Last semester I got a bit obsessed with court stenography—the transcript as a product of directed, selective attention and interpretation—and the tension between voice, language, and record. I started experimenting with different ways of capturing and documenting speech and questioning what counts as valid speech in different situations. 

During midterms, I shared some of my experiments with the looped-feed thermal printer. In this iteration it is connected to a speech-to-text framework (sometimes from multiple microphones) and prints everything it recognises as valid speech in real time. As the paper feed is an infinite loop, over time it starts overprinting on top of the previous transcript, and then again and again.

Thermal paper is used for many forms of administration. Its voice is bureaucratic, it’s a voice of authority. It prints receipts, reports, queueing numbers—ephemeral bits and pieces of social and legal identities that come and go just as the thermal paper itself is fading when exposed to light, heat and certain chemicals. 

Earlier this semester Sebastian and I spontaneously decided to team up to get unstuck together post Paris trip and pre midterms. We saw connections in our research—both of us looking at ritual, symbolism, and the interplay between the mundane and the sacral, as well as language and the corporeal.


Instead of working toward a single, sanitized output, we adopted the format of a conversation. We helped each other by thinking, talking, and making along—initiating a dialogue between our practices through brainstorming sessions and a series of quick, spontaneous experiments. Some of them revolved around vocal expression and genres of speech. We exchanged writing, read each other’s texts aloud, inhabited each other’s speaking and reading intonations, and created both literal and metaphorical feedback loops.

Apart from working with speech and feedback loops, Sebastian and I talked a lot about gesture and his research around hands. It made me think a lot about gesture and body language as parts of one's voice, also in legal environments. 

We also talked a lot about rituals and chanting, and how through repetition speech transforms into something else, losing its semantic meaning and becoming a different sort of utterance, or music. In psychology there's a term for this – they call it "speech-to-song illusion". So we did some experiments with recording and playing back speech, resulting in a feedback loop where speech slowly turns into a kind of ambient melody.

Over the course of this semester I auto-saved thousands of little speech-to-text snippets on my computer. One of writing exercises in Ali's class asked to create a text collage, so I assembled a poem/polylogue/play from a selection of decontextualised speech bubbles. I don't even remember anymore which conversation most of them come from, but I feel like it captured the spirit of that time period pretty well.

When we received Flavia's sonic diary assignment I installed an app on my phone, which was better than the native app because it could record in the background, while I'm using other apps. Then some time later my phone started lagging more than usual, and then it turned out that the app was triggering the recording every single time I used the volume button. I ended up with tons of unintentional, uncurated recordings (which coincidentally feeds very well into the rest of my research this semester).

As I was listening back to them (all unlabeled, of course), I was seeking anchors that would link the sound back to its origin - a friend speaking or a specific sound I would recognise. So I started thinking about memory and space, and how the recorded sound was not just recreating the routes I’ve taken and places I’ve been, but just like distorted memories it created new geographies, almost like a dreamscape, a labyrinth populated by ghosts. In an attempt to translate that, I created an infinite memory maze, where the recordings are co-exist in a non-linear, interwoven soundscape. 

(And here I'm also thinking back to the object dialogue I've written, where a Naughty Stenomask flushes someone's voice down the toilet and leaves it wander around in the canalisation tunnels) 

Another version of the maze, which I populated with words spoken in Clara's class last semester (which she recorded and shared with us). I used natural language toolkit (NLKT) to extract most commonly used words and phrases.

I was also playing around with reassembling the extracted bits into a single track: 

Some stuff I've written for Tina's in medias res assignment

Together with Seb and Doris we are (still) working on space design for the graduation. The main challenge is to accommodate sufficient space for performative works, for which we came up with a "fluid stage" structure with curtains, that can be opened and closed depending on the program. 

A /yet-to-be-finished/ collective documentation of the Paris trip. Each upload (photo/video/sound/text – we've all been keeping our own diaries throughout the trip) is a Tarot card with an assigned arcana, laid out in specific formations and depending on their position carrying different meanings. It got way too busy later in the semester, and the development slowed down, but we're hoping to still have a tarot reading session in the future (maybe when we organise a show in october). 

I've mostly worked on the code, interactions and the overall concept. Doris and Mel worked on the formations and the rest of the lore, as well as the drawing for back of the cards. Woory and Tine worked on the overall design.

Work outside of Sandberg was also aligning with my research into voice and record. Since late February we are collaborating with Tegenlicht and VPRO Medialab. So far it's been quite exciting, cause the job feels more like a residency – we are free to choose what we do as long as it's interactive and aligns with the current topics Tegenlicht journalists are working on. Also, we don't have to officially launch anything in this round – so we're working on prototypes rather than finished pieces, which means even less editorial control from VPRO's side. 

At the moment, Tegenlicht are working on a series of episodes about the rise of AfD in Germany. We scraped recordings of AfD members that were published on Bundestag website and created a vocal search engine, which allows to gain insight into how often, by whom and with what emotional connotation the searched words were used (in the prototype the records only date back to 2022, which could potentially be much longer). 

Portfolio with my work from the previous semester is still online here:
https://portfolioassessments.pointer.click/