We have more video on our hands these days than we know what to do with. Search has moved past the wunderkind position and is now a confirmed deity in the online universe. However, one of those nagging things we’ve overlooked in this dynamic industry of ours is that video really isn’t a very well tagged form of content.
The tagging that happens on YouTube and other video searches is decent, but not robust in any way. Have you ever searched YouTube for a “tearful scene” and seen the results? Not what I would call specific to the keywords.
And if you try to drill down, it just gets worse. In fact, you’ll be the one having the tearful scene after a while.
There’s more to it than that. What about the idea of a deeper definition of a tearful scene? What about why there was a tearful scene? Yes, context is vital, but not evident in a non-verbal expression of an idea or a dramatic scene.
Online, words are the hooks that we hang technology on.
As odd as that may sound, we’re condemned to exist this way until we can define experiences at the level of abstraction rather than associative definitions. It may be easy to find an actor but can you find a scene that they played by describing the narrative of the scene?
This may seem somewhat beside the point to some people. If you want something, think of the word that it’s associated with, search it, and find the video clip. Easy.
That may be easy now, but as the content juggernaut grows, searching will continue to adjust to the volume. At some point it’s not just a matter of scale, it’s about context and narrative.
Searching isn’t good at dealing with abstract terms, like “beauty.” Yes, you can drill down and refine the search, but there are so many variables that it becomes pointless.
Searching for video content reminds me of the Rolling Stones lyric, “You can’t always get what you want, but if you try sometimes you just might find, you get what you need.” OK, Mick and Keith aside, I think I’ve made some sort of point.
How we get there is the real question. I’m sure there are some amazing minds that can figure it out. As I’ve said many times before, users may not have evolved enough to use this theoretical search solution yet. Then again, maybe they have.
We need to study the way people describe a thing they’re looking for, without using specific words, and try to build a search result that works.
The questions would have to be specific to video, or motion picture content. Like: “Describe the scene in ‘Blade Runner’ when the killer robot dies on the rooftop in the rain.” Then, ask: “What was the mood of that scene?”
That’s what Ask.com was known for, but for some strange reason it didn’t catch on. Maybe it’s time for a search algorithm reunion tour. Or search algorithm 2.0.
Searching streaming video content must become more accurate, right down to the scene level. That may seem crazy, but as we get a higher level of content length and complexity we’ll see the need increase every day.
One thing’s for sure: all hope is lost that I’ll go back into my VHS tapes to find that one part of that movie I taped and wanted to see again.
Election 2016 is already like no presidential race before it, and one of the most striking aspects of this year’s race is the disparity ... read more
Video consumption keeps increasing and Facebook is serious about a video-first world, encouraging us all to explore its full potential. Ian Crocombe, ... read more
Mike Andrews Ph.D is Chief Scientist (Forensiq) at Impact Radius, and is carrying out some fascinating work around digital marketing and ad ... read more