Back to Parent

Outcome

Move up Delete Move down

Intent

In our investigation, we explored the intersection of prediction and technology through a device embedded with machine learning. Through the lens of tasseography, we explored how we could use mysticism and belief to imbue rituals with meaning - magnifying the mundane action of stirring a teacup into an omen of the future. Tea rituals are ancient and persistent, and the act of reading one’s tea leaves has long been practiced throughout medieval Europe. Traditionally, tea reading is predominantly an exercise in interpretation. Yet we wanted to see how data-ifying the process, and imbuing it with mechanical intelligence could be used to guide one’s interpretation differently. And perhaps, rewrite the future.

Move up Delete Move down

Context

Our project was heavily influenced by Auger Loizeau’s Real Prediction Machines. In our discovery process, we frequently discussed data and how it could be used to make uncanny predictions in our lives. This parallel to forecasting rituals fascinated us, and we found that Loizeau’s use of sensors to gather data and make predictions was very compelling. We chose tasseography due to its very in depth ritualistic process. We felt that the physical artifacts involved in that ritual, including a cup, stirring spoon, and even the tea itself lent themselves well to technological integration. Dongtao’s insight that we could shine the screen through the bottom of a glass mug solidified our decision.

Move up Delete Move down

Prototype/Outcome

Our device uses the Arduino Nano 33 BLE Sense and an Adafruit 1.47" Round Rectangle TFT Display. We also utilized LED strips in the construction of the device.

The device works by pulling audio from the Microphone included on the BLE Sense board. The audio we are classifying is the sound from stirring the cup, which then is classified into a “Neutral,” “Slow,” and “Fast” stirring. These inform which images appear on the Display.

The images on the display were created in Photoshop. Since we knew the dimensions of the screen, we set a fixed canvas, downscaled reference images into it, recolored them as appropriate, and passed them into an online software called Ditherlicious. This software created dithered, 1-bit pixel art that could be passed into the TFT Display on a SIM card.

The device itself was designed in CAD, using precise measurements of the cup - which was purchased pre-made.

 


Using this 3D model, we cut the shapes from black acrylic, and assembled the final device. While it was challenging to hide all of the wiring and lights, the final result looked very polished. The only downside? Black acrylic tracks a lot of fingerprints. 

Move up Delete Move down

Process

When we chose tea leaf reading as the basis for our investigation, we knew there were several key aspects of the ritual we wanted to remain untouched. Primarily, we felt the act of peering down into the cup was crucial to the embodied experience of tea leaf reading. Secondly, we felt that giving the user the opportunity to interpret their own destinies was crucial to the experience of tea leaf reading. With these guiding principles in mind, we began our explorative process.


Our initial plan was to classify certain aspects of the user’s actions while performing the tea ritual, including tea temperature, speed of stir, and how full the cup was. However, we soon felt that this level of quantification stripped some of the mystique away from the tea ritual. Not only that, but that much classification would be beyond our programming capabilities!


However, this line of thinking brought us down a different path. Dongtao came up with the idea of using the screen beneath the teacup, and generating a pattern based on microphone input that a user could interpret, much like tea leaves in tasseography. This ultimately shaped the physical design of the product - which included a base that engulfed the cup, expanding its form and elevating it from the table. And though we ultimately shifted away from the tea leaf pattern idea, this form persisted to the final prototype. 


Speaking of the screen, we decided to instead approach the omens more literally - using classic tasseography symbols. Michael created these screens in a pixel art style - inspired by the mystery game Return of the Obra Dinn. We felt that presenting users with the omen in this way would be more engaging, and give users unfamiliar with the practice the ability to simply interpret, rather than attempt to identify symbols in what would essentially be random noise. 


Alongside screens, we were bringing the device to life using laser-cut black acrylic. To ensure a seamless experience, the device was carefully designed with the exact measurements of the glass cup we were using. There was a specialized cutout for the specific model of screen we were assigned as well. We decided to add LED strips to the device as well, to lend a more mystical, immersive experience. 

The device ultimately used the microphone and Edge impulse to classify how quickly the user was stirring the cup with the spoon, and used that input to inform the (still mostly random) omen. 

Move up Delete Move down

Open Questions and Next Steps 

While our project played with the ritual of tasseography, we think there is still much to be explored around the interpretation in this process. How can we present a leading visual without outright showing one? Does the visual need to be leading at all? Furthermore, how might we draw the user in more to the experience? We found the lighting and stage-setting very effective to bring our users into the experience. How could we expand on that? Is there a way to set the stage without human intervention?

Move up Delete Move down

Reflection

In the end, we were very happy with the results of this project. While our grand ambitions at the start regarding our ML audio capabilities were not quite achieved, we think the effect created by the device, and the embodied experience of using it were very successful. And while we ran into some technical difficulties (having to recompile every demo was not ideal!), our feedback was very positive, and people seemed to enjoy the experience of ‘tea reading’ with our device.

In the future, there are some obvious technical challenges that should be ironed out. However, to expand on our scope and intention, I think we could do more to evoke the overarching experience of tasseography. Perhaps including animated lights, and more dynamic screens inside of the cup could have lent to a more exciting reading experience. Furthermore, we got some great suggestions to utilize the lights to illuminate the cup itself. This could lend some drama to the experience, which would again, lend to a more immersive experience.

These technical shortcomings aside, I do think that our provocation was very effective. At the very least, it has prompted many discussions and opportunities for improvement. If we had more time, it would be fascinating to run down any of these rabbit holes. 

Move up Delete Move down
Drop files here or click to select

You can upload files of up to 20MB using this form.

Add Rich Text Add Code Block Add Embeddable Content