Creators: In the world of the internet, 60 seconds is a long time. On YouTube alone there are hundreds of hours of video uploaded every minute. A new "experiment in algorithmic filmmaking" taps into these vast swaths of video content, and turns them into an ever-changing music video for Manchester band Shaking Chains' debut single "Midnight Oil."
6 comments:
The ever growing overlap between technology and art never ceases to amaze me. I can’t help but wonder how much manpower went into coding the website. I understand nothing about computers, but it would be nice to know whether it was a one man job as a sort of hobby or a job that programmers were contracted for. Regardless of how it was made I didn't anticipate it to be as captivating as it was. In the end I wasn't even listening to the music, just staring and the few seconds of each shot. While I was watching it I found myself conceptualising projects I could create. It was an interesting experience to have that many ideas pop up over and over again all while in the course of a short video. Maybe the program has the potential to be a fast paced pinterest in which you type the keywords you want and receive tons of information in only a few minutes for inspiration.
I decided after reading this article that I would have to watch this music video at least three times, to get a wide variety of pulled videos and to really test the algorithm. After reading the premise of the film, I was confident that it wouldn’t work. How could you possibly trust your art to mindless math and logic? But Ladies and gentlemen, I have been proven wrong. The video works every time and it actually means something. The artist juxtaposes disturbing images of war and sexuality and death with images of the earth’s and humanity’s beauty on the calmer chorus. It creates a very disturbing effect. I have now watched the music video nine times (though every time I’ve seen it there is always the same footage of someone putting pepperoni on a raw pizza towards the end of the video so I’m just going to keep watching it until that footage no longer appears because I’m curious). This is just incredibly amazing and a fantastic/weird idea that actually worked.
When I saw this article, I was expecting to be really impressed by this video. I wasn’t. I think the concept behind it is interesting, but I watched the video before I read about the vision, and at least for me, that message didn’t come across.
Now, I would like to see the code, but as far as I can tell, it really isn’t that complicated. I’d have to go through and time the “video switches” to test this theory, but I’m assuming that they have the video change at a consistent rate (not constant, just consistently aligned with the song) and then pull a random video (or a video with some particular sublist of keywords if they want a makeup tutorial in one place and a video of an animal in another place). However, as far as I could tell, the videos themselves didn’t fit with the song, just the timing of the video switches, and even that wasn’t super clean. I suppose it would be marginally impressive if it was able to generate a music video for any song by analyzing the music and finding where to put the video switches (and it is possible that this is analyzing the music and switching the videos at slightly different intervals each time, which does make it more interesting from a technical perspective), and it would be really cool if it could pick videos that seemed to go with the song. A program that could generate a seemingly natural music video for any song would be super cool.
But this isn’t even remotely that. It generates what to me felt like a “best of Youtube” montage (side note: I don’t think pulling clips from Youtube gives the feelings of snapshots of moments in time because certain videos in it (make-up tutorials, Rhett and Link, funny cat videos), feel distinctly “Youtube”, not like moments in life) that at least to me, didn’t actually seem to be a music video that went with the song. The idea is good, but I don’t feel like they actually executed it to its potential.
Huh… the thought of algorithmic filmmaking is definitely something I have never thought about before. If I am going to be completely honest with you I wasn’t terribly thrilled with what I saw, but the concept and the ideas are definitely there. I would be very interested to see the brains behind the project and see how it is pulling the clips and making the new videos. It really is fascinating to watch the creation of a new movie every time hit play. I even tried to trick the program to see how fool proof it really is. After seeing this article, I began to wonder about other possible applications and ways that this could be used in the film industry and also for other projects. Maybe down the road, we will see this in a software-based system that many people could use and have the benefit of using and creating with.
This idea is really interesting. I’ve always been curious about the interaction of patterns in art. I’ve always wondered if it’s possible to create an algorithm for art. I mean, some would argue that if you can create an algorithm for it then it’s not art, and there’s a point there. But I guess, in terms of audience experience, it’s about noting if they can tell the difference between art and something created by a computer. But none the less, this is a very cool concept. I especially love the idea that since this video is ever changing, the art sort of lives on even if everyone else has forgotten. You could come back in a hundred years and this music video would be completely different. I should also note that in general, the music video does feel like a music video. You have to be looking for it, but there’s a sort of progression from beginning to end.
Thinking back to, say, the 80's, when music video culture was way more involved, this contemporary algorithmic music video idea seems like it could be adapted to being significance back to music videos. People used to watch music videos every night on MTV and in the 2000s, when youtube blew up, watched OK Go and Beyonce millions and millions of times. But now a days no one I've encountered cares much about music videos, we don't listen to a song on the radio and think about the video. I can see this idea, which draws on contemporary "media overload" or a "spastic advertising"style, resonating with youth culture. It's newness and individuality are really appealing, but I don't think it's bringing back music video culture either. I think the unrelated videos could distract from the music and disassociate from the band. However, maybe it's a step in popularizing this forgotten form of entertainment.
Post a Comment