Blog

Deep Dream resources roundup

magiccitydd

I’ve been following the Deep Dream tool since it emerged a few weeks ago with the idea that this is much more interesting than a bunch of freaky psychedelic images. We might not be able to say exactly why it’s more interesting, but there’s something there, I’m sure of it.

So here’s a list of the stuff I’ve seen to date. If you know of anything else please let me know.

Explanations

Deep Dream is Blowing My Mind and more background info from Memo Akten is a great primer.

Google’s original explanation.

Why Google’s Neural Networks Look Like They’re on Acid – an attempt to explain just that.

As Art

Is it art? Specifically for Rich Oglesby’s answer.

Why Google’s Deep Dream Is Future Kitsch

Taking it further

Shardcore attempts to use it sparingly

Avoiding kitch. A concerted effort to get away from the puppyslug. Also has a good explanation.

Deep Dream Music Video

Code

Running deep dream on Windows and OSX. I was completely flumoxed until I found this.

Google’s deepdream code

How to make gifs and videos using DeepDream Animator

DeepDream Video

Online Deepdream services

Dreamdeeply.com

Focus

There are at least two broad meanings of the term “focus” in relation to photography. The common one is to bring the subject into focus, to make it not-blurry. Effectively it is not the subject that is out of focus but the focal plane which is either in the wrong place and needs to be moved closer to or away from the aperture, or is not deep enough and needs to be increased by reducing the size of the aperture.

The other meaning of “focus” is slightly more interesting. Here we’re talking about the subject of the photograph, that which we want to draw attention to. There could be many identifiable things in the image but we want the viewer to ignore them and focus on one in particular. We do this in at least two ways.

The first is to use compositional techniques, such as the Rule of Thirds and creative lighting, to leverage cognitive biases in the viewer’s mind telling them what is important and what is irrelevant. We build up the image, as a painter might, to create a story which leads the viewer through the image, causing them to pause as they understand the subject through the visual cues that lead them to it.

The second is through framing. Here we decide what to include and what to leave out. Photography is in many ways a subtractive art. We create involving photos by excluding many of the things that make up reality and then asking the viewer to put them back. What happened before and after is removed as we take a slice through time. And what is happening beyond the edges of the photo is excluded, to be ignored or imagined but never known.

The edges of the photo have become interesting to me. Being in the camera obscura you have a sense that you’re seeing everything, but then someone walks in to the frame and you’re surprised. Where did they come from? There isn’t supposed to be anything beyond the edges.

This blinkered view is embedded into photography. We zoom in and crop out to focus on the subject and the resulting image is better for it. This mimics how our brain processes the surprisingly wide visual field our eyes are constantly pouring into it. We are vaguely aware of things in the periphery but we mostly ignore them, focussing on a relatively small area in the middle. We mentally zoom in and crop, so it’s no surprise our cameras are made to do this too.

Wide-angle lenses do exist, but they look wrong. The barrelling of the fisheye is the most immediate problem, but they also push the background away, creating a distancing between anything in the foreground and its context. The GoPro, frequently mounted on the subject, actually uses this to its advantage, emphasising the thrillseeker and making their context seem vast and awe-inspiring. They know it didn’t look quite like that in reality, that it looks wrong, but we on our laptops are ignorant that a trade-off has been made.

To focus is to exclude, to create voids. And in doing so it is to invite speculation and imagination, for the brain abhors a vacuum and will fill it with ideas. And that is where photography gets its power from.

Systemic Photography

I had a couple of crystallising moments recently where something I read contextualised everything I’ve been thinking about over the last few years in an incredibly useful way. The first was Vilém Flusser’s book Towards a Philosophy of Photography which came up during a talk by James George at Resonate this year. He’d recommended it to his friend Alexander Porter who found it a revelation, completely changing his view of photography. I figured I needed to read it too.

It’s good, and made a lot of sense, but the crystals that emerged were still small. I knew there was something there but I couldn’t quite articulate it, or turn it into a working practice of my own. Then Alexander pointed to an article about Flusser by Kenneth Goldsmith (the keeper of UbuWeb and someone who’s been on my radar for a while) called It’s a Mistake to Mistake Content for Content which is so much better than the cleverclever semantic title suggests. Here’s a key bit:

Flusser claimed that the content of any given photograph is actually the camera that produced it. He continued with a series of nested apparatuses: The content of the camera is the programming that makes it function; the content of the programming is the photographic industry that produces it; and the content of the photographic industry is the military-industrial complex in which it is situated, and so forth. He viewed photography from a completely technical standpoint. In Flusser’s view, the traditional content of the cultural artifact is completely subsumed by the apparatuses — technical, political, social, and industrial — surrounding, and thereby defining, it.

Suddenly the crystals were huge, so big it was a little bewildering. I told Jenny about this and she seemed surprised it was such a big deal as surely I’d been thinking about this for years, but that was sort of the point. I had been, but I hadn’t fully realised it. This is what really interested me about “social media” before it became impossible to work in that area thanks to marketeers and other eejits. And this is everything than interests me about photography in the 21st century, about how context, and the lack therefore, is essential to understanding the power of an image, and how that context is deeply embedded in the systems that enable it.

(I should also add that I don’t agree with all of Goldsmith’s statements, especially those where he appears to impose a value judgement, but that’s fine. It’s his perception that has value here, not his conclusions.)

So this is the flag in the ground for the first stage of working with these ideas in a more deliberate way. Here is my plan.

I want to keep things simple. Diving in an unpicking the myriad systems that inform a photo on Instagram is insane, though I hope to get there eventually. As Goldsmith says, the best way to interrogate these systems is “to break the system by doing something with the camera that was never intended by industry.” To hack it, in other words. So I’ll be looking for where that happens, particularly emergent community hacks rather than those make a political or artistic point. Alongside this is finding glitches in the systems themselves and examining those. The little moments where the ideology of a system bumps into the technical or commercial reality and causes us to raise an eyebrow.

Systems, of course, are all around us, so I’ll just be looking at those directly related to the creation, distribution and consumption of photographic images and writing a bit about them on this blog.

And, finally, I’ll be ignoring the value or otherwise of the visual content in the images themselves. I’ve been collecting vernacular photographs from Instagram for a while now and it’s become a dead-end. The images are not the thing – it’s the context, and I threw that away. Idiot.

Mission stated. Onwards.

Notes on Sitting In Stagram

Siting In Stagram was produced in early February 2015. I reposted the same image over and over to Instagram to see what happened. I then bundled it up into a video which I also posted to Instagram:

1 -> 90

A video posted by I Am Still Sitting In Stagram (@sitting_in_stagram2) on

I submitted it to the art blog Prosthetic Knowledge for consideration and he featured it. A few days later it went “viral”, appearing on news sites and content farms all over the world. I was able to track 950,000 views based on the video some of the sites embedded though this could only be a fraction of the total. I have had lots of feedback and questions and as such have spent the last couple of months thinking a lot about this silly little artwork I made when bored with a head cold one February night. This post attempts to gather those thoughts so they might be used for future enquiry.

Attrition

Attrition is something that fascinates me. I see this as distinct from simply entropy, where something degrades on its own, as it implies the slow, steady action of an external force. Rocks do not become pebbles on their own – the attrition of the sea causes that transformation over centuries, one wave at a time. A step in an old church is work into a curve over centuries, one step at a time. Each step brushes off a layer of atoms to no visible effect, but in aggregate a significant change occurs.

This is evidence of activity, and it’s most interesting when it’s human activity. A seat worn by bottoms, a library book worn by readers, street cobbles worn by vehicles, vinyl records worn by playing. This attrition is all around us and is a way of measuring popularity. The worn seat probably has the best view. The worn library book is the most useful. Your favourite record is the one that sounds the worst. And so on.

Lossy copying

Before the digital media era copying was inherently lossy, in that each copy threw away a small amount of data. Photocopiers were the best example of this. Every office would have a faded, almost illegible photocopy of some cartoon or joke, the original of which was lost in time. See also tape-to-tape, VHS, chemical photography, screen printing – each copy, while perfectly acceptable, was slightly inferior to the original. Something, however small, had been lost.

Digital copies are supposed to be lossless, and when done from person to person they mostly are. Files sent by email, shared over BitTorrent or copied on a USB stick are, byte for byte, identical.

But media that is socially shared over services like Facebook, Twitter and Instagram is not copied in a lossless way. A photo uploaded to one of these services is optimised for delivery across that platform, usually using the JPEG compression algorithm, to reduce its dimensions and file size. JPEG is a lossy format in that it throws away unnecessary information. Most of the time this is not a problem, and the JPEG format is ultimately a wonderful thing, but like the waves on the beach, each copy throws something away and over time the changes to the image move from invisible to subtle to blatant. The same applies to mp3/aac for music and mpeg for video – they are analogous to cassette tapes and VHS video.

This distinction between making a lossless copy and sharing a lossy copy over a service is important. The latter is mediated, and that mediation introduces imperfections. That is what I’m exploring here.

Alvin Lucier

When I decided to use Alvin Lucier’s I Am Sitting In A Room as my framework it wasn’t a serious choice, but on reflection I think I can massage it into making sense. Lucier wasn’t so much interested in the technical phenomena of media degradation. He was interested in how this mechanical process could tell us something about the room itself.

Social spaces on the internet are analogous to “rooms” as they are containers for human activity and artefacts. This work is on Instagram but it is also about Instagram in a similar way to how Lucier’s work in about that specific room, and hopefully can be used to understand some general things about our online social spaces.

Instagram

Instagram is a division of Facebook, one of the largest companies in the world. This is important and there are many things that can be said about Instagram as a company, but this work does not directly deal with them.

What is interesting about Instagram to me, as a photographer, is how the platform is based entirely around photographs (and short videos, but those are effectively moving photographs). What text there is on Instagram is in service to the image, and as such the language of Instagram is visual. People are using it to communicate using images and developing dialects and slangs in that medium. This is massively interesting for a photographer.

But Instagram is also a mediating force. There are restrictions on what can be posted and how it is seen. Your photo has to be cropped to a square and will be rendered at the relatively low resolution of 640 pixels wide. Your stream of photos, by those you follow, can also only be seen on a smartphone or tablet, meaning the image is usually seen in the palm of someone’s hand, not on a large monitor. These restrictions are not bad, per se, but they do motivate creativity in how people use that service. If you want to communicate something specific you have to understand and work with Instagram’s limitations. This is what I mean by mediated.

Regramming

Most online social platforms have a method for sharing media within them. Twitter has the retweet, Facebook the share, Tumblr the reblog. These function to take someone else’s content and insert it into your personal stream. Instagram does not have this functionality, which makes sense as Instagram is about showing your own photos, not other people’s. But the usage of Instagram’s platform has spread outside this narrow limitation and there is a subculture of “regramming” on the service where people post copies of images found on other people’s streams. These images are usually image macro memes with captions which quickly communicate a message or idea and are a common way of group-bonding online. They also serve as a call to action, soliciting comments and faves from followers.

I became interested in regramming through my rabbits’ Instagram account which follows a hundreds of other rabbit accounts, many of which are run by teenage girls. This gives me a rare glimpse into how teenagers use images to communicate, something not usually afforded to a 40-something childless man. One meme in particular caught my attention – the “Which One Am I?” macro, where girls ask their followers to put them in a category. I’ve been collecting them for a few months and have them online here.

whichami-screenshot

These memes are spread through taking a screen capture (on the iPhone this is done by holding the Home and Power buttons down together til the screen flashes) and posting this screen capture as a new image. The process of doing so introduces compression artefacts which, along with inaccurate cropping, mean that each regram is of a lower quality than its original.

Brian Feldman named these low quality images in his now-seminal article The Triumphant Rise of the Shitpic. Some quotes:

“Shitpics happen when an image is put through some diabolical combination of uploading, screencapping, filtering, cropping, and reuploading. They are particularly popular on Instagram.”

“Perhaps most importantly, the Shitpic aesthetic could very well be the first non-numeric indicator of viral dissemination.”

“If you look at a Shitpic, you can instantly tell the level of virality by how worn it looks, how legible its text is, how many watermarks adorn it. You can count them much like you would rings on a tree. A pristine-looking meme engenders skepticism – “This can’t be that funny, it hasn’t been imperfectly replicated enough.” But when you see that blurry text, partially cut off by the top of the frame, and a heavily compressed picture… that’s when you know: This is gonna be a good-ass meme.”

The shitpic is fascinating not just on a technical level. It is a surprisingly fertile window into how digital images exist online to spread ideas and develop cultural norms.