The following scenarios illustrate situations where video-embedded documents might be handy and you can review the options in a lot more detail in this extended whitepaper.

• Large company IT departments are pushing employees to store their business documents in secure document management systems like SharePoint or Lotus Notes. These systems enforce enhanced security and offer document control safeguards that are difficult to replicate on standard web servers. Documents with embedded video coexist very nicely with these systems.
• For security reasons, the same IT departments may resist the proliferation of intranet or extranet web server applications. Even if a document could be hosted on a web server, policy restrictions or resource costs may make that difficult for people in business units and non-IT departments. Those individuals can author video-enhanced PDF or DOC files without requiring any special IT support.
• Although this problem is diminishing, companies with highly mobile workforces may still have people operating in places where web connectivity is marginal or nonexistent. Field service crews, overseas offices in developing economies and personnel co-located behind strict client firewalls may not be able to reach key company websites. They can carry a library of video-enabled documents on a hard disk, thumb drive, or CD.
• Most employees in large organizations use Microsoft Word and it would be convenient if all forms of publishing could be done with that tool. However, to work with other programs, Word documents often must be exported to another format. If you add video into the mix, that gets very complicated. One iPOV client found that, by converting Word originals to PDFs and embedding video, the resulting documents became much more universal. A given PDF document with embedded video can be stored in SharePoint, loaded into the Learning Management System, linked and made viewable online at a website, or emailed to parties outside the company.  One file, many uses.
• A similar situation occurred in a global multinational. Divisions around the world each had their own Learning Management System, some of which were old and not standards-compliant. If they published eLearning in the customary SCORM or AICC packaging it would have taken a lot of effort to import it into all of the varied, regional servers. However, every LMS (even the old and non-standard ones) could import and display a PDF. By packaging the video inside the PDF, the courseware effectively became a universal commodity.

Examples

(click on the images and captions to view PDFs with embedded video at iPOV’s website)

You can view more examples at iPOV’s website: http://www.ipov.net/content/rich-media-pdf

Gotchas

A fast-moving technology always generates surprises that initially baffle newcomers. These are the surprising limitations and gotchas that dictate what is actually possible. Accordingly, we begin this guide with a list of the key constraints that we have learned to work around:

1. Can I put a video into a Word Document and have it viewable and playable on the Page?
   No. You can create that impression if you insert a Windows Media ‘object’ into your document, but the video won’t actually be inside the Word file. It will be an external file that the document is linked to, and you must manage this linkage yourself. You can put a movie inside the Word file, but it will appear as a small icon that you must click to open. It is theoretically possible that a skilled programmer can write a macro or plugin to make this possible, but iPOV has not seen any developer activity in this area.
2. Can I put a video into a Word document and then save it as a PDF?
   Not reliably or well. iPOV is not aware of any tool that will reliably export a video-embedded Word document to PDF and have the video play correctly. First, if the video is in a typical, common movie format (MP4, AVI, WMV, MOV, etc.), it should first be converted into a Flash movie – something PDF exporting software won’t do.[1] Second, if the movie embedded in the Word document is already a Flash movie, the resulting embedded movie will have very limited play control (see point #3).
3. Can I add a Flash video to a PDF using Acrobat Pro?
   You can add a FLV or H.264 video or SWF animation easily enough, but the Flash Player doesn’t supply any play controls (start, stop, pause, tracker bar, volume, etc.). This is a big drawback in business situations where the viewer may want to review a key point or adjust the sound volume. To get good play controls, it is necessary to package the video and play controls together into an SWF file (e.g., using an external editor such as Adobe Flash) and insert the unified SWF file into the PDF.[2]
4. Why haven’t I seen more PDFs with embedded video – it looks really cool?
   Acrobat has supported embedding videos since version 6. The problem for business documents comes down to workflow – see answer #2 above.  In general if you want to edit the text of a document you will do so in Word and then re-publish it to a PDF.  At this point someone must individually re-insert all the videos from the old PDF into the new PDF. iPOV has developed a set of processes and procedures to help us do this quickly and correctly and we are working on tools to automate the process even more.[3]^{,    [4]}
5. Why not put the videos on YouTube and link them to a document?
   YouTube (and most sites like it) caters to people who want wide public distribution for their files. It has a very basic security system and it may be often blocked by corporate firewalls. Sites like these also tend to display ads that may or may not fit with your corporate image.

[2] The latest version of Acrobat Pro has an alternative way to import linked files, but it is finicky and prone to error. It is also more difficult to debug any broken links once the files have been imported.

[3] When documents are designed for large numbers of end users, it is often easier link to a video on a web server rather than distribute a packaged document.

[4] Note also that recent security changes have made some old documents no longer playable.

• Getting the knowledge out of the expert’s brain
• Embedding the knowledge in the learner’s actions
• Everything in between

Worse, these are intimately related. When experts and learners are at a distance from one another, it is incredibly easy to miscommunicate. It is equally easy for the message to get misplaced on the way. To combat this, instructional designers have developed a meticulous methodology to guarantee that the knowledge leaves and arrives safely. However, as change drives education and training to respond more quickly, that attention to detail is under stress. Increasingly, it looks like we are facing a choice: fast,  good and cheap- pick two.

How can you deliver knowledge efficiently?

While it seems new to the eLearning community, this dilemma is not unique. Goods producing industries and service industries have faced this challenge for decades. They have also found ways to resolve the conflict and deliver quality, speed and value. So why not approach the problem the way that Toyota makes cars? Why not adapt the principles of “lean thinking” to the knowledge supply chain – end to end. That takes some new ways of thinking and thinking about some new things.

The Principles of Lean Systems

The concept of a ‘lean’ business processes is at least a century old. Many credit Henry Ford’s Rouge River factory in the 1920s as the first major business process that was seriously lean. A ship with iron ore unloaded at one end of the plant and steel smelted from its ore emerged in the Model A automobiles at the other end. Nothing was allowed to impede the smooth progression from raw material to finished vehicle. In the years since, Ford’s vision has been refined. Now, many operations experts consider the Toyota Production System to be the gold standard for lean. For my taste, Spear and Bowen’s definition of Toyota’s approach reduces Toyota’s success to four key principles:

1. All work shall be highly specified as to content, sequence, timing and outcome.
2. Every customer-supplier connection must be direct and there must be an unambiguous yes-or-no way to send requests and receive responses.
3. The pathway for every product and service must be simple and direct.
4. Any improvement must be made in accordance with the scientific method, under the guidance of a teacher, at the lowest possible level in the organization.

Source: Decoding the DNA of the Toyota Production System, Steven Spear and H. Kent Bowen, Harvard Business Review, Sept 1999, p98

Lean processes are not unique to automobile manufacturing, or even to manufacturing in general. They can occur in a wide range of businesses, from supply chains, to services, to the military, and even healthcare. The essential ingredient is the passion to apply and follow sound system design principles. There are several versions of the principles. In their 1996 book Lean Thinking, James P. Womack and Daniel T. Jones outlined five principles, which I would paraphrase as follows:

1. Determine the customer’s definition of value.
2. Identify all of the steps in the process and determine their contribution to customer value.
3. Remove any avoidable steps that don’t create value. Redesign the process to eliminate currently unavoidable steps that don’t add value. Arrange the residual value-creating steps into a transparent, integrated flow.
4. Let customers pull valued items from the system – replacing them only as needed.
5. Pursue perfection in the process through continuous improvement.

Adapted from Womack and Jones

A Lean Knowledge Supply Chain

To translate these concepts to the world of eLearning takes a bit of imagination. iPOV has been exploring this for nearly a decade. To this point, we have identified a short list of critical tasks that have to be done:

• Visualize the entire supply chain: Think about the entire journey that knowledge must take from the expert’s brain to the learner’s actions. Don’t just think about how is managed, or how it is delivered, or how it is assessed. Those are all important parts. But they are not the whole journey. In order to improve the process, you must understand how the different tasks flow together and how an upstream task can impact a task far downstream.
• Identify opportunities to create ‘standard work’: It seems counter-intuitive that anything involving knowledge could be subject to standardization. However, while the knowledge itself is highly varied, the package containing it doesn’t need to be. Consider a training video. It doesn’t matter what the camera saw. In a movie, it is all reduced to a sequence of image frames played at 30 frames per second. Once it is a movie, an Einstein lecture is processed the same as a Madonna concert. You can’t standardized the content of the video, but you can standardize the processes that handle and distribute it.
• Measure the flow time: How long does the whole journey take? How long does it take to pin down the SME? How long does it take to get them to write/review the knowledge? How long does it take to get approvals? How long does it sit before it is released? How long does it take to schedule learners to view it? How long does it take to measure the effect? How long does it take for the results of those measures to filter back to the start of the process? If you don’t understand where the time is going, there is no way to pull it back.
• Identify and isolate the waste: This could be a blog article by itself (and probably will). In our optimized, technical world, it is really hard to find any process that is as wasteful as the one that delivers industrial knowledge. Delays, mistakes, rework, overwork, lack of attention, distractions, and misunderstandings are legion. Name an eLearning development and delivery project that didn’t undergo at least 4 or 5 edit cycles. To remove this waste, you must first identify it, label it, and understand it. Only then will you be able to get mad about it and do something to eliminate it.

iPOV has proven again and again that, if you can address these issues, you can develop high quality eLearning materials in record time. We love talking about how we do it. We love talking with other folks that are trying to do it too.

“describes a problem which occurs over and over again in our environment, and
then describes the core of the solution to that problem, in such a way that you can
use this solution a million times over, without ever doing it the same way twice”

Alexander, C. et al. (1977) A pattern language: towns, buildings, construction, New York: Oxford University Press.

Alexander proposed that designers seldom create designs from scratch. Instead, they recognize that the similarities between their problem and previous problems and the similarities that should be applied in proposed solutions. Alexander’s idea has since been eagerly adopted by software designers.

eLearning Design Patterns

In 2005, the European E-Len Project produced a report on 34 possible applications of design patterns to eLearning. Unfortunately, the underlying project has been closed – possibly because many of their “patterns” turned out be fairly obvious and trivial to any experienced teacher: e.g.,

• Management of on-line
  questionnaires
• “Management of on-line questionnaires”
• “Moderation of asynchronous online groups”
• “Student interaction in groups”

An older, but arguably more interesting, model can be found at The Pedagogical Patterns Project. This links to an effort that seemed to fade in the mid 2000s that produced a set of far more interesting design patterns. Many of these were aimed at classroom teaching, but they illustrate the ability to define a problem and pose a generalizable and believable solution pattern. Consider the following example from a book submission that the authors provide on their website:

SHOT GUN SEMINAR *
This pattern was originated by Alan O’Callaghan

You wish to have a group of learners discuss the results of literature research of diverse topics.

❊❊❊

A typical technique to broaden a group of learners understanding of a field or topic is to have them research the topic and select one learner to report back to the group as a whole via a presentation. Often this results in a good presentation, but poor discussion, as only the presenter is fully motivated to do the research.

[details omitted]

❊❊❊

Therefore, identify a topic for research and encourage everyone in the group to research the topic. Choose the presenter by random ballot at the beginning of the scheduled presentation time. Restrict the timing for the initial delivery.

[details omitted]

❊❊❊

Shotgun Seminar recognizes that learners are typically motivated by the fact that they are presenting rather than by the intrinsic nature of the research topic. By selecting the presenter at the start of the seminar itself, every attendee is equally motivated to do the research, and is equipped to enter the discussion opened by the presenter. The overall benefit of a successful application of the pattern, especially on a regular basis, in a regular time slot, is strongly beneficial to the development of a research culture.

[details omitted]

This “design pattern” is practical and widely applicable, yet devoid of any assumptions about the specific subject matter of the class. It could be applied equally to a course in psychology, engineering or business. This gives it tremendous power and utility.

We have been following this concept for some time. On the face, it seems to promise tremendous time and cost savings. We even tried to write and apply a few design patterns of our own:

• HareAndTortoise – Pattern for situations where expert can do task quickly, but explanations are lengthy. The pattern calls for making two videos: a) one that is the expert at full speed to show how it is done, and b) one that is long and thorough so a learner can understand the implicit logic.
• StoryBehindTheStory – Pattern for situations where a new corporate policy may encounter resistance. The pattern calls for an authoritative figure to record a short (30-45 sec.) war story explaining the issue or threat that caused the policy to be adopted. This video is played as an introduction to motivate interest and compliance.

Despite our efforts, we consistently hit a major roadblock. It doesn’t help if we follow a design pattern and our clients don’t. Unless we can get our clients to recognize, understand and respect the pattern, they will do what clients usually do. They will toss us material and tell us to sort it out. Worse, they may try to apply their own conception of a design pattern and ‘help’ us by delivering raw materials in that form. We are always begging our customers not to help us.

Reluctantly, we have come to the conclusion that design patterns are a fantastic idea that will never catch on – at least until there is broader industry collaboration on the construction and maintenance of these patterns. What if the commercial training and eLearning industry were to embrace eLearning design patterns? What if a group of major vendors and eLearning consumers got together and built a library of these documents? Imagine a world where eLearning buyers and vendors had conversations like these:

• “We need a course that is mostly StoryBehindTheStory, but with a few JustTheFactsMa’am patterns”
• “I think we can deliver what you want using mainly the HareAndTortoise pattern – with a Flash animation Roadmap on the front for guidance.”

Oh well, one can dream.

If not, check out the ATEN CS228.

The ATEN CS228 is a KVM Matrix switch with the following properties:

• It can control up to 8 computers (or 64 computers if it is daisy-chained).
• Two consoles can access shared computers simultaneously.
• It has audio (speaker/microphone) support to/from both consoles.

In plain English, this means that two operators can share up to 8 computers at a time and follow several collaboration modes:

1. They can share a single computer – one operator has full control while the other is a passive observer.
2. They can switch computers in seconds – as long as they don’t try to use the same device (in which case, see scenario 1)
3. They can pass control of a given device back and forth between them with a few keystrokes.

Consider the scenario where you need to train a new hire to use a complex software package (e.g., CAD, business intelligence, movie editor, etc.).

The idea is simple. Pair a newbie with an experienced professional and give the newbie some introductory tasks. Even the simple tasks will have a few tricky parts. That’s when the newbie calls across the cubicle and asks “How do I do this?” The mentoring professional hits a couple of keys to take control and demonstrate – while the newbie gets to watch and ask questions. A quick demonstration and it’s back to work for both of them. As the newbie gains confidence, the mentor can share more complex tasks – first by letting the newbie watch, then by letting the newbie drive under the mentor’s watchful eye. Eventually, the questions will subside and both will be working full time on value-added tasks.

The following figure shows how this compares with standard classroom and OJT. The intensity of blue in the diagram indicates the amount of revenue work that is being done. Dark blue is all productive work. Light blue is semi-useful. White is purely overhead devoted to teaching and learning.

You’ve seen multimedia that you would love to use in corporate communication. Unfortunately, the corporate IT department is very security-conscious. They resist installing a new server or web application inside the firewall, and they forbid putting corporate information on outside hosts. They ‘encourage’ everyone use the company’s secure document management system (.e.g., Lotus Notes or Microsoft Sharepoint) for all business communication. Is video out of the question?

iPOV’s Rich Media PDF

iPOV has refined a way to embed sophisticated, animated Flash videos inside a standard Adobe PDF file. The result is an integrated, thoroughly standard PDF. If you give us a Word document and a video (in almost any format), we will return a polished ‘Rich Media PDF’ like the following examples. Note: In keeping with their PDF format, these should be downloaded to a local disk and viewed from there. Download times will vary according to file size:

Right Click on Thumbnail and Save PDF to Disk in Order to View	Example Application
	Software Instruction Memo (1.7MB, 1 video clip) – It’s pretty common that a minor software problem will get in the way of productivity. Wouldn’t it be nice to record a quick fix or procedure change and send a totally unambiguous memo to let everyone know what to do.
	Video work instruction (2.2MB, 2 video clips)- This shows the correct method to perform the sort of factory or distribution task that is used throughout industry. The work instruction design contains two video clips. One shows the task being performed at full, expert speed. The second shows it being performed step-by-step to make the subtleties more evident. Click here to view an in document version.
	A video prospectus for a real estate property (10MB, 1 video clip) – Most of the information in this flyer is given in the embedded video clip. Even though the video has no commentary, it still gives a more accurate impression of the property than any text or still picture could ever achieve.
	Instruction sheet for a computer maintenance task (1.9MB, 1 video clip)- This video-enhanced document combines a step-by-step written procedure with an illustrative video clip. The listing ensures that the viewer has a complete guide to the tasks. The video clip ensures that the viewer can put the list of instructions in their proper physical context.
	Procedure for using software (14MB, 5 video clips)- Video clips with screencast movies of a software application have been added to a detailed step-by-step procedure. The procedure provides the accurate details for the task, while the videos provide situational context and a sense of the flow of the transaction. A viewer can remember the transaction flow far more easily than they can memorize a list of steps.
	How to change the oil in your car (3.7MB, 1 video clip)- This is a simple procedure, but it illustrates how a synchronized script can be added to the video clip to make the speech clearer. The companion script could also be used to add extra description or even warnings that are synchronized with the action.

Who would need this?

These PDFs will play on any computer where Adobe Acrobat Reader and Adobe Flash Player are installed. Presumably, anyone who wants to view a PDF will have Acrobat Reader installed and, according to Adobe’s statistics, Flash Player is installed on 98% of computer desktops. Companies can handle situations where a multimedia web page is not feasible:

• IT security policies make it difficult to put the video on a web site.
• Bandwidth limitations make it difficult to stream video to the target audience.
• The target computer is not connected to a high-speed network.
• Users are carrying laptops into areas with low connectivity.
• The company’s official documents must reside in a document management system like Sharepoint or Lotus Notes.
• The company assembles collections of documents to distribute on CD or DVD to stakeholders beyond the firewall.
• The company deals extensively with overseas offices that have slow connections.
• etc.

In each of these situations, a Rich Media PDF would offer an attractive alternative. This format will fit in many of those awkward places where a multimedia website cannot reach.

Requirements, Time and Cost

• The cost depends on the amount of processing that iPOV must apply to the video. To prepare a video (without additional editing) and place it in a PDF with customer-supplied text (e.g., with the formatted contents of a Word document) typically costs $75/page, or less in larger quantities.
• If iPOV is asked to apply its standard video transformations to the video clips, the cost will generally be determined by the transformations. Since the ‘flattened Flash movie’ is an iPOV standard output, PDF page publication fees are typically waived.
• Simple page constructions are returned in 1 to 2 days.

Bottom Line

Rich Media PDFs are a useful compromise between static text and graphics documents on one hand and multimedia websites on the other. They offer rich multimedia, with a level of portability and flexibility that a website cannot match. Finally, they offer rich media in a package that is widely accepted by security-conscious IT departments. If you have text documents and videos that you would like to display together, consider giving us a call.

The value of multiple voices

For small or one-off projects, it is reasonable to try to use one really good voice per role or language. As project size increases or the project is likely to evolve over time, however, it is often better to source 2 or more voices per role and alternate them among the individual clips. This practice requires more setup and project management (you have to coordinate 2 or 3 times the number of voice talents), but it can pay big dividends down the road:

• If you expect to make changes and edits, there are risks to designing with a single-source voice talent. If you lose access to the talent and have to bring in a replacement, it can be very jarring to the listener to suddenly hit a new voice when most of the project is done by one person. If, however, the project uses multiple voices from the outset, another voice will cause little comment.
• If you only use one voice, there will always be some listeners that find that voice unsettling. It’s ultimately a matter of taste. If you alternate voices, however, listeners are more likely to view them as simply ‘different’. This makes it easier to source voice talents that are very good – but need not be great. That little bit of leniency in voice sourcing can substantially speed production and restrain costs.

Adjusting the pacing of the underlying animations

If you are putting one voice on an animation or movie, you would expect that you would have to align the speech with the action. However, that generally occurs fairly naturally as you build the animation and edit the movie.

When you try to take the same movie and apply a voiceover in a different language, however, the new speaker will likely have a different timing and cadence – and possibly overall duration. For example, it can take 20-25% longer to express a given thought in German than it does in English. German has a lot of multi-syllabic words and more complex grammar constructions.

When the audio structure changes significantly, it is usually necessary to adjust the timing in the underlying Flash movie or animation to allow a bit more time or to cut out dead air. iPOV now does this on all of its work as a matter of course, and we have developed proprietary tools to help us manage this process. The tools include a web-shareable script editor with built-in Flash viewer, teleprompter, and voiceover recorder utility. We have found that small changes to pacing and cadence can make a surprisingly big difference to perceived quality.

Dynamic switching of languages

iPOV has developed a unique Flash player control (www.cosolvent.biz) that allows for on-the-fly switching of media. This can be used to give users the ability to select video clips and languages dynamically at page load. You can view a demo of our capability at this link.

In theory, we can post one video clip to a site, supply any number of scripts and audio tracks and let server logic or JavaScript determine what the viewer sees and hears. In practice, we generally change the Flash movie clips as well as the script and audio. We use different video clips because we usually adjust their timing (see above) to align with their respective audio.