I'd like to put forth a bold hypothesis. You have one of these in your skull and its an integral part of your visual system. Anyone who has experimented with psychedelic drugs has seen the intricate geometrical and fractal patterns. The explanation for these visions is that the brain works on feedback loops and the brain waves, not the individual neurons, are the currency of representation and thought. This idea has been rediscovered and elaborated independently several times by rational psychonauts since the discovery of LSD, most notably in Steven Lehar's PhD Thesis (1994) and James Kent's book Psychedelic Information Theory (2010). This feedback wave theory of brain function has basically been experimentally confirmed with fMRI data in the papers "Harmonic brain modes: a unifying framework for linking space and time in brain dynamics" and "Connectome-harmonic decomposition of human brain activity reveals dynamical repertoire re-organization under LSD" (both from Selene Atasoy 2017).
So back to my bold hypothesis. The optical nerve connects the eyeballs from way up in the front of the skull to the occipital lobe way in the back of the skull. The conventional view is that the placement of this part of the brain and the long nerve connecting to it is a mistake of evolution which can't be refactored out in incremental steps. My hypothesis is that the lengthy optical nerve isn't a mistake and isn't just a wire but is in fact an integral part of how vision works. The optical nerve is a delay line, with different paths through the optical nerve being more or less circuitous. The available paths through the optical nerve can result in simple transformations such as rotation, translation, contraction, diffusion (gaussian blur) and time-averaging (temporal gaussian blur).
What I'm suggesting is perhaps the lengthy optical nerve wires in our skull literally serve the same function as the long tube with the screen/cameras in this machine. Maybe the output images of this machine look nice to our brains because they are exactly the same patterns our brain is already looking for and thus satisfactory comprehension is immediate. The two parallel cameras on two parallel long tubes in this machine bears an uncanny resemblance to the eyes are situated in the skull and connected to the brain.
I totally agree there is something important going on with feedback loops and the brain (and vision, and the mind) - especially loops made of loops (made of loops).
I wrote something up around 2010 on my thoughts about all this stuff when I was first contemplating the HD feedback device. You might find it interesting - you can read it here: https://walkswithdave.tumblr.com/post/38950021331
I recommend Douglas Hofstadter's books Gödel, Escher, Bach (where I first read about video feedback) and I Am A Strange Loop.
"The core intellectual claim, then, is much the same as that of Gödel, Escher, Bach: namely, that a proper understanding of Gödel's proof helps us to see that life, mind and self are all constituted not by biochemistry but by the higher-level patterns that biochemistry makes possible. In particular, human selves are abstract self-referential (reflexively looping) patterns that arise spontaneously out of the meaningless base of neural activity.
These patterns are real, albeit abstract. And they have real causal power, even though they are epiphenomena generated by the brain. They affect other patterns within the mind-complex, and they loop back into the brain itself. Unlike the complex patterns generated when a video camera is focused on a television screen showing its own output (an analogy repeatedly used to demystify the notion of strange loops), the human-self patterns can determine changes within the system's hardware. But this mind-brain causation is nothing "spooky" or supernatural. It's an emergent consequence of the physical complexity of human brains."
Hebb claimed short term memory was a “reverberation,” based on Lashley’s notion of interference patterns and standing waves in the brain. Karl Pribram details a fair amount in his last book.
Yay to Lehar’s work! Haven’t read his thesis, though. If you like Atasoy, check out Adam Safron’s Self-Organizing Harmonic Modes.
Tangentially, perhaps, but I wouldn't say that is the main argument.
I see the main argument as:
If the biology doesn't make sense within your current hypothesis, it is most likely the hypothesis which is wrong.
and the LSD argument is not about scientists taking drugs, but about scientists paying attention to evidence which comes from taking drugs. Which is very reasonable. Certainly as reasonable as other ablation studies, which are commonly accepted as good scientific practice (when kept within ethical bounds).
If some scientific discoveries were already made with of LSD, would you know? Its not like the scientists would admit to how they made the discovery in the methods section.
This looks like what everyone who's ever owned a video camera and a crt tv has automatically done within the first second or so of owning them, point the camera at the screen! :)
It's a nice effect though, so it's neat that someone actually built something around it.
I started by pointing a camera at CRT TVs many years ago, and for a long time wanted to be able to do that with HD monitors (and, of course, HD TVs don't have the all-important knobs needed to create controlled feedback).
It was finding the HD Panasonic field monitors a couple years ago with the hue / brightness /contrast / saturation knobs for only a couple hundred dollars used (they were initially a few thousand dollars) that made this project possible.
Ever play with analog synths on a CRT? You can make some pretty incredible structure by chance, with waves. (Of course, you can make any image possible with wave combinations, but the chances are awfully low)
Beautiful imagery, reminds me of playing around with fractal software as a kid.
You could probably get a wider variety of image by adding more nonlinearity to the mix - As far as I can tell, this system is just iterating rotation and scaling, which are linear maps.
Ah this is just lovely, nice work. I had a primitive idea for something like this about 20 years while in media arts school, but sadly never followed through with this. Had no idea such lovely visuals could come from this!
Well often when you build something in the software world we are told not to go learn a new language to build it but just stay with the tools you are a master of, so probably this guy was a master of the tools they chose and hence it was not made as software that would have necessitated learning new tools.
However the other thing is that as a system of analog tools it may be able to achieve all sorts of effects that might require a lot of processing power, development time, and sources of entropy for randomization to match. In short I think not just that what was built was easier for the person who built it, but might also be significantly cheaper comparatively.
on edit: of course they can't distribute their machine the same way you can code but that's the tradeoff
I wasn't saying what you did was easy or cheap, but wondering if it might not be easier for you to have made it this way than programming it, or that despite being expensive to make might it not have been more expensive to build as software and if software would have been able to achieve the same things as easily.
I wrote this in the comments below, but - asking why it's not done in software is like asking a painter why they don't just use a camera and take a photograph.
Ok fair point, but I was answering someone who was probably a software developer or something similar and pointing out that there were also potential 'practical' reasons from the standpoint of software development why this could be considered a more effective way of achieving an end. (as far as if my surmise is correct can't be sure)
It depends on the software. Anyway there's just different ways to create automata. There's no reason why one way is better or worse than others. Painting is no better or worse than photography or sculpture, etc.
I've thought about simulating it in Blender. Not because it's be cooler. But, because I can't afford nor do I have the time to build such a glorious kinetic sculpture.
I think analog buttons to control rotation, zoom, contrast, ... are more easy to use and tweak until you get a nice image than a slide bar or a number in a form.
Probably the monitor to camera part is easier to do in software, but as other comment says the author probably is more familiar with monitor and cameras and understand better how to tweak the values to get nice images.
I'm not sure if you're just joking here (I suspect you are) - but, I say no computers because a computer didn't create these images, you know, with software. Yeah, there's electronics involved, of course.
Great, beautiful work BTW! Good choice using field monitors instead of consumer screens but consumer cameras usually perform image compression, even on live HDMI outputs. Not sure if that applies to your cameras but am curious if high frequency detail would become more apparent between RAW/compressed with all these feedback loops amplifying tiny differences.
So... I read Hofstadter on video feedback systems, and I experimented with IFS-style fractals in software. I never connected the two until I saw this machine in action!
Fantastic piece of art. I love how it makes the process of iteratively drawing things => fractals into a physical, tangible thing.
So back to my bold hypothesis. The optical nerve connects the eyeballs from way up in the front of the skull to the occipital lobe way in the back of the skull. The conventional view is that the placement of this part of the brain and the long nerve connecting to it is a mistake of evolution which can't be refactored out in incremental steps. My hypothesis is that the lengthy optical nerve isn't a mistake and isn't just a wire but is in fact an integral part of how vision works. The optical nerve is a delay line, with different paths through the optical nerve being more or less circuitous. The available paths through the optical nerve can result in simple transformations such as rotation, translation, contraction, diffusion (gaussian blur) and time-averaging (temporal gaussian blur).
What I'm suggesting is perhaps the lengthy optical nerve wires in our skull literally serve the same function as the long tube with the screen/cameras in this machine. Maybe the output images of this machine look nice to our brains because they are exactly the same patterns our brain is already looking for and thus satisfactory comprehension is immediate. The two parallel cameras on two parallel long tubes in this machine bears an uncanny resemblance to the eyes are situated in the skull and connected to the brain.