Wednesday, August 04, 2010

Bandwidth and Information

I decided recently that I had "stopped growing musically." You have to be a Catholic flower child from the 60s to inflict that kind of guilt and deprecation on yourself over what is meant to be a hobby--something you do for fun.

(not from the 60s, but you get the picture)

So I hit upon a plan. I found this great dobro player, Martin Gross, who has a terrific YouTube channel. So my plan is to learn one song of his every month. OK, all happy again now that I am "working" at having fun.

As I've been working on "Blues Stay Away from Me" I've made a couple of observations.

The old way of learning a song was to play it on the record player and keep hacking away at it until you figured out how the person was playing it. Essentially, not much has changed except that with YouTube you get the video channel as well as the audio. And honestly, that makes it easier, but not in proportion to the orders of magnitude increase in information that the video makes available. Anyone who's worked on televisions is well aware of the difference in bandwidth between the video signal and the audio. There's just a lot more information in the video signal, and Mother Nature is an exacting accountant (the cost for transmitting information is bandwidth--the more information you are hauling, the wider the highway has to be).

Seriously, if you had to choose between learning a song by just listening to it without seeing the video, or watching it without hearing the audio, you'd be much better off just listening.

Kind of ironic seeing that the video has a ton more information in it. So what gives?

Well, most of the visual information is irrelevant. The color of the guitar, the spacing between the strings, the freckles on Martin's hand, etc. The most important information is what fret he is putting the slide on and what strings he is plucking. Since this is not a split-screen video, those two pieces of information are at opposite ends of the display and it takes a bit of replay sometimes to figure out what he's doing.

It might be true that a picture is worth a thousand words, but apparently a K of sound is worth a Meg of video.

About this time you're checking the header of this blog, thinking it was supposed to be about user experience and user assistance stuff. Well, it made me think about screen cam versus written procedures. Does the same thing apply here?

I think it does. I've felt for a long time that if the only thing we have to say is click this and type that, then a video is not the way to go (and a LOT of software videos are of that variety). Lot of bandwidth for just a little information. I wonder if Tufte's concept of chart junk and data to ink ratio can be applied to useful info/bandwidth analysis. Things like tone and physical manipulation in motion seem to justify the kind of bandwidth that video carries. I don't think of this is as a transmission efficiency issue, no more than Tufte was trying to save ink costs. The human bandwidth and ability to focus is more at issue here.

Plus, it's easier to scan a written procedure to get to that snippet of information I need than it is with a video.

So the point is twofold:
  • Written words are still an incredibly efficient channel for conveying information. Quit beating yourself (or others) up if you consider yourself a writer and that to be your primary channel. "I am technical writer, hear me roar."
  • If you can afford to throw a video or two into the user assistance, do something worthy with that bandwidth.


Techquestioner said...

I, too, think video is overkill for many types of information transfer. At concerts, or during a televised music performances, I sometimes close my eyes to concentrate on listening to the music.

When someone is learning a new process, they want the step-by-step instructions and a way to confirm that they are in the right place as they proceed at their own pace. For this, static pictures or diagrams interspersed with the steps may be all that they need. Having to correlate the instructions with a video may be too much of a focus shift for some processes (or for some people).

For your musical example, maybe all you need is the audio, the tabulature for the chord fingerings, and maybe a picking diagram.

Bruce H. Johnson said...

Video might be excellent for steps involving motion, such as a graphic drawing program. I could show node editing and moving.

Also, could depend on the target audience. A simply-produced video on how to outgrabe the momraths with hand positions (and proper posture) would possibly be better and faster than the most eloquent prose for a foreign-language audience.

Michael Hughes said...

Back to the music, I'm still pondering if video without tab is better than tab without video. Of course best world is audio, video, with tab :-)

Ken Hilburn said...

This is a very intriguing concept you have proposed: that we can have "video junk" that actually prevents us from gleaning what we really need to learn from an instructional video.

I suppose the part that's the most difficult to account for is the capacity of the viewer's brain to filter out the so-called noise and collect the pertinent information. In effect, we have, to some extent, trained our brains to ignore things such as freckles or spurious mouse movements. My guess is that the user's ability to auto-filter that "junk" is somewhat directly proportional to their experience level within the context of the video. This is probably the same with regards to Tufte's chart-junk concept. Experienced musicians can learn by watching a "single view" video and experienced application users can learn by watching a mouse move around on the screen. However, beginners probably won't make it because they can't filter the junk.

Thanks for posting!

Atmospheric said...

Mike, I agree with many of your points. However, I wonder how much this point of view might be somewhat generational in nature. One of my struggles as a technical writer is how to entice folks to actually read what I have written BEFORE they get themselves into trouble.

While you and I (being about the same age) might be comfortable with cracking a manual or searching help for keywords, I think a decidedly younger generation is going to opt for the video, despite its inherent waste of bandwidth.

So then, which method is REALLY the more efficient way of communicating: efficient compact prose that few people read or a video that might be more widely viewed?

I don't have an answer in mind. I'm just asking the question. And for the record, I've not yet started embedding video, but I'm sure thinking about it.

Ted Kuster said...

Thanks to Youtube, I fully expect to have this banjo thing down in time for the First Annual STC Summit Bluegrass Jam, next year at Sacramento.

Joe said...

The big problem with video for me is that it often forces you into it's timeline.

If I want to learn how to use Swype on my Android phone, I am forced to go to the website and watch 10 videos each around 3 minutes long.

In all honesty, all of this information could have been written in 2 paragraphs and an 8 item, bulleted list, easily read in under 3 minutes.

I think someone could make a killing if they would come up with an Evelyn Wood Speed Video Watching course.