1 / 3
Tchaikovsky Reimagined is an experimental theater project that explores how classical cultural memory can be re-experienced through artificial intelligence and generative media. Transforming three canonical ballet excerpts by Pyotr Ilyich Tchaikovsky into evolving immersive environments, the work investigates how machine perception reshapes the relationship between sound, image, memory, and space.
Classical ballet historically encoded emotion through disciplined bodily choreography; in Tchaikovsky Reimagined, these movements are no longer performed by dancers alone, but reinterpreted through statistical systems that continuously generate, dissolve, and reconstruct visual form. Viewers no longer observe ballet from a distance; viewers enter its atmosphere, follow its rhythms, and inhabit its continuously shifting generative worlds.
The visual pipeline uses DayDream’s Real-Time StreamDiffusion framework to generate evolving AI imagery shaped by the emotional, rhythmic, and atmospheric qualities of Tchaikovsky’s ballet compositions. These images are transformed into three-dimensional particle environments that respond to real-time orchestral audio, with musical dynamics shaping motion, density, spatial transformation, and environmental behavior. Through this process, choreography shifts from embodied performance into computational motion and atmosphere.
Rather than reconstructing ballet historically, the project recontextualizes cultural heritage through machine perception and immersive media, asking whether cultural memory in the age of AI can move beyond preservation to become a living, generative, and continuously transforming experience.
Yuqi Hu
She is an interdisciplinary AI researcher and creative artist studying how humans and intelligent systems perceive, represent, and generate the world.
