Category: Capstone

Posted on

Capstone 2: Entry 2

An Unbiased Written Review of a Public Exhibition of a Capstone Project as Written by its Creator

In the same vein as the most famous critics will rave about the paintings of Mark Rothko, poetry critics will someday expound upon the virtuoso poet, Azariah Kierson-Galeano, that was able at the young age of 23 to produce an algorithm that produced poems on the same level as famed haiku author Basho Matsuo. The poems were both enigmatic and surrealist combining the best qualities of the short form poems with a certain Dali-esque flair.

If you are lucky enough to be in the same room as the working Turkish Typewriter prepare to be delighted by the simple, yet tactful selection of words. The project is an ambitious look at how we are monitored in everyday life as well as how our words can be twisted and turned in meaning. There is no better example of subversion in fabricated form yet to date.

Exhibition goers gave such praise as expounding on how the project utilizes the Analog and moves away from today’s heavy saturation of computers whose form lacks function. Praise was also given for the 1980’s storeroom aesthetic, which perfectly encapsulated the viewer’s feelings about monitoring and the surveillance state we are entering.

After users began experimenting with the machine, they quickly took to explain it to other users, who, in turn, spread the news like wildfire. Using focused phrases, users began to move away from simplistic poems and began to utilize concise phrasing to produce works of art aided by the wonderous machine.

Posted on

Capstone 2: Entry 1

Progress Report

The 50% marker for my Capstone project is fast approaching. Looking back it is important to take note on what has been accomplished thus far. Additionally, pains should be taken to document where challenges arose and where corners were cut in order to create a more feasible project.

Interaction

But what is intended for this project? How will people interact with the final installation piece?

Looking back, my initial idea was for the project to be presented as part of an art gallery whereby people’s discussions of the art would be recorded, parsed, and reorganized into a poem. This poem would ideally serve as a memento of the art gallery. However, in order to be part of the art gallery, it would necessitate that the project itself was art. As modern art is defined less by what it is and more by what represents, the “why” of the project still needed to be fleshed out.

The artist’s statement of the project would read something akin to:

What are poems and why do poems make us feel things? Why are poems’ value derived so much from the author? Can poems exist in a vacuum? Without an author, this poems attempt to encapsulate the zeitgeist of the temporial moment.

Code

Continuing from this artist’s statement the code would have to create poems that were able to capture the essence of one’s speech without understanding the meaning of the speech itself.

Most of my code progress is documented on my Github. As such, I have mostly been refining my code and making sure that it works as intended. A large part of the issues with the coding was related to the lack of structure in poems and poems being perceived as poems solely because someone says it is a poem.

The code works as follows. Speech is heard by the system’s microphone, using a wrapper for pyaudio called speech_recognition. The audio is then sent to either Sphinx, Google Speech-to-Text, Houndify, or IBM using the same library. However, there is currently an issue with IBM’s service and the library. The returned audio is then parsed into a PoemGenerator class, which uses two variables for the corpus with stop words and a stop word free corpus. A random word is chosen to end the first line. The rest of the line is chosen via Markov Chains until the line is at its syllabic length. The next lines created in a similar manner with their last words rhyming with a previous line. Rhyming is done using nltk.

Technical

On a more functional level, I have completed an Eagle file for connecting an Arduino to the typewriter. I have also looked at tables hoping to find one that is about the right size to sit the typewriter on and have the computer and microphone under.

Posted on

Capstone: Entry 10

Prototypes 3: The Final Entry

Technical

For this set of prototypes, I looked further into how to hook up an Arduino to my typewriter. At first, I tried to set up a matrix assuming that one of the pins was for power and ground. Assuming this, I wrote out the matrix for the lettering as matric in Arduino. This wouldn’t work and I will likely move others to a four-pin output.

An additional note about this is that there are several keys – SHIFT, CAPS LOCK, and CODE – that need to be pressed in conjunction with another key in order to work.

All pins connected

This didn’t work as the keyboard was mostly mechanical and half of the pins were actually grounds. I switched over to using transistors. Given the number of transistors needed – at one per key, I looked into buying PCBs and how I would wire up the 30 transistors and pins needed. I ran into additional problems as the 1980’s typewriter’s ground was in the back of the machine and not built as a pin into the IC.

The code to run this was fairly easy. The one problem was testing how long the button should be pressed in order to only output one keystroke. I found the system works between 50 and 300 ms but 1s will fire multiple times and 10ms will not fire the key.

Circuit running

I started to learn EAGLE and other CAD software to create a keyboard that could be attached to an Arduino. this PCB would have

If I was to use a regular Arduino I could add an increased number of transistors and create a binary code to switch between various numbers. This would require more coding than is totally necessary.

Role

I also worked a bit on the role of the projects. Beyond just being an art installation, I thought some more about where it would sit and what sort of people would be near it. If I were to install it in a nearly silent gallery, the generator would not always have a large enough corpus to read from and would start to spit out very generic poems based on the starting corpus. Conversely, if I put the machine in too loud of a place, the ambient noise would be too distracting for the machine to run and complete sentences could not form in the corpus.

Thinking more about the statement of the machine, I looked into putting the project in an open mic night and having the poem create a meta-poem about the night. Alternatively, I could read the poem as my own at the end of the night and see if I could pass it off as the genuine proof of human emotion.

Posted on

Capstone: Entry 9

Prototypes 2

This round of prototypes was to expand on the first round and look into what worked and what didn’t work. Mostly, I worked on the techical implementation of the project and less on the role and look and feel. The role, I feel is fairly settled as an art installation piece. On the other hand, the look and feel will be heavily dictated by what is technically possible and what parts need to be incorporated into the project. Additionally, as the project is not handheld nor physically interacted with, the look and feel is not a primary factor in its usability.

There were two parts to my prototyping: coding and fabrication. In the last round of prototypes was able to write a Markov text generator that produced poems based on a large enough corpus. For this round, I worked and using speech recognition. This was a lot easier than I anticipated as most of the machine learning required for speech recognition was already done by Google and I was able to just use their API.

Link to github:
https://github.com/akierson/capstone-prototype

The fabrication was a bit harder. The typewriter I had purchased was luckily in perfect functioning order and all of the electronics and mechanical parts did not appear on their last leg despite being several decades old.

Testing out the typewriter and realizing I rely heavily on the backspace key

After testing that it worked, I tried to look for schematics on how they were wired. I was hoping to bypass the keyboard and only use the carriage for printing but I was unable to find any and the manufacturer has yet to contact me back.

I did try to hook up an Arduino to the carriage wiring but the Arduino was did not provide enough voltage and I was unwilling to try the 100V+ that some people online suggested.

Next, I took apart the keyboard to see if using solenoids was possible with the keyboard. I had expected the keyboard to have a mechanism more similar to today’s mechanical keyboards, but it instead had a rubber pad and keys without springs, instead relying on metal tabs to push the keys back up. The keys also have individual cases for a peg underneath them which I could potentially use to house a solenoid instead.

The smallest solenoids are about 12mm by 20mm with a 4mm pinhead. The head could easily take the place of the key peg and, with some sanding, the solenoids could replace the peg housing.

Posted on

Capstone: Entry 8

Prototypes

Most prototypes should answer a question: Is this feasible? Will people hate it? Does this project achieve its purpose? For these prototypes, there were three areas that needed to be prototyped: role, look/feel, and technical

I created some early musings about the project. I was trying to decide on the best method of using a typewriter as a printer and how best to write poems from streams of words.

Parsing incoming words and early ideas on project goals
Algorithm flow and solenoid adaption on typewriter keys.

Role

For my role prototype, I looked at art installations that interact with the viewers and created a short storyboard from those. The idea was that the typewriter would appear apart from the gallery until the viewers leave and notice that the typewriter has made a poem just for them.

Story board of art viewers in gallery

Look/Feel

For the look feel prototype, I purchased a word processor and set it on my desk for a few days to decide if it was cool enough. I also looked at the Media Archaeology Lab for more refined typewriters such as the fully mechanical ones. I decided that the typewriter should ideally be connected to the computer housing in a permanent way.

Next, I sketched some tables-housings. The table design should ideally match the typewriter. If I am able to get a mechanical typewriter, the table should be a refined one but, if I can only get a word-processor, a plastic table is almost preferred.

Sketches of table-housings
Mockup of sketches

Technical

For my technical implement, I did several iterations on the same algorithm of using Markov chain text and the nltk Python library. The algorithm would take in a string of text, the longer the better. I used the Sugar Hill Gang’s “Rapper’s Delight” and the full corpus of Shakespeare’s sonnets.

I first started with haikus as 5 -7 -5 to test out the syllable library. Next, I made basic sonnets of 14 lines with 10 syllables each in the style of Shakespeare. My last step was to add rhymeing in the form of ABAB-CDCD-EFEF-GG. I ran into some trouble because occasionally words don’t have rhymes, but in a further prototype or the final iteration this could be achieved through recursive programming and chaning the parameters each time. This would allow for less exact rhymes and for the program to rewrite lines that don’t fit into the sonnet.

https://github.com/akierson/capstone-prototype

Posted on

Capstone: Entry 7

What do prototypes prototype?

  • Create prototype for audience
  • Should show roles, implementation, look and feel

For my prototype, I would like to make a working prototype of converting a data stream into visual graphics. Specifically, I would like to work toward an Authenticity app and the fashion side of this. Though it is nebulous, I think by using sample data from Google trends and Twitter data I could compile the relative popularity of common brands such as Nike and Adidas by urban area.

By creating a focused data set and having set desired outcomes, I could work towards a more nebulous idea of assigning regional popular to various items using data scrapped from various websites. There will be the added challenge of having to backpropagate my data based on future sales releases and data from the aforementioned companies.

However, if scarpaing the data from Instagram, Twitter, and Google proves harder then expected (as they are tightening their APIs), I may also look into fake fashion writing by analyzing top fashion blogs and associated images and counterfeiting them. This coulld also be incorporated into taking pictures of viewers and spitting out a fashion/tablod piece on them.

Posted on

Capstone: Entry 6

AAR: Case Studies

I looked into several projects that circled inthe general direction of my project. I am still missing one idea but the other three are almost fully fleshed out.

Klout

The first case study is Klout. It was an early attempt at codifying someone’s personal impact on social media with a single numeric metric. There were several issues with this. Critics of it came from both sides a the social media gamut. The proponents of social media said that codifying social media into a single number showed users the man behind the curtain and that metrics should be more holistic. They believed that users would modify their behaviors if they were shown how the direct impact of their actions in a “Hawthorne Effect”. Critics of social media worried that giving users the tools to enhance their social media presence would lead to things like the Cambridge Analytica incident in 2016. Of course, they ended up being right.

I also found it interesting to look at how, while heavily criticized for trying to gamify social media, Klout changed how social media giants track data and use data. Facebook and Instagram changed their timelines to increase attention. Facebook also started to offer data tracking to companies in exchange for their data. Other apps have also tried to capitalize on the void by appealing to “influencers” on Youtube and Instagram.

Deepfakes

My second case study is on Deepfakes and GAN. GAN, generative adversarial networks, were first proposed by Ian Goodfellow in 2014 as a new way of training neural networks, whereby a generator neural network is taught to try and convince a detector neural network it is legitimate. He made the analogy to counterfeiters and police. Deepfakes, a portmanteau of “deep learning” and “fakes”, arrived two years later as fake celebrity porn. A Reddit user by the same name had usedGANs to mask the face of celebrities onto adult actresses. This opened up the whole conversation of what lines this crossed, what it could be used for, and if we could ever trust images and videos again. Since then, interest in Deepfakes has waned outside of porn studios.

I found it interesting to look at as it is the cutting edge of creating fakes and can be used to fool all but the most trained eye. While making a porno of the founding fathers would be interesting, it is also interesting to look at how it had a sudden spike in intest and then people forgot about it as though they couldn’t be fooled.

PhoneStory

The third case study is on informing the public. I wanted to look at PhoneStory by MollieIndustria. PhoneStory tells the story of creating an iPhone with a series of dark minigames. It was quickly removed from Apple’s AppStore but has since been added to Google’s Play Store. I wanted to look at how the public can be informed about the error of their ways in non-aggreivating ways.

Subversion

For my last case study, I wanted to look at something subversive. While most of the previous case studies can be construed as subversive, it was not their explicit intent. I wanted to look at something that subverted the commercial narrative. I looked into Garbage Pail kids and the uthors previous venture, Wacky Packages. However, the pictures were often not political and solely sold for their shock and wae. I was also ooking at other rtist’s efforts to create pieces to raise convecern about McDonald’s , The Catholic Church or others. Lastly, I am looking at other commercial attempts to high jck diasters or others to subvert the narrative. in their favor.

Posted on

Capstone: Entry 5

50 + 50 Ideas

Ideas from Class:

  • RSS Feed of Deaths
  • Map showing estimated deaths in real time
  • Interactive art piece showing an individual’s impact of consumer usage with pics of dead children
  • Measuring app using dead children as a measurement for products
  • Narrative of phone creation
  • Surrealist look at consumer usage – substituting people with cats or something
  • AR for statues, war memorialsgiving wiki pages and kill counts
  • Generative art showing trade routes of products
  • Generative art showing the digital impact
  • Music device with triggers on social media impact – likes, retweets etc.
  • Music triggers for war tolls
  • More accurate war memorials
  • AR showing massacre locations
  • Crime feed in social media look
  • App alerting users to places where crimes took place
  • AR HUD
  • AR HUD making us all look like sims
  • AR HUD showing social media data of people
  • Brain mapping cognitive biases
  • Art project showing company slogans and public images with enviromental impact attached
  • Recreation of ads showing death toll
  • AR showing everyone’s indirect kill count
  • Social media posts whenever someone contributes to killing the enviroment
  • Social media posts whenever someone contributes to slavery
  • AR showing products that are indirectly supporting slavery
  • Masks made of corporate items
  • Chains made from nestle procuts to show their support of slavery
  • AR surrealism app mimicing LSD
  • phamtom rings on phones to gaslight user
  • Non-art installation of a room/the installation room
  • AR showing people’s clothing impact
  • HUD with clothing prices with Amazon tie in
  • Chrome extension to show products kill count on Amazon etc.
  • Generative terrian game
  • RPG fighting corporate villians
  • Get companies to support a non-existent video project – with a video project at the end showing the company’s enviromental impact.
  • Impossible monopoly-esque game about corporate take over the app
  • Phone ad blocker
  • Generative hat maker
  • Generative Clothing maker
  • Generative art in the style of HR Giger
  • Platforming where the terrian is based on tweets – the player has to tweet enough and get enugh replies to continue getting land to run to
  • Platformer where the in game currency is created by social media clout
  • Likelihood someone could commit a crime and get away with it based on social media clout
  • Death Metal bluetooth hi jacking somehow replace localized notification sounds with death metal
  • Gernerative tweets based on someone’s social media that people might tweet after their death
  • Obituary creator from people’s social media
  • Calculator for meauring someone’s authenticity
  • calculator for the originality of ideas
  • Rap generator

Ideas from 2nd round

I found the task interesting and that most of my ideas became centered around a few central themes. Most of my ideas were then variants on these themes. I wanted to work with social media and people’s digital footprints, I wanted to have some sort of generative process either with machine learning or algorithmic, and I wanted to shame people for being consumers. I found that since most of my ideas were variations on similar themes, it was hard to create 100 distinct ideas.

I thought most of the feed back I was got was good. People seemed to like the HR Giger headphones more than I anticipated. I was also surprised that some of the feedback on the two subversive ideas were positive and that the person might actually use it. It surprised me as the rest of the feedback on it was positive that it would be good commentary and an nice art piece or that it was terrible.

Posted on

Capstone: Entry 4

50 + 50 Ideas

We brainstormed for an hour to produce 05 ideas. Afterwards we brinstormed again for 50 more ideas. MOst of of my ideas had similar themes, including, the usage of personal data, subversion, and our dark Lord Satan.

5 Ideas

These are my five best ideas:

HR Giger Headphones

I have always liked HR Giger and while I have mae a few pieces based on his owrk I have not created any technical creations. I think headphones would be a great way to display this style because of Giger’s frequent use of mechanical elements.

Sculpture by Giger
Sketches of possible head phones

Authenticity App

Have you ever wanted to appear like a unique authentic individual without having to actually put in the work to have discerning tastes? Introducing the Authencity App which will give suggestions based on your location and social demographic. This app will suggest clothing, songs, and photos that are becoming popular but not in your location so you can appear cutting edge and like a cool kid.

This app is more of a subversive idea based on two of my good friends, one who gives all the appearance of having discerning and unique tastes while actually just knowing a guy in a large metropolitan area who keeps him in the know. The other likes to give the appearance of having really unique tastes, abhoring anything commercial and, despite mostly shopping at H&M, won’t be caught dead in a foreign country with one of their bags. I think it would be an interesting look into how being an authentic and unique person has become a pursuit in and of itself instead of the interesting things and events being the pursuit.

Music Remixer

I have always wanted to be able to play an instrument and be good at making music but I have no rhythm. Music, however, can be made using fairly set rules. In this project, I would create an application that could take in a sounds clip, song, or YouTube video and spit out a song in the style of remix music such as dubstep, techno, and house music.

While the music may not sound good at first, with some discerning humans judging it, I think with several iterations the machine could learn to produce modernly okay YouTube poops.

App for Crime Awareness

this project was inspired by a project that someone had done in my design foundations course. The idea is to use the police feed in a given area to put markers where crimes have happened. Whne user’s are then inthe vicinity of where the crime happened, it would alert the user.

This would be part activism, part informative as it could be generalized to all crimes or specifically, hate crimes, sexual crimes, etc.

Post Optimizer

Back in my formative years, my friends and I ran bots on Instagram and Twitter to gain followers from people who would #f4f. This project would take the gamification of social media to a new level by giving optimizations to comments, posts, and hashtag usage.

I liked this one because it’s awful and I hate it, but the challenge of it could make it worth it.

Posted on

Capstone: Entry 3

Why?

Being a cynic, I found it hard to watch Simon Sinek’s talk and take the selling seriously. I have a hard time believing that people actually make products with deeper intent. Between my father having me read business books that extolled the virtues and “bending the truth” to customers and the increasing shift toward corporate virtue signaling, I have become jaded from this selling of ideas and identities instead of people making informed decisions about the actual reality of products.

Corporations, am I right?

What is my why? Why do I get out of bed in the morning? Why do I go to my classes? Why am I even in college? Why do I make things?

My “why” is a draw to pull things apart and look at them in a Sartrean light that shows them for what they are outside of some contrived context. Sometimes this comes across as callous, sacrilegious, or perverse, but that’s all subjective.