Review: Video Demos of Software have been used for 10 years; now they are poised to takeoff
Feature: Bandwidth, new software, and the need to train quicker + better drive the market Software is in constant flux. I am fond of saying to my professional friends that if they think their field is subject to constant change and retraining, come on over and see what IT developers and their users have been subject to in the past 40 years - hundreds of major languages, databases and tools plus major process and paradigm shifts are the "disruptive" norm. Just take a look at what demoware looked like just two years ago.

Camtasia Studio, one sample of a video demo creation tool

First and foremost, demoware tools are able to do screen captures of window or browser software in action. Once those captures are made, the demoware tools then allow refining those screen demos, with highlights, captions, and in some case quizzes and redirected training. In addition, the software allows users to bring in videos or Flash animations developed from external sources to enhance the training. For example, a demo course used videos of the trainer; then at the end of each step doing a summary of what was learned and then a preview of what the next steps in the demo would convey.

This demo course was based on a live training session with an expert. It had the advantage of being distributed over the LAN, on the Web and also by DVD - so that users could tap into the course material as conveniently as possible. But it also had embedded into it a final quiz that the user submitted securely back to head office to get credit for taking and "passing" the course. It is these types of capabilities that mark video demoware. But video or animated demoware have been a part of the training scene for the past 10 years at least - what has changed to make video demoware and their creation tools more viable?

In this overview I examine three trends that are changing the viability and attractiveness of video demoware. First, there is storage and particularly bandwidth. Storage has become so cheap that 100 to 200GB disk drives are the norm. So a video demo file at .5 to 2GB which five years ago was imposing, is no longer so. However, for distribution of training material that size was still limiting until the recent bandwidth revolution. Now online access (as opposed to sneaker mail distribution by CD or DVD) to video demoware presents much less of a problem. And we see this in vendors as diverse as Apple, BEA, and Xara distributing sales pitches or help demos in video or streaming file format. So bandwidth enablement and the consequent new ways of delivering training demos/help is factor one.

The second factor is that the demoware itself has evolved capabilities and become cheaper and capable of delivering in more modes. This is what this review will focus on. Specifically we shall see how in two years the video capture, editing and publishing capabilities of video demoware creation software has vastly improved. Finally, the third factor is that the need for training, more effective and quicker training, is not decreasing but rather increasing. The result is that organizations are looking for more effective ways to deploy their training dollars - and video demoware will be in that mix. But first, lets look at some of the critical enabling factors for video demoware in a bit more detail.

The Bandwidth Explosion

The Economist in its April 28th to May4th 2007 issue has a special report, A World of Connections, devoted to the topic of the bandwidth and related wireless revolution about to unfold. Here are the wireless basics:

As the table clearly shows there is a wide range of devices, prices, and bandwidth coming available over the next 6-12 months just in the wireless world that is going to bring tremendous bandwidth not just to PCs but to just about any device - PDA, cellphone, kiosk, you-name-it. This effectively does away with the last 1/4 mile barrier to wiring up households and now people on the go. So for example, libraries in Ontario are becoming free Wifi islands for people with laptops or PDAs who want to connect to the Net at 3-6Mb/sec while using many other information resources available in the library. And of course municipalities are starting to think in terms of WiFi, WiMax or other wireless as a basic utility infrastructure like water and electricity as in the case of Philadelphia and San Francisco.

So ingredient one, large bandwidth is becoming available to not just PC laptop users, but also PDAs, mobile phones, tablets, and any other "wired" up device. This is important because demoware which captures and plays back live screen events, demands large amounts of bandwidth. Simple screen captures that use animation rates of 5-9 frames per second requires for VGA screens 1000 x 768 pixels/screen * 7 frames/second = 5.3Mb/sec capacity for uncompressed images. For highest fidelity demoware using video frame rates between 24 to 32 frames second the VGA requirements bump up to 20-25Mb/sec. This is where the software enablers come in.

Demoware , Compression and Codecs

Even with the enormous bandwidth that is coming available, 5 to 25 Mb/sec/connection can quickly swamp even the heftiest transmission system. And that still does not take into account the complexity of melding together all the different components of a screen presentation - audio, voiceover narration, vector drawings or captions throughout, extra charts and objects inserted to make a point, and finally control sequences to pause, quiz or otherwise control the demo/presentation.

So the demoware software is essential for 3 reasons and the corresponding steps. First, the demoware most often provides the capturing capability. Demoware "films" your screen for you. This is the raw, uncompressed animation clip or video. In the edit step, demoware takes the clip or video to the next stage by allowing users to enhance it with highlights, captions, and quizzes and other training finishes as required. The result is an edited training film (or sales demo) which has to be compressed and adapted for delivery on disk, DVD, FTP server, iPod among the most popular formats.

The last step is production. This production step compresses the demo into presentable format - usually Flash or some video format. This is where specialized codecs in demoware takes the screenshot movies, compress and format them appropriate to the medium for distribution onto a CD/DVD, live streamed demo, web Flash file, or later rebroadcast/redistributed "movie" downloads. So in output publishing there are two purposes. First, compress the demo down to a size so that it can be efficient ly distributed. Second, put the demo into a format and then deliver it to either disk, CD, DVD, mobile device or other appropriate platform. And as we shall see, many of the software programs vary in what platforms they are prepared to deliver to.

As anyone who has worked with JPEG or GIF images knows, there is a trade-off associated with codec compression: size versus quality - the less compressed the better the quality; the more compressed - there is loss of quality in various ways - typically noise and other compression induced artifacts. This is is known as lossy compression - because the original video or clip has permanently lost some of its original detail. However, there are less powerful compression methods that have the desirable trait of preserving all of the original - these are the lossless compression methods. The advantage of lossless compression is that if any re-editing or or reuse of the clip or video has to be done, the original detail is available.We shall see this theme and its trade-off of preserved quality versus degree of compression for efficient storage or transmission as a recurrent issue in demoware.

What in fact, the codecs do in terms of compression is quite remarkable. In lossless compression, the original image can be fully and accurately restored to its original state/condition after being decompressed. The trade-off in lossless compression is speed of compression/decompression versus the amount of compression possible. Typically for animations that is 9 frames per second for VGA quality (1024 x 768). Lossless compression can deliver a 3 times compression ratio at these rates - thus the codec is able to keep up with the incoming 9 fps animation; anything faster and it must buffer or deliver on a non-realtime basis.

So lossless compression is effective for animations at 9 fps or less and reduces the transmission requirements by a factor of 3 - so assuming Web colors (256 hues = 1 byte per pixel). Animation is going to require 1024 * 768 *8 bits for color = 6.3MB/sec before codec lossless compression / 3 = 2.1MB/ sec after lossless compression. This is a substantial improvement but still leaves a lot of video data to be broadcast and stored. Enter lossy compression.

Just Good Enough

Lossy compression is based on the idea that if you allow for losses which cannot be seen by most viewers one can up the compression ratios quite dramatically: 10 times with audio, 20 to 300-1 for video, and 6 times for still images - all with barely perceptible loss of quality. However if those media are repeatedly edited and saved then accrued generation loss can occur and will be distinctly noticeable. But the whole trick in video demoware is to make just one compression with the finished product using high-powered lossy methods. The result is a copy that is barely different from the original but reduced in size by a factor of 10 or much more. This lossy compression step can take 2 minutes or much longer to do depending on the size of the video screen, still images used , the demo's length and its complexity. Multiple hour renderings are not unheard of.

Obviously in some circumstances such as video conferencing such rendering times are impossible and so either screen size or image quality has to sacrificed. But in the case of most other screencasts or demoware which are targeted for LAN, website or DVD use, "Just Good Enough" rendering that take an hour or more to produce are still worth the while. The quality can be maintained at a fairly high level while the download time can be reduced by factors of 30 to 300 or more.

Lets see what that means in terms of our benchmark VGA screen going at mid range video rates - 28 fps:
1024 * 768 pixels * 1 byte for color * 28 frames per second = 22MB/sec. original size
22MBsec * 60 sec /150 compression factor = 8.8MB/minute compressed.
However, let me tell you that for codecs from Techsmith, ON2 and some Flash videos I am seeing compressions even better than that - they consistently deliver 2 to 2.5MB/minute of good video quality - which is about 4 times better than our average. However, because our eyes do a lot of interpolation and "filling in the blanks" the individual still images may be noticeably degraded in quality. See the Moscow State University video bench marks here for all the details on video codec compression measurements.

We are still not out of the technical woods yet - in video demoware there is another important technical factor related to compression and ease of use. This is the capture method used.

Capture Method

we have been spending so much time considering codecs and compression because they determine how big and what quality of demoware we can make. But there is another factor that impinges on that decision. Notice we have been including animations at a comparatively slow 2-9 frames per second in our discussions. There is a reason for that. One of the neat facts is that using inter frame optimizations based on comparatively slow animations, one can further compress video and media presentations such that our 2.5MB/minute video compression can be reduced by a factor of about 6 plus or minus 2. This reduces our VGA quality demos down to a cost of 300KB/minute - very attractive to Web-based sales or training demos. Hence the appeal of animations driven demoware from produced with event captures.

However, there is a major quality/fidelity versus size trade-off that occurs right during the screen capture or recording that is very important distinction among demoware programs. There are two ways to do screen captures 1)event driven or 2)fixed frame rate(usually 24fps or greater) . Many demo programs use event driven screen captures. In event driven screen capture, the program waits to take a snapshot based on events happening on the screen - typically mouse movements or click, keyboard clicks, or change in the state of a component. Actually even when a window (Linux, Mac, Windows or whatever GUI) is still - no mouse movements or keyboard operations taking place - there is a simmering of GUI events taking place such as callbacks, screen refreshes, alerts, components changing state, etc. This is the key problem with event capture. Normally these events are not captured - and sometimes in recording games or collaborative PC sessions that is a problem.

In contrast, fixed frame rate video captures at 24 frames per second does capture these externally driven events which are becoming more common with Web 2.0 AJAX, Flash, and Java applications. But to taking those 24 frames per second has the drawback of significantly larger demo files in the case of fixed rate video captures. So now the last major piece in the puzzle of which video demo software to get is in place. And as always the best choice depends on what you want to use the demoware for and what type of screen events are being captured. Here is our demoware rule of thumb:

1)If the size of the demoware must absolutely be minimized then consider event capture demo software. But realize extra work may be required to capture external events as most demo software is unable to pick up those callbacks, alerts and/or changes of state;
2)if the ability to capture all events on screen with high timing fidelity but less stringent quality of image requirements, then consider using fixed rate video capture software that does lossy compressions;
3)If the timing fidelity and quality of the clip images must be top notch; consider using fixed video capture with lossless compression;
4)If realtime events such as video conferencing are required use animation or reduced rate video capture software with fast compression methods.- also considering decreasing the capture screen size and/or the video image quality;
5)Finally if speed of creation of the demo is important and there are few or no external events(typical of many PC non-game and non-web programs) and quality in terms of time and graphic fidelity can be just good enough - then we are back to event capture demo software.

The key insight here is that event driven screen capture programs monitor the stream of GUI events and snap a picture of the screen for certain events, say a mouse click, mouse movement, or keyboard press. Since the GUI event, (a mouse click) provides a wealth of information on what and where the event happened - some screen capture programs are also able to automatically add onto the captured screen a text caption saying "Click on button here" or add color highlights to the text fields that should be filled in. It is these auto hints (which can be easily embellished or deleted in the demoware's editor later) and the precise capture of events that is the attraction of event driven screen captures.

But there is a downside. Sometimes the event-driven screen captures miss important events such as a game move, or a callback auto update of a screen common in Web 2.0 apps. Other times the event-driven capture can look herky-jerky, like animated films, because they miss/ignore GUI events of interest. As we shall see this is a major trade-off between event-driven versus fixed frame rate video captures. The attraction of event-driven capture is this ability to auto-label the mouse and keyboard events while reducing the screen captures considerably - thus delivering a much more compact demo. However, fixed frame rate videos (at 24 or more frames per second) miss no GUI events including menu or option box pulldowns, drag and drop operations, and Web based collaboration or alerter popups and displays.

Fixed frame rate captures just take a movie of the screen at a fairly high rate - say 20-32 frames per second. This has the advantage of not missing events occurring on the window and providing a sense of fidelity in recording screen, menu, mouse and other GUI events that is very close to natural. The movements and screen changes are close to WYSIWYG-What You Saw Is What You Get. Of course, the disadvantage is the huge size of the video capture files. Even with compression, a 1 minute clip can be several megabytes in size (if not dozens of MBs) depending on the size and complexity of the screen image. The following is a sample of an Event Capture.
For example, the Hearts card game captured with Adobe's Captivate 2 program at 540x480 in screen size resulted in a 0.5MB/minute demo file (see screenshot above). But an event capture of a Corel Painter screen at 1000 x 800 resulted in a 4MB/minute demo/file. Finally that same Corel Painter screen when captured using video frame rates with several different demoware programs, all of the captures averaged slightly less than 20MB/minute in demo file format. If doing screen captures can result in files of such large size why do it ? Well we have already seen that documentors have a wide array of technical compression strategies available. But the other compelling factor

Training Demand

Well first blame it on You Tube and the TV generation. 50-60 years of television have had their effect - reading is down, watching TV, videos, and movies is up. In fact, studies show that learning material presented in animated or video format is more effective in speed and retention of the material over many other learning modes including reading, audio-only, and various work exercises. In addition, video has been found to be as effective as live lectures in various training situations. . Finally there is the incorporation of demoware in the many learning and knowledge management systems. Now couple these observations with the fact that Training is big business - according to Training Magazine well over $50B in the US alone and one can immediately see why video demoware as a key part of eLearning would be the source of rapid development and improvements over the past five years.

In the IT arena, training is particularly important as IT software follows a variant of Moore's Law - every 3-6 years the software you use will have changed such that at least half of it will be unrecognizable in comparison to the original software. Now add to this mix a flux in the software used brought on by mergers, acquisitions, and changes in internal processes - and one has a major learning problem. Couple this with a decline in enrollment in computer science of 40% for the past five years and you have demand for ever more IT training coupled with a decline of those with the the skills to do the training. And finally, training is not easily out-sourced. So you have the basics of demand for better training.

But there are more factors. IT Manuals for many categories of software are rarely read except for getting started and getting out of an immediate problem. At the same time support from software vendors varies widely in quality and costs. So organizations are looking for more effective ways to do their IT and broader training. Hence the interest demoware - that can capture and convey well the essentials of when and how to use the various programs on hand.

Finally, demoware has some important convenience factors. On a CD or DVD it can be run at work or home - and rerun, paused, and digested just the way the user wants. Customized training. Yet with some of the demoware we review here, control is not lost. Built into the software can be postings and quizzes that verify the user is going through and learning the concepts properly. Then, like game software, as a user verifies through their quiz answers they are ready to go to the next level, the program then presents them with a series of new learning options. In addition, these results can be collected and sent securely back to the organization so that training progress can be monitored and evaluated. So some training developers will be looking for capabilities beyond capture and demo refinement. So lets take a look at what is available in the market place.

Demoware Goes Video

We have noted on our graphics website that video software like Adobe Premiere Elements or Ulead Video Studio have gotten so good that they are being used to generate video demoware. There is a problem. None of the popular or even high powered commercial video editors can do PC screen captures. And as we have seen the screen capture is a critical factor in demoware creation. Fortunately, there are a number of of good Demoware programs to meet these requirements. The table below list some of the key tools:

Demoware is a rich but primarily Windows centric market. Mac and Linux users will have to use video capture programs like Mac's iShowU, SnapzPro, or Screenography and Linux vncrec, vnc2swf, or Xvidcap and then import the video files into one of the Windows demoware programs which is readily done. Publishing and output is also a challenge because you may have to take the same basic demo show and produce 2 or 3 different copies depending on target usage: Web, PC, CD/DVD player, etc. So many readers will want to proceed to the second part of this review which looks at the capabilities of demoware in much more detail for their capture, edit, and publish capabilities. Meanwhile, the 4 top tools are indicated with a brief statement on why they were picked. But again, see the detailed reviews.

In the editing domain there is really quite a spread of features but the basics are a slide/image/clip viewer and/or a timeline view where users can add captions, notes, hotspots, images, and more clips or video. The range of capabilities here is quite broad and each program has some special or exclusive features. Finally, the publish options are another area where the tools have quite a mixed range of offerings - yet as we have seen this can be most important. The net result is that Adobe Captivate, even at $599, is the most versatile and comprehensive tool. Captivate is packed with downstream development features; and can be highly recommended in most demoware situations. Nonetheless, users will have to choose carefully among demoware packages for the one that most closely meets their needs.