, networking equipment, and all connecting wiring.”

For permanent room setups IBM purchased some of the AutoAuditorium systems originally created in the early- and mid-1990s as a research project at Bellcore (Bell Communications Research) and then later commercialized under the Foveal Systems brand.

Apart from these failed robotic experiments many non-automated and some partially automated concepts were wending their way to market, all of them deploying similar technologies to accomplish the same thing: Capture video of the instructor (talking head) and show that in one window of a web browser viewer interface and also capture images of digital rich media from PCs (like

10

PowerPoint), document cameras and white boards, and display that in another window. The same web browser interfaces had space left over for links to messaging, quizzes, and searches should those features be included.

While some designers were pursing one-button lecture capture others were taking a hard look at the extent to which remote students should participate in the learning process. Should they be able to take electronic notes for later review and for searching key points in the lecture? Should they be able to email questions as the lecture progressed? Should they be able to collaborate and write on a whiteboard just as easy as they could have gone to a blackboard in the classroom? And just how important is it to capture the lecturer’s talking head?

Debate began back then and continues today about the importance of the talking head. That video uses the most bandwidth leaving less left over for the rich media visual aids. Does seeing the instructor do any good anyway? Given the fact video was difficult to transmit over the web the first efforts at lecture capture learning centered on transmitting audio-only with whiteboard graphics, essentially a digital blackboard, so remote viewers could see and hear what the instructor was illustrating.

Back again to Silicon Valley. While Xerox’s PARC was inventing key components like GUIs, multiple windows and small-form computers that enabled lecture capture boxes and viewer displays, it was also doing whiteboard research. If someone did the genealogy of most all present-day commercial lecture capture products it would probably lead back to PARC’s lesser known whiteboard experiments.

PARC’s whiteboard research began in 1987 partly funded by, who else, ARPA, the US military. It was not beneath Xerox’s dignity to accept some government handouts. Lots of ideas were kicked around. One was to be able to capture writing on a whiteboard directly onto a computer and to save it before someone erased it. Another was to write with an electronic pen on a computer screen and have it appear on the whiteboard. Finally, why not connect whiteboards in different rooms, even different cities, so that people could collaborate over the internet without

being in the same room? By 1994 Xerox was selling its electronic LiveBoards with sixty-seven-inch diagonal displays that

users could access or write on from across a room with special electronic pens. They cost more than $40,000 each. (In 1996 Xerox’s MeetingBoard became an add-on to Microsoft’s NetMeeting.) This distance collaboration was just another form of distance learning, a thought that became immediately apparent at campuses across the United States. Why not have a LiveBoard in a classroom and instead of a student having another one in his or her dorm room they would merely access the event on a PC, and learn passively rather than participate? So instead of whiteboard to whiteboard, active to active participants, like PARC did it, it could be whiteboard to PC.

A slew of video lecture browser and meeting capture and delivery research projects began popping up at universities all over the place. One was the Unibroswer headed by Michael Danchak at Rensselaer Polytechnic Institute. He called the concept “guide on the slide.” A box with video of the instructor appeared on the screen next to a box that showed the PowerPoint slide or other rich media the instructor was talking about. (Later Danchak called Unibroswer “old hat” and told me he prefers a relatively new method of using cartoon-like avatars to roam the screen and point out stuff on the slides.)

In 1996 Cornell developed a similar concept called Lecture-Browser. The Cornell Lecture Browser, which was also used by Berkeley as the BMRC Lecture Browser, showed slides with a synchronized RealAudio and RealVideo stream. It used JavaScript to pre-load the slides into the browser

11

cache to reduce delays in slide transitions. All slides had to be converted manually by the author to a browser-supported image format.

In about 1996 Georgia Tech brought out Classroom 2000 (later called eClass) that transmitted a lecturer’s audio and his or her Xerox Liveboard material over the web using a web browser interface. The initial emphasis was on whiteboard sharing and student/viewer note taking, not on lecturing or capturing the body language of the lecturer. Eventually they added the option of showing video of the lecturer in a small window on the screen.

Dr. Gregory Aboud, who was one of the chief developers of whiteboard-centric Classroom 2000, told me he thought his invention was the first of its type but that he didn’t seek any patents on the concept because he was not interested in commercialization and also thought earlier work at Xerox’s PARC “was going to trump any patent claims.” Georgia Tech acknowledged concurrent projects like ADEPT at Stanford, Bell Laboratory’s AutoAuditorium (same lab that invented the transistor) and Cornell’s Lecture Browser that did automatically generate a table of contents that provided synchronized indexing into the slides.

Carnegie Mellon developed its Coursecast system in about 1998, a software program developed by two students. In 2008 the university commercialized Coursecast as Panopto. There was also some collaboration on a similar product developed in Germany called virtPresenter. (Coursecast was not designed for live webcasts but could operate in three modes: lecturers utilizing PowerPoint presentations, lecturers using screen capture presentations, and two-camera presentations for faculty who prefer using blackboards.)

In another Carnegie Mellon department yet another lecture capture product was germinating, but initially not as a video capture front end but as a video management back end. What today is known as Mediasite was part of Carnegie Mellon’s heavily government- subsidized Informedia Digital Video Library Project. Yes, and here we go again, one of Informedia’s financial benefactors was ARPA.

Carnegie Mellon’s Informedia first developed technologies to search library video archives and after that to search worldwide television newscasts for government intelligence gathering. The university commercialized its video management technology as ISLIP Media, Inc. (Integrated Speech, Language and Image Processing) which in turn opened the web site MediaSite.net in 1997. That media site offered search, preview and purchase of online video and audio content. It was used by NBC, BBC, Turner, IBM, the US Army, Navy, Air Force and Livermore National Labs. In March 2000, Mediasite was bought by Sonic Foundry for an estimated $8M and was later modified for lecture capture and integrated with Sonic Foundry’s front end appliance.

Dissatisfaction with the cost and proprietary nature of emerging commercial lecture capture products led some universities to form an open source consortium. Participating universities jointly

developed free lecture capture and video management software. Georgia Tech that did eClass, UC Berkeley that developed a variation of Cornell’s concept, MIT, the Stanford School of Medicine among others formed the Opencast community that sticks its free, open source software online that anyone can download. (The free software is not for the faint of heart, however. Its setup requires experienced IT personnel.)

Virtually all of today’s lecture capture products, whether commercial or open source, use the same basic viewer interface concept. All have similar looks on the remote site viewer’s

12

computer web browser interface: a small window with video of the presenter next to a larger window with the lecture PowerPoint slides, whiteboard or other rich media. All of them went down the same path a path so widely traveled it’s dug a fairly deep rut.

Another common trait is that none of the primary lecture capture products include the required AV equipment as standard or optional equipment. It’s left to the university buyer to sort out microphones, audio mixers, cameras, lighting, etc.

One reason for the exclusion of AV components may be that a lot of systems are sold by resellers and systems integrators. They might want to profit from the markup and labor for designing and installing the AV side of a complete system.

Another contributing factor could be that IT people, generally, know damn little about AV. Every major lecture product on the market today originated out of the IT not AV space. And it shows.

Is It Too Late for a Course Change?

Should Others Follow the Mediasite Leader? Stanford, like most traditional universities, is heavily “siloed”. It’s divided into corridors of

power, walled-off colleges and departments, bunkers from which turf wars are fought. Protecting old teaching ideas can more important than advancing new ones. One instructor at Stanford may make good use of state-of-the art classrooms while another goes blah, blah blah in front of the blackboard even if the classroom is packed with presentation technology like document cameras, electronic whiteboards, maybe even video cameras.

Even UCLA, which was one of the four original members of the ARPANET fell into cross-silo squabbling when ARPA first provided funds to link three computer centers inside the school. Observers “watched in disgust as the project fell victim to academic bickering: none of the three UCLA centers really wanted to work with the others…” (The Dream Machine, M. Mitchell Waldrop.)

My guess it is not much easier today for universities to innovate except that now the internet is poised to do to universities what it is doing to newspapers. The silos can be voluntarily dismantled. Or they can come crashing down. Traditional universities can begin deploying cross-campus lecture capture. Or wait until it’s too late.

Even though lecture capture was invented in the United States, it might be end up that foreign institutions of higher learning deploy it best and widest. Consider the story of invented-in-USA integrated circuits.

Before integrated circuits found their way into newly invented personal computers and later into lecture capture appliances, they had already begun replacing vacuum tubes in consumer electronics like stereo receivers and television sets.

Even though ICs were invented in America, it was the Japanese who better commercialized them, eventually wiping out the television manufacturing industry in the United States which clung too long to vacuum tubes. For hundreds of thousands of GI’s who served in Vietnam, Japanese consumer electronics were no longer cheapo and failure prone (like my first Japanese transistor radio in about 1957) but cutting edge, inexpensive, high tech with their IC circuitry, and reliable.

Take the thousands of GI hooches in Vietnam. No air conditioning. Crude construction. Few barriers to monsoon rain coming in sideways and swirling dust coming through every crack. Even so, fancy Pioneer, Panasonic, Sansui, Kenwood stereos, speakers and tape decks, Sony Trinitron TVs, transistor radios, dirt cheap, often layered with dirt, populated just about every hooch. No running water or indoor toilets. But electronics that could blow the hooch doors off. Good morning Vietnam! Good bye American consumer electronics factories!

13

Vacuum tubes may be to traditional universities as transistors are to online schools. Traditional universities may cling to their old ways until it’s too late.

While perhaps most of the blame goes to change-adverse universities a lot still has to go to lecture capture makers for the way they design and market their products. Between universities not knowing how to buy the technology and makers not knowing how to make it and sell it no wonder hundreds of thousands of lectures continue to disappear into thin air.

Because the Mediasite with its lecture capture appliance and server software claims to be the market leader and probably spends the most on marketing, its design and sales tactics deserve the most scrutiny and criticism.

Sonic Foundry of Madison, Wisconsin is a publicly traded company (NASDAQ ticker symbol SOFO) and claims its Mediasite product is the “uncontested global leader for rich media webcasting and knowledge management” for internet delivery of campus lecture videos. Media technologist Paul Riismandel calls Mediasite “the 800-pound gorilla of the lecture-capture world.” Sonic Foundry boldly predicts “every classroom, every meeting hall, every conference room will have Mediasite in it, just as you see projectors now.”

When the Mediasite appliance came to market in 2003 it was not the first commercial box to capture, stream and display video of a presenter and rich media. Pinnacle’s StreamGenie beat it to the

punch in 2000. StreamGenie’s viewer interface switched between full screen VGA (rich media) and full screen video (talking head) where the Mediasite box showed the two streams at the same time in separate frames. The ill-fated Communicator 1000 and Tegrity WebLearner boxes also preceded the Mediasite ML Recorder. StreamGenie went off the market about the time the similar Mediasite appliance debuted.

For classroom capture Sonic Foundry began with a software-only product that was installed on a PC. A 2002 press release said “Sonic Foundry and Boxx Technologies (high performance Windows 200 workstations) also are working together to offer a combined solution for Sonic Foundry's MediaSite Live(TM) rich media presentation system.” Thereafter Sonic Foundry brought an actual appliance to market, the ML Recorder.

Mediasite quickly became the best and most commercially successful box and spawned lots of knockoffs like Accordent’s Capture Station released in 2005. Others followed: Winnov’s C-Box, VBrick's Rich Media Studio, nCast’s Telepresenter, 323Link’s EDUcast, Echo360’s capture appliance, Qumu’s Capture Studio, and Matterhorn’s open source make-it-yourself capture appliance. All of these appliances are, essentially, single-purpose, small form-factor computers running Windows or Linux operating systems. They could all be housed in normal computer cases but, instead, are housed in smaller, custom-made boxes or off-the-shelf shuttle boxes. (Besides the appliances there are lots of software-only solutions like Tegrity and Panopto that schools do load onto regular PCs.)

While the idea of capturing lectures and displaying presenters and visual aids in separate frames in a web browser window originated in the early and mid 90s, as explained earlier, Sonic Foundry developed the electronic wizardry to accomplish synchronization without having to preload slides plus added features. It was all inside a relatively compact, portable box that anyone could buy on the commercial market. Even if it carried a steep $25,000 price tag.

But the classroom box was only part of the package. By itself, unlike some competing concepts, the Mediasite box but itself could only record for later on-demand playback, not steam live. Sonic Foundry concurrently developed server software based on its MediaSite acquisition from Carnegie Mellon, to achiever live streaming, to manage the video files for on-demand access and to add features like restricting access, searching, polling, Q&A, semi-automated closed captioning, and reporting.

14

Sonic Foundry did not start out in lecture capture. It was forced into it. Sonic Foundry was founded in 1991 in Madison, Wisconsin by self-described “weirdo software

artists” as an audio editing software company. Its founders developed various types of digital audio editing products, thus the word sonic in its name. (I used its excellent SoundForge editing software for several years before I replaced it with Audacity.)

Sonic Foundry’s founders with their successful, innovative products succumbed to the temptation of going public. They got aboard the 1995-2001 dot com express when greed and wild speculation spurred investment in high tech startups in the hope one would turn out to be the next Microsoft or Mosaic. Sonic Foundry went public in 1998, initially on the New York Stock Exchange and then moved to NASDAQ. It began buying other companies right and left and pushed the combined value of its shares up to $2 billion before the bubble burst. (Today that value, its market cap, hovers around $50 million making it a micro or nano cap/penny stock.)

“Tulip bulbs don’t last,” said Sonic Foundry CEO Rimas Buinevicius while explaining the dot com crash and this company’s free-fall stock price in 2001. He admits it had been the strategy of Sonic Foundry to “capitalize on the Tulip Bulb craze” referring to the speculation mania in Holland in the 1600s when the price of tulip bulbs skyrocketed tens of times over actual value. When Holland’s tulip bulb market crashed thousands of investors lost life savings and homes. (The tulip bulb craze is featured in the 2010 movie “Wall Street: Money Never Sleeps.”)

One of the companies Sonic Foundry bought in 2000 for $8 million was MediaSite, that commercial spinoff from Carnegie Mellon University’s Informedia project. Taxpayers had subsidized Informedia’s growth to the tune of many millions of dollars as the school developed ways to store and access video. Some of the research revolved around audio speech search, that is the ability to search not based on typed text but on the spoken word recorded in audio or AV files using so-called speech recognition software.

The dot com crash almost sunk Sonic Foundry. But unlike some of the highest dot com fliers that were blown to smithereens, it actually had some proven, valuable assets: the audio editing stuff and a video transcoding company it had purchased. In 2003 Sonic Foundry sold the products that were generating income, paid off several million in debt and ended up with a nice nest egg to buff up the company it had bought earlier and stuck on the shelf: MediaSite.

Sonic Foundry management had a vision: lecture capture’s time had come. Buinevicius said they were “Henryfording—mass producing—the way of making lectures available on line.” Except rather than start out with a Model T they went straight to making a high end model. “Ultimately you view us as the Mercedes, the BMW of webcasting,” Buinevicius said. Other times he refers to Sonic Foundry as the Cadillac of web video management, comparing his product to that luxury car, for example, when he disparaged the quality of YouTube video after Google bought it. (YouTube is a free way of doing what the costly Mediasite server software does but without power steering, power brakes, and electric windows.)

Sonic Foundry was betting the farm on its embryonic product, the media management backend, to which they would add the capture appliance front end to provide an end-to-end solution. But it was left with a company name—Sonic Foundry— and a product name—Mediasite— that had next to nothing to do with event capture.

Carnegie Mellon’s MediaSite could perform spoken word search as one of its tools to search large audio and video archives. That’s where, say, a math student could open a video file of a lecture on geometry, type in a key word like “congruent”, then instantly go to the places where the instructor used the term. And then watch the clips.

15

Lots of companies had dabbled in video speech search with varying degrees of success: Virage, Nexidia, StreamSearch (Comcast), Singingfish (AOL), SpeechBot (IBM) along with universities—VMR, Cambridge and WebSeek, Columbia.

For a while Sonic Foundry demoed the feature it called Multi-Modal search it on its web site. As a search tool it seemed to work well. But the demo was quietly removed from the web site and Multi-Modal search was never made a part of its Mediasite package.

Why? Maybe because it is not 100% accurate, not nearly as accurate as text searches. But more important, the computing power and time to do it is considerable. Apparently phonetic search was not worth the added cost of the option along with the processing time it takes. It might also have something to do with the fact Carnegie Mellon’s MediaSite product apparently did not use speech mining technology that it developed on its own but rather used IBM’s ViaVoice. ( In 2003 IBM sold ViaVoice to ScanSoft, now Nuance, which owns the competitive product Dragon NaturallySpeaking.)

The bigger pot of gold at the end of a rainbow would be accurate, automated speech-to-text transcription. This goes beyond phonetic search. It’s where a computer rather than a person “transcribes” words that are spoken on audio and video recordings, eliminating manual transcription. Thus fully computerizing functions like closed captioning.

Some products can do speech-to-text but not with near the accuracy needed for lecture and courtroom capture. The legal industry would love to replace expensive shorthand stenographic court reporters with audio or video recorders and speech-to-text software to create transcripts of depositions and court proceedings. (Many courts are switching to electronic record-making anyway leaving it to human transcribers if text is needed.)

I did a consulting project for the Atlanta-based Nexidia, Inc. whose speech recognition/analytics/audio indexing/ mining software can does spoken word search but not speech-to-text transcription. Nexidia’s speech search is very cool. But so far only governments and large businesses can afford it. Affordability for the masses could still be years away.

Mediasite does offer closed captions as an option that produces an accurate text transcript of classroom speech and also permits a search of that spoken word transcription. But its closed caption option bypasses any technology Sonic Foundry has on the shelf.

Mediasite farms closed captioning out to the company Automatic Sync Technologies (AST). How does AST do the transcription? With computers?

No. With old-fashioned manual labor. They use people to listen to the audio and type what they

hear. It takes about three days to generate a text file that then can be also be used for search besides the on-screen closed captioning. AST also does closed captioning for Mediasite competitors Echo360, Panopto and Camtasia. (The Mediasite closed captioning is partially computer automated in the sense that the audio file is automatically uploaded to AST and then, after manual transcription, is automatically downloaded and synched to the video.)

Where Mediasite’s lecture capture system may have been somewhat state-of-the art in 2003, today it’s outdated kludgeware. Its four most serious flaws are high price, poor rich media capture, defective viewer interface concept, and the tying or bundling of its proprietary front end and back-end products.

16

High price. An appliance for a single classroom lists for about $25,000. (But they usually sell for half that price.) And that does not include the expensive server software for live steaming and other features, AV equipment like cameras, microphones and audio mixers, installation costs and pricey, virtually mandatory annual service contracts. Those costs may be low enough for some universities that just want to equip a few rooms. But it is probably prohibitive for most universities that might want to democratize the benefits of lecture capture by deploying it campus-wide.

Poor Rich Media Capture. The second, significant, perhaps fatal flaw is the way Mediasite records, streams and displays the rich media, which is its main raison d'être. Instead of capturing PowerPoint slides or whiteboard writing in full motion like it does the talking head, Mediasite merely takes jpeg snapshots of them. It can only capture still (not animated) images at 3-4 frames a second. It takes rich media capture at at least 15 but better yet 30 frames a second (fps) to replicate full motion. That means Mediasite captures virtually no rich media action: no PowerPoint animations, no smooth writing on a whiteboard, no video clips. (Some users do deploy a clunky workaround to switch rich media to the video presenter window. But because of the window’s small size users have to accept less clarity to get the motion.) That’s probably why most of the Mediasite presentations you watch have rich media images that are lifeless and well out of close synch with the presenter. (NCast Corporation—Telepresenter and 323n Link—Educast--as examples, sell appliances they say do stream full motion rich media.

Rumors persist that Sonic Foundry plans to introduce full motion VGA capture but if it does it will be following not leading.)

Defective viewer interface concept. Sonic Foundry followed previous, IT-centric methods for displaying the rich media and talking head at the so-called far site, on viewer’s displays. Monkey see, monkey do. By sticking the presenter and

his or her visual aids in separate frames the presenter cannot directly interact with those aids as they do in the classroom. The result: stiff and lifeless presentations in the online version. It’s like lecturers are locked in a prison cell and Sonic Foundry threw away the key. Adding insult to that injury is the additional clutter on the Mediasite viewer interface which distracts viewers and wastes perhaps more than half the screen space. Mediasite’s developers were probably preoccupied with data transmission, with herding 1s and 0s and less concerned about communicating words and images. (Entirely missing in traditional lecture capture’s DNA is the influence of visual communications experts who know how to effectively capture the light and sound waves before they are encoded and transmitted. That’s why an IT geek may look at the Mediasite user interface and conclude nothing is wrong.)

Product Tying. The fourth major flaw with the Mediasite concept is a university’s requirement to buy the front end box and back end server software as a proprietary bundle. Yes, one reason the two were tied together is to make their operation more seamless. But there is an ulterior motive. Once a university goes to the expense and trouble of setting up Mediasite server software it becomes financially locked in to Mediasite for further expansion, because Accordent or Echo360 boxes, as examples, will not run on Mediasite server software. (Accordent chose not to deploy the vendor lock-in strategy and made its back end, media management system compatible with other media creation systems besides its own.)

Universities could choose from many other commercial and open-source media management

systems like Kaltura that are media agnostic, not limited to one box brand. Not to mention delivery platforms like YouTube University and iTunesU which are agnostic and free, although they lack the whistles and bells some schools may want.

17

What about Mediasite’s seamless argument? After the federal government sued Microsoft for bundling Internet Explorer with its Windows operating systems a settlement permitted Microsoft to continue providing both in the same package. But the deal also required Microsoft to share its application programming interfaces with third-party companies. That’s why competing search engines like Google and Firefox work seamlessly with Windows. The seamless argument would fade away if the link between Mediasite’s front and back ends were not proprietary.

The list of Mediasite’s less major conceptual flaws goes on:

• All or nothing package. The failure to offer many of its features à la carte so that a university does not have to pay for features that it does not want or need.

• Financial instability. Sonic Foundry, as a publicly traded company, is a so-called micro cap. It has flirted with being delisted from NASDAQ for years. Since it introduced its Mediasite products in 2003 it has only eked out a small profit in two quarters in 2010 while racking up millions in losses over the years.

• It’s failure to innovate. The Mediasite box, despite lots of minor improvements, is still very much the same basic design it was eight years ago. There have been no revolutionary changes. Same old. Same old.

• Unethical, possibly illegal anticompetitive marketing tactics such as bid-splitting, sole-sourcing and bid specification rigging involved with sales. Those tactics rob universities of truly competitive bidding which could result in lower costs and better products. There probably are some customers who actually want and use every one of the Mediasite

Recorder’s features. They would have checked every box even if the features had been optional. But the high cost of the feature-rich system might help explain why Mediasite does not sell in the tens thousands a year. (Sonic Foundry only sold 1,023 capture units this last fiscal year.)

Maybe another PARC parallel sheds light on Mediasite’s bloat problem: PARC, on behalf of its Xerox parent, developed the Star office computer

system, sold as the Xerox 8010. When it came to market in 1981 it was far and away the best office computer available. But its designers made three critical errors, according to author Mitchell Waldrop. First was its “rampant case of feature-itis” that left the Star loaded with “every neat think they could think of.” Second, it was a closed system, all hardware and software had to come from Xerox enabling the company to “lock in its customers.” And, third, “the designers

passed up several chances to do something simpler.” That same year, 1981, IBM introduced its Personal Computer. It only

had a fraction of the Star’s features: No on-screen icons, no mouse, and no highly graphic display. The IBM was much more difficult to operate and came with the crash-prone Microsoft MS-DOS operating system. But it was far cheaper, about $3,000 depending on options, compared to the Star’s $15,000. And unlike Star, the IBM’s innards were wide open; third party vendors were free to write and sell software for it. Plus Microsoft was free to sell its DOS to what became a spiraling number of IBM clones.

Let’s see how the Mediasite recorder stacks up to that scenario: It is also overpriced. It also has too many features. It is also a closed system that thwarts development of third party apps. And it also locks in customers. Just like the Star Mediasite does enjoy some sales. The Star sold just 30,000 units over its 8-year life. IBM sold ten times that number in just over half the time. And that does not count sales of the wildly successful clones like Compaq.

18

It’s why my V-Chassis was designed to be low-priced, is an open box with off-the-shelf components, and comes as a base model with extras optional and no mandatory annual service contract.

So does that guarantee V-Chassis’ success against the Mediasite? No. IBM did not anticipate the demand that was pent up for a cheap PC. At the moment it

seems that demand for lecture capture systems is weak at best. Demand is latent, not pent up. There are very reasonably-priced software-only solutions on the market yet they are not being snapped up like hotcakes. Something other than price is holding back demand. And even though the less expensive, software-only solutions are harder to set up and configure relative to their hardware brethren, they are vastly easier to use than the first IBM PCs, which flew off the shelves anyway leaving millions of users trying to figure out what a DOS prompt was, how to move a cursor around the screen without a mouse and what nonsensical commands they needed to issue.

The Utah Lecture Capture Procurement Investigation

Vietnam in the 60s and early 70s was a war zone. Even though I’d been taught in basic training

to kill, luckily I never had to shoot at anyone. And no one shot at me unless you count crude rocket attacks (misguided missiles) that were Vietnam’s equivalent of IRAQ’s pesky and lethal IEDs. Being there did heighten one’s sense of survival, an experience that might pay off back in the real world.

It did help prepare me for my work as a legal consultant where bitter fights between competitors spilled into the courtroom. Like the work I did on behalf of Caldera when that Utah company sued Microsoft. There was lots of bullet-dodging with that case. The Caldera v. Microsoft civil antitrust case ended up in a $200 million-plus settlement in favor of Caldera and Novell co-founder Ray Noorda but not before both sides suffered a lot of casualties.

Bill Gates may have looked and talked like a computer nerd but he conducted business like a robber baron. His company bought out or crushed a lot of its competitors. The mantra at Microsoft in the 90s was “kill the competition.” The word “kill” showed up in Microsoft emails, in judges’ rulings against Microsoft and in utterances like this one from Seattle reporter James Wallace: “[Bill Gates] not only wants to win, but he wants to kill the competition. He wants to bury the wounded.” One of many key emails in the case was one from former Microsoft VP James Allchin who wrote, “We need to slaughter Novell before they get stronger."

I did some of the investigating in Germany where Caldera’s DR DOS operating system had been dominant over MS DOS, because it was better. But Microsoft’s illegal tactics killed off the superior DR DOS as the prevailing player in Germany.

When it came time to bring a tech product of my own to market I had a fairly keen eye for anti-competitive behavior. The very first lecture capture request for proposals (RFP) I looked at in early 2010 bore the hallmarks of a rigged bid process. Weber State University in nearby Ogden, Utah wanted to buy two or three lecture capture appliances, “Cadillac” systems as one of the review committee members called them. It seemed to me no coincidence that Sonic Foundry CEO Buinevicius described Mediasite exactly the same way. “We like to really describe ourselves as the Cadillac of the industry,” he would say.

It looked like Weber State favored, even decided on Sonic Foundry’s Mediasite before the RFP was issued. I learned Weber was demoing the Mediasite appliance and server software leading up to the issuance of the RFP. I spoke to Utah’s Mediasite rep who had lost a Weber State bid two years earlier to Echo360. She told me she already knew Weber planned to buy Mediasites. How would she know that before any bids come in? Unless the fix was already in.

Weber was not my first encounter with bid rigging in Utah government.

19

As a legal consultant I helped attorneys use trial presentation technology such as helping prepare the PowerPoint opening for the Caldera v. Microsoft case and lots of others. (Ironically our side was going to use Microsoft’s PowerPoint during the trial and attorneys for Microsoft were not.)

In the of course my work in Utah courtrooms I discovered that the courts had issued several million dollars worth of no-bid contracts for courtroom electronic recording equipment, right under the nose of the Chief Justice of the Utah Supreme Court who administers the courts. I figured if illegal, anticompetitive procurement could take place within the court system it would not be surprising if it could happen at one of Utah’s public universities.

Another indication the Sonic Foundry might play procurement hard ball is the company’s war-like marketing attitude. Buinevicius once told a group of college students that Sonic Foundry faces a world-wide battle. “It’s about engaging the enemy, winning the war,” he said. If he wanted the company to survive he said he had to think about strategy as war.

In 2007 Sonic Foundry stuck a spoof video on YouTube titled “The Slashing.” It told the story of three competing webcasting solutions that responded to a call to meet in the woods. The three—“Kludgey The Homebrew”, “No Backend” (whose bare butt is exposed), and “Lack of Navigation”—confront a sinister Slasher in a black, hooded robe named MediaSite. The Slasher smashes in Kludgey The Homebrew’s head in using a Mediasite Recorder—blood spewing everywhere. Then the Slasher lassos No Back End using RCA cables choking the life out of him as blood spurts out of his mouth. And then he throws a Mediasite software CD like a frisbee that lodges in Lack of Navigation’s head, killing him.

The slasher video isn’t quite so funny in light of the dirty tactics Sonic Foundry uses in real life to, figuratively speaking, smash its competition.

When the Weber State RFP specifications were released it was even clearer Weber already favored the Cadillac system I had been told about. The university wrote into the specification that the review committee would not considerer price during the first review round. That would have eliminated Panopto, free lecture capture software, for example, because it would not have met as many of the specifications as Mediasite.

One way of killing competition is to subvert pro-competitive procurement laws, to figure out ways to put one’s thumb on the scale purchasing departments use to weigh bids and proposals. Like find someone in, say, an IT department who would be evaluating lecture capture alternatives and has some influence on how bid specifications are written.

Bid splitting is one way to rig the process. If specifications for a product could be split so each part came in under the minimum bid requirement price, then open bidding could be bypassed altogether and the preferred product bought outright without facing competition. At least two California universities appear to have purposefully split its acquisition of Mediasite recorders and software in order to escape looking at alternatives.

Sole-sourcing is a method several universities have used to dodge competitive bidding. Without much effort I found that the University of Wyoming, Dixie State College and Uniformed Services University (USU) all awarded sole-source contracts to Sonic Foundry’s Mediasite. Sole or single sourcing can be a legitimate way to avoid the cost and delay of a bid process when it’s clear only one product meets requirements. But how can Mediasite be sole-sourced when it has perhaps a dozen competitors? USU, for example, already had some Mediasite recorders. When it came time to expand it justified buying more Mediasites without going out to bid because no other product is compatible with Mediasite’s proprietary design. That’s vendor lock-in in action.

20

Bid specification rigging is yet another way to favor a particular product or vendor. It’s also called “tailor-made tendering” where specifications are “tailored” or fashioned to favor just a few or even just one vendor. It involves collusion between a vendor and an insider who helps formulate specs. It requires trickery and deception, some way to make it appear that the specifications promote a wide range of bidders when, in truth, only one is favored.

IBM, before it put small, personal computers on the map, once monopolized the mainframe computer market. After years of being worn down by Justice Department trust busters, IBM had to adopt policies for its employees and resellers to reduce or stop any procurement collusion involving bid specifications. One book described accusations “leveled against the giant in the trade press” that IBM reps rigged bid specifications by means of tailor-made tenders.

“Collusion between sellers and employees of the purchaser is a more important threat to competition in the computing industry than collusion among sellers. For example, if an employee of the purchaser is convinced that an IBM computer would be best for his company, or for him personally, he might be tempted to confer with IBM to ensure that the specifications put in the call for tenders are requirements which only an IBM machine could adequately meet.” (The Impact of Publicity on Corporate Offenders by Brent Fisse and John Braithwaite, 1983.) Bid specification rigging is a fine art. The spec sheet has to look like a procurement is wide open

when it’s really closed. The trick is to bury a couple of requirements that only one vendor can meet among requirements that lots of vendors can meet. And to know someone on the review committee will focus attention on the specs no other competitor meets.

Is that what Sonic Foundry does? Split bids? Rig specs? Engineer sole sourcing? A former Sonic Foundry sales manager told me that’s what exactly they did. “They (Sonic

Foundry) are very good at creating bid specs and bid spec templates so if a customer really likes it they will walk in and say here’s the language and things you need to use if you want to make sure that we’re part of the bid spec,” he said.

I told him it seemed to me Sonic Foundry found a way to get university insiders to write specs to exclude competitors, that once the specs are put out to bid it’s almost certain that Sonic Foundry is going to win:

Answer. Sure. Question. How did that work? Answer. It’s an old IBM trick. They’re putting kill points into the spec. Question. What are kill points? Answer : Kill points are, basically, it’s an old IBM trick from the 60’s. What they’re doing is that you always put things into the spec that you know your competitors will not have.

The former salesman said he had a template that was written by Sonic Foundry employees in

Madison that he could show or give to someone at a university who has influence over specifications, to help get them insert the kill points. After that, he said, anyone else bidding would become “column fodder”, that is a bidder who merely helps a university create the appearance of open procurement because it needs a certain number of bidders to participate.

At IBM’s peak in 1952 it pulled 64 per cent of the electronic data processing revenues. Big Blue was not just suspected of colluding with spec writers. Another complaint was that it also purposefully tied its hardware and software. Connecticut professor Marc A. Triebwasser wrote:

Another practice that IBM used to maintain its customer base was bundling. It would

not only provide the hardware, but also the software--all as a single package. As a company

21

became dependent on this software, it could only use IBM equipment to upgrade. If a company wanted to use a computer built by another manufacturer, it would have to invest in a totally new software system Of course that is what Sonic Foundry does with its recorder hardware and server software. The

University of Wyoming, for example, began specifying only Mediasite in bids after making initial purchases. “Because we’re not an Accordent shop I can’t buy Accordent appliances and run them though my Mediasite sever,” the Wyoming employee over lecture capture said. “It’s proprietary. One won’t talk to the other.”

Question: So once they do get their foot in the door they pretty much have a monopoly? Answer: Yeah. In the book Dealers of Lightning Michael Hiltzik wrote that early on manufacturers did not see

any advantage to marketing machines even remotely compatible with their competitors. “Once IBM sold a system to United Airlines it could rest assured that the frightful effort or rewriting software, retraining staff, and moving tons of iron and steel cabinets around would make United think long and hard before replacing its system with one made by, say, Honeywell,” Hiltzil wrote.

Hiltzik concluded that “standardization has helped make computers a mass market phenomenon.” “It allow users to be reasonably confident that a program bought off the shelf will work properly regardless of who manufactured their computer just as they know they will find the accelerator and brake pedal in the same location regardless of whether their car is a Ford or Chevrolet.”

Was the Mediasite system purposely designed to be incompatible with competing devices so it’s difficult for universities to mix and match different brands? So once Sonic Foundry had it foot in the door swung wider open?

Duh! Sure. Sonic Foundry’s foot-in-the-door sales strategy is aimed at locking universities into the

proprietary Mediasite concept for ongoing expansion. Stock research analyst Jeffrey Walkenhorst (www.commonstocksense.com), who owns Sonic Foundry shares, sees Mediasite’s high price and its proprietary, lock-in strategy as pluses:

“ Sonic Foundry’s version/platform upgrades and broad Mediasite offering extend market

leadership and further entrench existing customers that are expanding campus footprints,” Walkenhorst blogged. “We believe Sonic Foundry’s high gross margins indicate the proprietary nature of Mediasite technology. (Emphasis added.) I wondered if Walkenhorst knew whether Sonic Foundry was using unethical or illegal tactics to

achieve those high margins. I asked the question via email and he answered:

Packer Question: What do you know about rumors about how Sonic Foundry gets some of the sales it does at some universities where it sometimes, purportedly, lobbies potential buyers, like a university IT or distance learning person who meets them at a trade show, to influence their purchasing departments to write specs in Sonic Foundry’s favor? Is that ethical or legal? If that’s one of the ways Sonic Foundry competes could that expose them to risk if they get caught trying to rig specifications in their favor?

Walkenhorst Answer: Probably true. Nothing wrong with this. It's the Dale Carnegie way, How to Meet Friends and Influence People.

22

Meeting friends and influencing people. Certainly Sonic Foundry is not the only company and Mediasite the only product where sales reps try to make friends who have some say with purchasing decisions. And then try to influence them. It’s why they make sales calls and give demos. Certainly they want the specs of their product included or covered by a bid specification.

Where is the line crossed? Probably when a vendor tries to get an insider to tailor specs that exclude competitors and include features that are proprietary and promote vendor lock-in.

Just a few months ago, at the close of the 2010 fiscal year, North Carolina State University decided to expand its existing lecture capture across the campus. It already had 50 Mediasite appliances. Did it look to Accordent or Echo360 or perhaps a dozen other possibilities to expand? No, it purchased 80 more Mediasite recorders, all in one lot. With annual maintenance fees running at something like $1250 per unit, that gives Sonic Foundry an ongoing revenue stream of $162,500 a year every year. It’s NC Sate worshiping at the altar of Propritarianism .

You’d expect the NC State bonanza to have been trumpeted by Walkenhorst on his web site. But Walkenhorst is no longer discussing Sonic Foundry in connection with what he calls “global trends in …lecture capture.” On November 24, after he had promised some news about the company, he announced he was limiting his analysis of Sonic Foundry. “We hoped to provide additional commentary headed into the company's earnings report last Thursday morning,” he wrote. But he said “along with our colleagues, we received information last week that limits us from blogging on the information technology (IT) sector.” He did not say what the information was. Could it have been the fact Mediasite sales are under investigation in Utah?

Would North Carolina State or the University of Wyoming lock themselves into buying a motor vehicle that is so unique and proprietary that it becomes infeasible to buy other brands for subsequent purchases? University transportation fleets are usually made up of competing makes like Fords, Chevys, Toyotas and other cars that can peacefully coexist in the same motor pool. Why should lecture capture be any different?

Taking the car analogy one step further, what university would buy one brand of car that had the brake pedal left of the accelerator pedal and another brand where the pedals were reversed, creating user confusion? Rather than craft bid specifications for lecture capture systems that favor over-priced, underperforming systems universities should draft specifications that favor lower priced, cross-compatible products. Specs like that would broaden the range of potential bidders.

Here’s a list of rules and specifications I recommend for the IT components of an AV/IT lecture capture system in order to foster competition and cross-product compatibility:

• Front-end lecture capture appliances/PC software and back-end media management

servers and software must be specified and procured in separate RFPs or RFQs. • Media management systems—back-end server software—must be compatible with

most lecture capture appliances and software-only applications and procured separately from the boxes that feed them.

• Lecture capture hardware and software must be designed so that proprietary media management systems are not required for internet distribution.

• Captured video must be compatible with open source media management systems, like Matterhorn and Kaltura and with social media content delivery such as YouTube and iTunes.

• Capture components must be designed so that trained AV or IT personnel could service them (replace and upgrade parts) so that service contracts are not mandatory, just optional.

23

• Bidders must submit prices for their optional, annual-fee service contracts and for per-call tech support and service, disclosing the full and true cost of ownership. (Pay per call phone support must be an option besides annual service contracts.)

• All products must disclose the length of warranty including firmware upgrades and the cost of such upgrades after the warranty expires.

• The time and nature of all contact that vendors and resellers have with anyone involved in formulating bid specifications must be fully disclosed to purchasing departments and to all potential bidders.

• All products must be given equal opportunity to demo if any pre-purchase trials are conducted or allowed. All contacts any vendor has with any university employee must be logged and disclosed.

• If a university wants to conduct a long-term test of a product, anything beyond a one-day vendor demo, it must buy the product it wants to test. (Some schools do that. It discourages foot-in-the door and bid spec rigging tactics while lowering marketing costs sales costs that are passed on to other buyers.)

• For campus-wide installations at least two lecture capture brands—if the price tags are reasonably close—must be procured to ensure ongoing competition, to encourage innovation, to promote product cross-compatibility and to avoid vendor lock-in.

Reinventing Lecture Capture Give AV a Chance

Present day lecture capture originated in the IT world and its progression has been dominated

by IT think. There is not even a hint, at any point along its evolutionary trail, of any significant input from visual communications and AV experts. That may be why AV components have to be tacked on, after the fact, with all major hardware and software lecture capture offerings.

Another is the apparent absence of any good American studies into the effectiveness of the dominant multi-pane viewer interface design, a design which is clearly an IT creation.

It seems German researchers come closest to asking the right questions, the main one being whether showing the presenter and his or her rich media in separate frames a good way to communicate. “Many lecture recording and presentation systems transmit slides or chalkboard content along with a small video of the instructor,” points out one German study, stating the obvious. “As a result, two areas of the screen are competing for the viewer's attention, causing the widely known split-attention effect. Face and body gestures, such as pointing, do not appear in the context of the slides or the board.” (Anthropocentric Video Segmentation for Lecture Webcasts, Gerald Friedland, a German who teaches at Berkley and Raul Rojas, Institut für Informatik, Freie Universität Berlin.)

The Berlin research built on a University of Osnabrück study, “To see or not to see: layout constraints, the split attention problem and their implications for the design of web lecture interfaces,” Government, Healthcare, & Higher Education (E-Learn '06), pp. 2937–2943, Honolulu, Hawaii, USA, October 2006.

Some American studies tap dance around the split attention question problem without thoroughly researching it. For example, work of Dey, Burn, & Gerdes, 2009 Published online: 22 January 2009 Springer Science+Business Media, LLC 2009 and Moreno, R. & Mayer, R. E. (2000). Meaningful design for meaningful learning: Applying cognitive theory to multimedia explanations. ED-MEDIA 2000 Proceedings (pp. 747-752 which considered issues important to lecture capture:

24

“The split-attention principle states that learning is impaired when the presentation requires the learner to mentally integrate disparate sources of information before the instructional material can be rendered intelligible. The spatial contiguity principle states that it is better to present words and graphics integrated rather than physically separated. The temporal contiguity principle states that students learn better when corresponding words and pictures are presented simultaneously rather than successively.” I wonder how widely known the split-attention effect is among campus lecture capture decision

makers and among American makers of lecture capture products. How many users of lecture capture systems know about it? Do any of them care?

I knew about the split attention defect because my background is AV and communication not IT and funneling 1s and 0s. When I taught broadcast news at Brigham Young University I stressed the importance of presenters attaining then maintaining a close connection between what they say and what they show, a concept lost on most if not all captured lectures I watch. I taught how to control and direct attention not split it. Later, when I taught similar courses at the University of Dortmund in Germany, I cited research conducted by Dr. Hans-Bernd Brosius, in particular his paper “Text-Bild-Korrespondenz und Informationsvermittlung durch Fernsehnachrichten, ” Rundfunk und Fernsehen, 42, 171-183.)

Brosius studied the consequences of failing to connect narration with video in television newscasts. I met him at the University of Munich a few years later while consulting for a German

television station in Bavaria and encouraged him to update and repeat the study. Sadly he didn’t. Some of his research methods he used for broadcasting should be applied to lecture webcasting. It’s the type of research Sonic Foundry and others should have funded had they been really interested in advancing the state of the art.

Television news reporters and anchors were broadcasting visual content long before web-based, distance learning interfaces existed. Over the decades many broadcasters have been perfecting ways to display rich media on television screens and how to connect the video and audio

(usually scripted narration) for maximum comprehension and retention. At the same time millions of viewers have become accustomed to the way news is typically broadcast. (Imagine watching a television weathercast with the weather map in one frame and the weathercaster in another.)

Yet eLearning and videoconferencing developers decided to go their own way. Instead of merely adapting simplified television news production methods, they started with a blank sheet of paper. The multi-window/monitor format became the standard, a format that is as ill suited for the presenter in front of the camera and for viewers in front of their displays. When webcasters decided to reinvent the wheel, their version came out square rather than round.

Once developers decided to stick stuff fin separate frames, perhaps going back to PARC days, everyone else followed suit, the blind leading the blind, apparently figuring that’s the way it ought to be done.

Over on the broadcast side the technology used to display visuals undergoes continual change. Since the fifties there has been an evolution of products broadcasters use for simultaneously displaying presenters and data. Take weathercasting, for example. In the 50s and early 60s weather presenters simply wrote temperatures and sketched fronts on grease board maps. It was a simple, effective, real-space approach.

With the advent of computerized weather graphics the television industry turned to virtual space solutions such as chroma key compositing (aka blue/green screen effect) and digital video effects

25

(DVE) to electronically insert weather data into a frame with the presenter so the two elements appeared to be together in real space.

While choma key is still widely used it has given way in part to rear screen projection and then, as size went up and price down, to plasma and LCD displays. Projection screens and flat panels enable the presenter to actually see what he or she is talking about so they better connect with their rich

media. When an anchor (or teacher) turns and points to something on an LCD display the image it is really there.

Over-the-shoulder flat panel displays may enable be more effective webcasting than broadcasting. News anchors are generally more experienced and better trained to deal with the difficulties of using key screens. Plus their narration is usually fully scripted and further aided by teleprompters.

Classroom instructors on the other hand, generally interact far more with their visual aids. are also a good fit for webcasting. The

monitors are affordable and, unlike chroma key and DVE, their content can be seen by the presenter, affording a much more intuitive and natural presentation.

Chroma keying is sometimes used for webcasting. One clever product, Serious Magic’s Visual Communicator, enables the production of television news-like webcasts using green backdrops.

My patented system and method for webcasting rich media (patent No. 7,733,367) takes a page out of the television news book and adapts the concept to the classroom. Rather than capturing rich media from a laptop, for example, that is input (via a VGA connection) to the lecture capture appliance the laptop is a display. The images on the display along with the presenter are captured by a camera. The camera, in turn, is connected to the recording/streaming appliance.

This system and method transmits the presentation over the internet with one synchronized signal as opposed to the two unsynched signals transmitted by conventional methods.

The idea sprang from my experience as a television news reporter, university instructor, consultant for German television and legal consultant. My design process began with storyboarding

(sketching) what the remote viewer would see on his or her display. Then I worked backwards to devise a system that would achieve that look.

It’s the way the California television set company Express Group and I worked together to design news sets for German television. We sketched the first design for a set in Hamburg on the back of a Burger King napkin in downtown Hamburg. We started by storyboarding what viewer’s would see and then creating a studio set and establishing camera positions accordingly.

The first application of my patented system and method was in the legal arena. The DepoDisplay variant captured, transmitted and displayed the attorney, deponent and exhibits (rich media) in a single frame. For the first time ever depositions could be conducted paperlessly, eliminating the time-wasting paper shuffle.

26

I designed and built the first system for the Utah law firm Prince Yeates then later another for the firm Siegfried & Jensen (S&J). The latter was set up in a replica courtroom that I also designed.

The S&J DepoDisplay system was built around a custom conference table that had built-in monitors, a swing-out laptop tray, an auxiliary fill light, and a mount for a 42” LCD screen. The AV setup consisted of two PTZ cameras, two Winnov streaming recorders, a Marantz digital audio recorder with true confidence monitoring for backup and a Polycom videoconferencing codec.

I also designed a portable version of the room system so paperless depositions could be recorded on the road. That design, called Pack & Go, evolved into my V-Chassis product, the first all-in-one lecture

capture system on the market. It’s a completely self-contained with a recording/streaming encoder, audio mixer, camera, LCD display and ancillary devices. It was designed to be robust and inexpensive. I compare it to the VW Beetle in its day.

V-Chassis can work with my patented lecture capture system or with conventional rich media capture appliances with DVI/VGA input such as Winnov, VBrick, nCast, 323Link, Qumu, and Epiphan’s upcoming Matterhorn appliances. (V-Chassis could also be configured to run non-VGA appliances like Viewcast’s Niagara and software-only solutions like Tegrity and Panopto.)

No Microsoft appliance is on that list although some run Microsoft operating systems and some stream with Microsoft’s Windows Media Encoder. Despite years of fiddling around with lecture capture concepts, Microsoft remains on the sidelines. What happened?

When I finished my Vietnam tour of duty in June, 1970, I was sent to Ft. Lewis Washington for processing out of the military. At that time there was no sprawling Microsoft campus in Redmond, an hour’s drive from Ft. Lewis. There wasn’t even a Microsoft. Bill Gates was 15, attending Lakeside private school in Seattle. Lakeside bought a computer terminal that was connected to a mainframe a DEC PDP-10 owned by General Electric. Gates and his eventual partner, Paul Allen, would type in commands, wait for the distant computer to process them and then the results would spew out on punched paper.

Gates would go on to become the richest man in the world. Over the years his employees were continually working on ways his products could augment education. Hundreds of thousands if not millions of his operating systems are used by schools around the world for administrative purposes and teaching.

But Microsoft’s quests to bring lecture capture systems to market failed. Repeatedly. It abandoned efforts to bring its complex, one-button system to market.

Last year Joshua Kim, a blogger on the insidehighered web site, urged Microsoft to simply buy itself into the lecture capture game. Kim, a director of learning and technology at Dartmouth, tried to goad Microsoft into weighing in: “Microsoft is in danger of becoming irrelevant at the front-lines of learning and teaching.” “Our (lecture capture) community would benefit if we could help Microsoft

27

invest some of the $30 billion it has in cash in educational technology companies and projects,” he wrote.

Kim seems frustrated with lecture capture products that are too expensive and/or too proprietary. He welcomed the consortium of universities that is developing free, open-source lecture capture software under the Matterhorn name. “Matterhorn will push the consolidation of the lecture capture market, one that is currently fragmented and confusing,” he wrote. “The threat of Matterhorn in the very near term to the proprietary vendors may be small, but if Matterhorn gains traction the existence of a viable open source alternative will force a roll-up in this industry.”

But Kim also figured Microsoft could change things overnight by buying out a proprietary product and making it open source, even super cheap. “Buy into the lecture capture market,” Kim told Microsoft. “You have plenty of options. Tegrity, Echo360, Sonic Foundry, to name a few. “

Microsoft, Kim wrote, could use its cloud infrastructure to sell related applications that could subsidize the cost of lecture capture hardware and software.”

“Offer institutions the option of low-cost lecture capture in exchange for publicly

sharing campus related learning captures. Monetize all this by building up a robust educational channel on www.bing.com. There is an opportunity to leap frog past iTunesU and YouTube/EDU with an integrated lecture capture and publishing platform. Advertise against the lecture content, generated by your low-cost but high service lecture capture platforms. Use successful platforms like the XBox Live service as another route to distribute and monetize lectures.” Kim’s suggestion apparently went into one Microsoft ear and out the other. One reason might

be that a buyout might have led to Microsoft dominating the lecture capture market and attracting a new wave of antitrust lawsuits like Caldera v. Microsoft. He’s right, though; lecture capture does need some sort of big shakeup. A game changer.

In the meantime Microsoft has quietly reentered the lecture capture space by suggesting learning institutions adapt existing Microsoft products. Microsoft’s October 2010 developer’s conference introduced “The Future of Lecture Capture.” Cameron Evans, CTO for Microsoft Education in the United States, conceded, “Microsoft has not produced any specific lecture capture products for the market.” But he said the technology used for the conference “is simply a Microsoft Silverlight application built with Microsoft Expression and Visual Studio 2010.” “The bottom-line, he said, “is that anyone can build a similar application for lecture capture fairly easily.” (Note: The standard version of Expression is supposed to eventually replace Windows Medias Encoder.)

I sent Evans an email seeking a more detailed explanation and wondering how Microsoft’s new lecture capture initiative tied into its past efforts.

“Don’t look for any comparisons between any work we did decades ago, lots has changed since then,” he replied. “While the lecture capture capabilities are available in Expression and IIS (Internet Information Services) there is no out-of-the-box solution for this, at least not yet.” He suggested that “PowerPoint 2010 has the ability to record lecture speaking notes into a PowerPoint movie... if you a need simpler solution that is out-of-the-box.”

Videoconferencing

Videoconferencing with the exception of certain types of training is a poor distance

teaching tool. It is what its name says, a way to conference via video. Otherwise it might be called videoteaching. Video streaming, with a few exceptions, is a superior way to distant teach.

It’s simple. With streaming online students can access video on their PCs, iPads, etc., wherever they want and, if on-demand, whenever they want. With videoconferencing distant

28

students have to consume (waste) time and fuel to drive to remote VC sites that require special equipment (a codec). They have to watch when and where they’re told.

Lecture capture is less expensive. A special box is only needed at one end, not both. And even though streaming is not “face to face” and far site learners are not mic’ed, students can still interact live or later via texting, email and discussion boards.

(Because videoconferencing is a good meeting tool I designed a codec into the Siegfried & Jensen system. I personally own two VC codecs. Videoconferencing is also an option for my V-Chassis system. But I don’t recommend educational customers.)

By definition lecture capture devices support the transmission of rich media to distance learners. Otherwise it would just be video of a talking head. Rich media transmission is an option with videoconferencing codecs. Polycom, calls its dual screen system Presenter+Content and Tandberg, the second largest VC maker, calls its rich media delivery system DuoVideo. The major videoconferencing brands display lectures at the far site the same way as most lecture capture products: In separate frames, one for the talking head and another for the visual aids. Instead of displaying on web browser interface VC uses either a split screen on a single display or two displays.

Most universities use videoconferencing for meetings and perhaps some lectures. But most of them probably build their elearning around lecture capture, not videoconferencing. Utah, on the other hand, built its statewide distance learning program on VC. The Utah Education Network (UEN) connects

public universities and K-12 schools in a network that supports some lecture capture, some online collaboration but mostly videoconferencing. It has a vast array of 600 VC end points as part of what it calls its Interactive Video Conferencing Network (IVC).

In 2006, as I was fine tuning my lecture capture system and method, I

approached UEN’s steering committee about a chance to demonstrate my system and discuss lecture capture’s superiority over their videoconferencing paradigm. In an email to committee members I wrote that “part of the UEN solution involves the use of dual monitor videoconferencing systems and document cameras, for example.” “I believe there are better ways to lower the cost of delivering on-line courses while making it much easier for instructors to produce the courses and for learners to comprehend and retain course content.”

In April, 2006 I gave the UEN’s Technical Services Subcommittee committee a PowerPoint presentation, opining that UEN was going down the wrong path with its emphasis on videoconferencing.

Panel member George Miller later emailed saying UEN had no need to further explore lecture capture. “Utah has spent millions of dollars and over 20 years of experience in developing EDNET, Utah Education Network, The Utah Electronic High School, the Electronic Community College, and the UEN Satellite System,” he said. “The unique teaching strategies employed by our distance learning teachers combined with unique hardware is combined into an effective distance learning network that is second to none in the world.”

29

That’s quite a divergence of opinion. He thinks his VC-centric system is the best in the world. I think it’s a white elephant.

Since Miller’s email Utah continues to pour hundreds of thousands down its videoconferencing rat hole, throwing good money after bad, while perhaps only 10%—who knows the figure?— of the student population is regularly served with the technology that also wastes time and oil.

Four years after my PowerPoint, instead of starting over, UEN is tacking lecture-capture-style internet-based tools onto its IVC concept. I guess a little lipstick on a pig can’t hurt.

George Miller, who sang the praises of his face-to-face IVC network four years ago, today sings a little different tune. In a podcast (not via videoconference) he tells Utah teachers that they need to add what he calls web-based mediated technologies onto IVC classrooms. That geek-speak seems to describe a method of using social media where interaction between teacher and student (and among students) does not take place face-to- face but over the internet, perhaps not even live. Yes, the same concept that works far better with lecture capture than videoconferencing but the UEN wants to pin that

tail on their donkey. Miller apparently came to believe that face-to-face, the main benefit of UN’s IVC, may not be so

important after all. He says students who are reluctant to participate in a face-to-face environment may more readily take part in web-based learning. “Students who typically contribute little or not at all in a face-to-face classroom, “ he said, “have been shown to contribute significantly more in a threaded discussion.” And they can interact in web environments such as text messaging, electronic discussion boards, etc.

Why didn’t that occur to him five years ago, or earlier? Or a few million dollars ago? The UEN bureaucracy does have to defend the time and money wasted on its VC concept by

adding lecture capture or at least social media to VC rather than replacing VC with lecture capture. UEN should, instead, admit its mistake, cut its losses, dump VC as a primary lecture solution, keep a few codecs for meetings and embrace the newer, simpler, less expensive, greener, more democratic technology that puts lectures on iPads instead of on VC displays that cost students time and gasoline.

In the meantime Utah colleges and universities that are UEN members are left leaderless in their attempt to adopt lecture capture. At Weber State, for example, it was an individual department without the participation of UEN or the university administration that purchased Echo360 in 2008. Even apart from its participation with UEN videoconference Weber also has some VC it deploys on its own for distance learning.

Aside from whether videoconferencing is a good teaching tool, the method would also benefit from my patented system and method of delivering rich media for its meeting function. The major videoconferencing products, like their lecture capture counterparts, also deliver video of the presenter and—as an option—images of his or her visual aids in separate streams, and display them in separate frames.

The disconnect between presenter and visual aids compels viewers’ eyes to dart back and forth between the talking head window and the rich media window. Which screen are you supposed to look at? And when? The presenters seem trapped in the frame, looking for all the world like mug shots with moving lips, with no way to direct the viewer’s gaze.

Conventional VC: dual streams, dual displays.

DepoDisplay/UniDisplay: single stream, single display.

30

Polycom, Tandberg and others license the technology for their dual stream rich media transmission from Avistar Communications Corporation in San Mateo, California. Avistar has a portfolio of video and rich media collaboration technology patents that it has held over the heads of the main videoconferencing makers until they cried uncle and agreed to pay royalties.

Turn back the clock again to PARC in the early 70s. PARC scientists perfected if not invented most of the methods to display graphics and video on the same screen using what they called “tiles” but Microsoft now calls windows and others call frames or panes. At the time it was working on the tiles concept one of PARC’S best known and accomplished scientists was Robert Metcalfe.

Metcalfe invented the Ethernet in 1973, later co-founded 3Com (modems, routers and switches; now owned by HP) and among other posts became a board member at Avistar, holder of those rich media patents. (Metcalfe was also on the board of Forethought when it was sold to Microsoft and became PowerPoint. Metcalfe was friends with Bill Gates until he knocked Gates’ anticompetitive practices. Maybe that’s why Metcalfe now uses Apple’s Keynote rather than PowerPoint. He evangelizes online video and more recently turned his focus to promoting more and cleaner energy and joined the University of Texas at Austin as a professor of innovation.)

Metcalfe’s first-ever local area network was used at PARC to connect several of the world’s first practical PC—the Alto— to each other as well as to the world’s first laser printer.

I emailed Metcalfe wondering why VC makers like Tandberg would pay his company, Avistar, royalties for what I thought is a flawed multi-pane concept. I wrote, "I told (Tandberg's attorney) that ‘it's pretty much as if they are fighting over a patent for vacuum tube radios.’ They're bickering over a bad way to present content.’” I also attached a copy of a lengthy white paper I wrote about problems using multiple windows and the advantages of my single frame UniDisplay/DepoDisplay concept.

Metcalfe e-mailed back: “Your DepoDisplay seems like a very good idea." He thought it might attract investor support.

We’ll see. V-Chassis: A Garage Company

A garage company, according to Forbes magazine, is a “low-budget startup” but “in truth, most

began in kitchens, living rooms, or the back booth at Denny's.” A lot of great products were home brewed whether in the garage or kitchen. The point is they were not whipped up by existing, large companies that care more about preservation than innovation.

Perhaps the world’s two most famous garage companies are HP and Apple, both in Silicon Valley. I already mentioned the one located at David Packard’s house where he brought in an old Stanford classmate, Bill Hewlett, and during their free hours they would fiddle in that garage in search of new product ideas. Such as automatic bowling-pin setters. But it was their audio oscillator that received an unexpected order from Walt Disney engineers who

were working on the movie Fantasia that launched HP. The other famous Silicon Valley garage belonged to the parents

of Steve Jobs. There, in Cupertino, Jobs and his friend Steve Wozniak built the Apple I computer, releasing it in July, 1976. It sold for $666 and customers still had to add a power supply, keyboard, TV monitor and, if they wanted, build their own case like the one in the photo. But it was the Apple II, released a year later, that ran the world’s first killer app,

31

the VisiCalc spreadsheet, which made their computers more popular than anything offered by established, large companies.

The Apple steamroller kept gaining traction with the PARC-inspired Macintosh, the first successful commercial product with a graphical user interface (GUI). The Macintosh design was finalized after Apple engineers were given guided tours and in-depth briefings by PARC engineers. The Macintosh success, in turn, forced Microsoft to adopt GUIs for its Windows operating systems.

Michael Hiltzik whose book chronicled the history of Xerox’s PARC, analyzed why that company, that once had a monopoly on copiers failed to commercialize most of its innovative ideas. “Xerox executives did not define their business as making copiers, but rather as making money,” Hiltzik wrote.

Sonic Foundry, like many publicly traded companies, is driven by pressure to turn a profit. Had Sonic Foundry’s priority been the video capturing, recording and streaming live events—among them lectures—it would have developed and brought to market some of the innovations its competitors now have. It could have invested more in research, not only just to improve its flagship product but also to develop better, cheaper ways for presenters to webcast with rich media.

The development of my Depo/UniDisplay product began in my study. I started designing it for the legal space with the goal of helping my lawyer clients make a record of depositions in a way that would cost less yet be more effective than traditional shorthand or even simple video recording. I was aware of competing products and methods one of which was Mediasite. The advantage Mediasite had over most of competing audio and video-only courtroom record-making product is that it also captured the exhibits.

My first impression back then: Mediasite worked sort of okay in the courtroom. But it was way too expensive. And the way Mediasite displayed the exhibits/rich media and displayed on the screen was clunky, Plus courtroom exhibits unlike classroom PowerPoints had to be loaded manually. (At least the Mediasite could capture the exhibits and had optional streaming. Competing courtroom appliances like FTR, CourtSmart and JAVS couldn’t record exhibits and couldn’t stream.)

The existing, deficient products led to the creation of my now-patented DepoDisplay method for capturing legal proceedings. (UniDisplay is the name of the same system and method I use for the lecture capture space.)

The first DepoDisplay system at law firm Prince, Yeates & Geldzahle captured video of the lawyer, deponent and exhibits onto a Sanyo casino/security-style digital recorder. Proceedings were streamed live via a laptop with Windows Media Encoder.

Many of the law firm’s depositions took place away from the office at opposing counsel’s conference rooms. So I put together a portable version I called “Pack & Go”, a system that could fit in a carryon case. But with each use the operator to unpack, set up and hook up a bunch of individual components: camera, hard drive recorder, monitor, backup audio recorder, audio mixer, microphones, projector and rear projection screen and laptop.

Both room and portable systems functioned quite well. They produced depositions that were far more effective at communicating witness testimony at a cost much lower than plain stenography. The technology could help attorneys cut the duration of depositions and trials even as it improved depositions’ communication effectiveness with judges and juries. Plus clients could be billed less to cover depo costs.

But it did not take long for me to figure out there is little demand in the legal industry for cheaper, faster and better. Most attorneys bill by the hour. Why would they want to spend money on

32

technology that shortens the duration of depositions and trials—thus reduces their billable hours? It seemed most attorneys are more interested in making money than delivering better justice and providing more value for their clients. After the Prince Yeates installation only two systems were sold to other law firms.

At that point I parted ways with a partnership that had been formed to market DepoDisplay in the legal arena. I turned my attention to marketing the same systems under the UniDisplay moniker in the education sphere. My former partners joined a new company, iSys Global Solutions that developed its own lecture/legal capture system built around “multiple computer processors for capturing, processing, storing and uploading as many as 7 video and 28 audio feeds.”

The good news in the education world: Unlike the legal arena dozens of universities are very interested in lecture capture. The bad news: Competing products were already flooding the space and defining how they thought lecture capture should work. It was practically a given, for example, that the best way to display the present and rich media was in separate frames.

None of the products back then came packaged with necessary AV equipment. All that had to be sourced and hooked up by the customer. For schools with small AV and IT staffs I figured there ought to be an all-in-one product that combined all key AV and IT components in a single box. Thus emerged my V-Chassis concept. (V stands for video and chassis is a “framework to which the components of a radio, television, or other electronic equipment are attached.”

The very first mockup consisted of taped-and-glued-together pieces of cardboard. After that I built several generations of prototypes in my garage, without the backing of any outside companies or venture capitalists. If this was going to be a low cost solution it could not be handicapped with a lot of debts while the concept still on the drawing board.

Part of the development involved selecting and testing components that would be inexpensive, reliable and harmonious. A special wiring harness and power supply were engineered to help reduce clutter.

The first prototype was equipped with a Winnov-brand encoder. It was the same brand I used for the second law firm room installation. Winnov provided good tech support. It worked quite well but was too large, too heavy, a bit too hard to use and too expensive. Plus Winnov was just introducing a lecture capture model that included VGA input for rich media, deflating their interest in my alternative concept.

The next prototype was designed to accept the smaller, lighter, a bit less expensive, and easier to use Viewcast Niagara 2100 streamer/recorder. But Viewcast’s marketing people were not much

interested in the lecture capture space. I needed a similar device that would be even lower priced. I didn’t care if it had fewer features. My goal was to retail a completed V-Chassis for under $10,000 and the encoder is the single most expensive component.

That’s when I heard about I-O Corporation here in Utah, a company that had a contract to build encoders for the Mormon Church. The church’s purchasing agents wanted a device that could webcast services from chapels scattered throughout the world. But even the price tag of the reasonably-priced Viewcast that I had used for a prototype had scared them off. They needed something way less expensive. And so did I.

I-O Corp took the church’s challenge and developed a no-frills device that retailed for under $1200, less than a third the price of Viewcast’s cheapest offering. No extras also meant no hard drive and a low-end mother board.

I told I-O management that’s what I needed except a V-Chassis encoder must have a hard drive and a tougher mother board. Thus emerged the I-O ProEdge encoder built especially for lecture capture.

33

It is much cheaper, lighter and easier to use than even the Viewcast. It can create either .wmv or Flash video files. And it has one-button upload to YouTube.

Flexibility and consumer choice are hallmarks of the V-Chassis concept. If a customer has no interest in a V-Chassis itself he can buy a ProEdge encoder alone (retail $1,300) and configure the AV part on his own. And a customer who wants a V-Chassis but with a Winnov, or Niagara, or V-Brick, or Matterhorn, et al, can order that. A customer who prefers my UniDisplay system and method over multi-frame can buy the requisite license ($500) and equipment (a plasma or LCD display and stand).

Summary

Lecture capture needs a shakeup. Otherwise it will stay off course, in a rut. Several stars would

need to line up to democratize lecture capture:

• University presidents need to pull their heads out of their asses, take charge, and institute top-down lecture capture implementation beginning with specifications that promote competition, open source or at least cross-platform compatibility.

• Encoder box vendors/makers should demand uniform, fair specifications for university purchasing departments. They should also work together to insure cross-product compatibility among themselves and between themselves video management systems.

• The Department of Justice and state attorneys general should investigate and prosecute illegal bid splitting and bid specification rigging schemes which may turn out to be fairly common.

• Congress and state legislatures should help fund lecture capture pilot programs for universities and K-12 schools.

• Sonic Foundry, the so-called “uncontested market leader” which has been more of the problem than the solution, simply needs its ass kicked.