Wednesday, October 24, 2007

Why You Should Use Video

HCI researchers need to use videos more frequently to illustrate prototypes and concepts. An important part of doing research in HCI is communicating ideas to others in a way that they can use your ideas or your findings to further their own work. Writing and publishing journal articles is the conventional way of delivering this knowledge to a wider audience; however, in recent times, a conference paper has become an accepted means to convey knowledge in HCI--primarily because the research field moves so quickly. Yet a well-produced video can deliver the same message to an audience quicker, with more punch, in a more memorable way, and with less investment on the part of the audience researcher. It also acts as an archival mechanism, allowing "demos" of prototype systems to be shown long after the prototype has ceased to function. A video can perform all of these functions with ease--all it requires is the effort to create the video itself.

In this article, I cover the benefits of a good video for the audience and for the researcher. I use this to motivate a way of thinking about your video, and then provide directions on how to create and deliver a video for impact. Finally, I point to several examples of polished and unpolished videos -- all of which have had an impact on how I think about my own research.

Videos are good for the audience
  • It is a low-investment vehicle with potentially high payoff. An audience member is committed to viewing the video for only as long as the video lasts: thus, if a typical video length is 2min-5min long, then s/he only needs to invest 5 mins watching your video. This time commitment is orders of magnitude shorter than the amount of time s/he would need to commit to reading a paper about the same ideas. A well-produced video conveys the essence of the same ideas as a paper would (the paper expands on those ideas)--yet ultimately, what an audience remembers is the key points--the same points that should be delivered in the video. In sum, it is low-effort investment that pays off ultimately with the same ideas the audience would get if they read your paper.
  • It is easy to understand. Videos are usually easy to understand because the voice-over conveys the information in a conversational tone. When writing papers, it is easy to lose this conversational tone and get muddled in the details. Because videos tend to be brief, they generally get straight to the point -- thus, it is easier to grasp the essence of the idea, whether it is the problem space you are addressing, or the solution space you are presenting. When giving a talk or writing a paper, you ultimately write abstractly to cover a broad problem/solution space; yet with a video, you are forced to be concrete, be it to capture a specific instance of the problem, or a specific instance of the solution. This concreteness allows people to more easily see and identify with the problem/solution.
  • It can be viewed at any time. Videos have the advantage of being played at any time. They do not require the original author to provide a voice for the idea (as in a presentation). Thus, they can provide a complementary archival function of a paper about the same idea.
  • It can be replayed. In a talk, the audience usually cannot ask a speaker to pause, rewind, and repeat what was just said -- as a consequence, if you lose an audience at one point during your talk, you may have lost them for the rest of your talk. In contrast, a video can be rewound and replayed at the audience's convenience: if they don't understand what you said, or want to see the demo again, they can do so. As an audience, or "learner of the idea", this is a very powerful mechanism.
Videos are good for the researcher
  • It forces you to focus on what's important. Because your video is a one-sided conversation, it is important to convey the idea in a meaningful way. Like a good elevator pitch, creating a video forces you to be specific about what is important in the message or idea you are trying to convey. The video helps by making concrete the details that you leave out of your narration -- in so doing, it actually "says words" on your behalf! This means that your narration ends up focusing on the important ideas that you are trying to convey.
  • It forces you to be concise. Storyboarding your video forces you to be concise about what you want to say. Nothing is worse than a boring video -- it's less painful than a boring movie, but painful nonetheless. By making a video, it forces you to convey a small set of important points, therefore forcing you to be concise about what these points are, and how to communicate these points. Making a video is good for you, the researcher.
  • It provides you with an archive (a.k.a. you only need to make it work once). Staging a demo for visitors tends to be extremely stressful: demos frequently suffer from "demoitis" -- things that worked moments ago fail when you need to show someone right now. Sometimes, the only person that knows how to make the demo work isn't around, or even worse, code rots -- the longer since you last ran the demo, the less likely it will run right now. By producing a video, you only need to make the code work once. After you stop running code every day, futzing around with it, fixing bugs--always, all bets are off: there is a chance it may never work again. The video is like proof that it worked. Once.
How to think of a video
  • Understand there are different kinds of videos. Different kinds of videos need different kinds of treatment. Sometimes, the audience is a close familiar -- someone who understands the problem space or solution space, and you're just trying to convey a sketch of the idea. Other times, the audience is someone who is doing HCI, but isn't focused on the particular problem you are working on. Finally, sometimes videos are intended to be shown to funding sponsors, who barely even understand HCI. Like giving a lecture, you need to position your video differently depending on the audience. I suggest erring on the side of believing your audience is ignorant when making a video -- it forces you to be more explicit about vague concepts, and perhaps more importantly, could allow your video to be pitched to several audiences. (It has been my experience that once you make /one/ video, you are unlikely to go back and make another.)
  • Your video is an extended elevator pitch. In an elevator pitch, you get 30s to convey your idea. With a video, you have people's attention for 2-4mins. That's actually a lot of time, but you still need to use it wisely. You want to make sure that there's content to see or listen to, so don't waste them: the video acts as your proxy -- upon watching this video, someone will judge whether you are interesting or not. Don't waste that.
  • Your video is a prop to convey the idea... so convey the idea. Say it at the beginning, say it in the middle, and say it at the end. Don't forget to say the idea. Sometimes, you just use the video as a prop, where you actually watch the video with someone else, and narrate in person. This is a good idea too, but remember that a good video can stand on its own, so make sure the video still conveys the idea if you are not around to babysit it.
How to get started
  • Start shooting. There's no better way to get started making a video than to start shooting. You don't need anything more than a digital camera in its "video" function. Heck, you can just take photos and string them together--this technique is called "stop-motion" animation (my favourite stop-motion video). You'll learn what angles make best use of the video frame, and which ones can't even see the computer screen or interface.
  • Familiarize yourself with a simple tool. Upload your photos and movies to a simple tool, and start playing with that tool. I use Windows Movie Maker -- it's not powerful, but it is very easy to play with and to get things going quickly. You'll learn which features of that tool will help you to tell a story, and which are superfluous.
  • Make videos. Finally, make the videos! Put in the cuts, use the transitions, and everything under the sun. Save them as movies! You need to go through the entire process to tell understand how to tell a story with a video.
  • Show your videos to others, and watch them yourself. Finally, send your videos to others, and get their feedback on those videos. Get them to tell you what worked, and what didn't work. Ask them whether they understood the video, so you know how you need to change the story telling that you do. Don't forget to watch your videos yourself -- it is sometimes quite surprising what you learn from watching your own video.

Tips: in the order that I learned my lessons
  • Front load the information -- get to the point. This is important. If your video is over 5min long (even 4min long), it is probably too long.
  • Take the time to get the narration right. Make sure it's concise, to the point, and doesn't waste a listener's attention span. Nothing's worse than taking the time to make a video that no one watches.
  • Take the time to get the shots right. Position the camera, the people, the computers, and whatever props you are using to get the best shots that you can. Make each frame tell part of the story -- don't waste it, and take the time to get the shot right. If you start editing later and realize the shot is wrong, it takes a lot of effort to set up the system and lighting conditions again to get continuity. There's nothing worse than to tell an actor to bring the same clothes they were wearing that day so you can reshoot a shot.
  • Get multiple versions of the same shot -- from different angles, too. Take multiple versions of the same shot -- this will help you cut and crop for continuity. Get different angles on the same shot, too (just redo the scene with the camera in a different spot). Get 3x the amount of video you think you need for your video -- this will help you pick and choose later.
  • Use cuts and meaningful scene changes. Use the pauses and scene changes to move from topic to topic -- just as you would use a slide change in a real talk.
  • Use subtitles to underline a point. Subtitles are a powerful tool -- used sparingly, they can really convey the important, key concepts of a video.
  • Use background music. Use subtle music and turn the volume down -- background music can cover up some of the pops of a bad clip or narration.
  • Storyboard it if you have time. If you've got the time, do up a storyboard. But story-boarding is boring! Just grab a camera and start shooting!
Good videos: polished and unpolished
  • Nerf Email Notifier: Low production value, but an all-time favourite -- conveys everything you need to know in the first 10s... "Oh, I have mail."
  • Casual interaction in a hallway: Low production value, but is a great clipped video that shows you exactly what happens in the hallway
  • AmbientROOM: High production value, and demos a room that has a lot of cool features. Notice how they use cuts and zooms to focus your attention on certain aspects of the scene.
  • Tagged objects at Xerox PARC: High production value, and demos lot of interesting ideas.


Blogger Christopher Collins said...

If you wrote that in a minute then I know for sure that you are not human.

October 25, 2007 at 7:52 a.m.  
Blogger tangaroo said...

I am actually a robot. Hahaha.

Anyway, here's a link to a screengrab tool that can be used to make quick videos for passing to email or IM buddies:

October 28, 2007 at 11:35 p.m.  

Post a Comment

<< Home