2022
SHIT(T)Y Vol.1
A theater performance with real-time visuals & ML-images
CLIENT
Analogtheater
in cooperation with Studiobühne Köln
from/with
Dorothea Förtsch, Lara Pietjou, Ingmar Skrinjar, Sophie Roßfeld, Hanna Held, Isabella Kolb, Daniel Schüßler.
TEXT
Emsemble
STAGE & COSTUME
Eva Sauermann
DIRECTION
Daniel Schüßler
VISUALS & AI-IMAGES
NEOANALOG
TECHNICAL LEAD, AI-IMAGES & CAMERA
Thomas Vella
MUSIC & COMPOSITION
PRODUCTION- AND DIRECTION ASSISTANT
Hanna Held
DIRECTION ASSISTANT
Isabella Kolb
MANAGEMENT
PR
OUTSIDE EYE
Tim Mrosek
INTERN
Nabi Wenke
Photos
Nathan Ishar & Ingo Solms
SUPPORTED
by the Cultural Office of the City of Cologne, the Ministry for Culture and Science NRW and the RheinEnergieStiftung Kultur
»A booming sound performance about Cologne, life in the city and the Me in space.«
Daniel Schüßler
Black Void
The set resembles a black cube lit by two projections – a wall and circle in the center.
The introduction takes place outside the black box.
Real–Time
ML-images are put together in various ways. Using TouchDesigner we manipulate them in real-time.
The visuals on the wall give context, atmosphere and are also used to light the room.
The floor projection in the center structures the room, lights the actors and is used as stage.
Within the machine hallucination – Driven by Ben Lauber’s musical compositions.
Using TouchDesigner we were able to produce parallelly. It gave us the flexibility to change and tweak everything up to the last minute.
Zooming out through the city with a ML »outpainting« workflow and exponential scaling.
ML-City
The images that resemble a subjective view on the city are – without exception – made with a Mashine Learning solution.
The prompt: »A big office building in the shape of the Cologne cathedral«
Generated images are extended by adding more and more objects.
Synthetic skin promt: »close up of old furrowed skin with a scar, photo«
These »typical« street scenes are by no means real. They resemble the subjective view of the Machine Learning algorithm, based on how it conceives our world.
Process
We used OpenAI’s DALL·E 2 to create the images, TouchDesigner to bring them to life and TouchOSC to steer the whole show from one iPad: Audio, lights and visuals.
Related projects
2022
SHIT(T)Y Vol.1
A theater performance with real-time visuals & ML-images
CLIENT
Analogtheater
in cooperation with Studiobühne Köln
from/with
Dorothea Förtsch, Lara Pietjou, Ingmar Skrinjar, Sophie Roßfeld, Hanna Held, Isabella Kolb, Daniel Schüßler.
TEXT
Emsemble
STAGE & COSTUME
Eva Sauermann
DIRECTION
Daniel Schüßler
VISUALS & AI-IMAGES
NEOANALOG
TECHNICAL LEAD, AI-IMAGES & CAMERA
Thomas Vella
MUSIC & COMPOSITION
PRODUCTION- AND DIRECTION ASSISTANT
Hanna Held
DIRECTION ASSISTANT
Isabella Kolb
MANAGEMENT
PR
OUTSIDE EYE
Tim Mrosek
INTERN
Nabi Wenke
Photos
Nathan Ishar & Ingo Solms
SUPPORTED
by the Cultural Office of the City of Cologne, the Ministry for Culture and Science NRW and the RheinEnergieStiftung Kultur
»A booming sound performance about Cologne, life in the city and the Me in space.«
Daniel Schüßler
Black Void
The set resembles a black cube lit by two projections – a wall and circle in the center.
The introduction takes place outside the black box.
Real–Time
ML-images are put together in various ways. Using TouchDesigner we manipulate them in real-time.
The visuals on the wall give context, atmosphere and are also used to light the room.
The floor projection in the center structures the room, lights the actors and is used as stage.
Within the machine hallucination – Driven by Ben Lauber’s musical compositions.
Using TouchDesigner we were able to produce parallelly. It gave us the flexibility to change and tweak everything up to the last minute.
Zooming out through the city with a ML »outpainting« workflow and exponential scaling.
ML-City
The images that resemble a subjective view on the city are – without exception – made with a Mashine Learning solution.
The prompt: »A big office building in the shape of the Cologne cathedral«
Generated images are extended by adding more and more objects.
Synthetic skin promt: »close up of old furrowed skin with a scar, photo«
These »typical« street scenes are by no means real. They resemble the subjective view of the Machine Learning algorithm, based on how it conceives our world.
Process
We used OpenAI’s DALL·E 2 to create the images, TouchDesigner to bring them to life and TouchOSC to steer the whole show from one iPad: Audio, lights and visuals.
Related projects
© 2022 NEOANALOG studio for hybrid things & spaces