top of page

In Just Noticeable Difference, dreams are no longer private, individual experiences or visually dominated spaces; they form a perceptual system shaped by sound, memory, and fiction.

 

This project constructs an experimental loop of dreams—music—memory: Over 60 days, two dreamers and the composer record their dreams daily. The composer extracts sonic elements from these records, evolving them into music, which is updated every five days. Before sleep, the dreamers listen to the composition, enhancing their auditory perception within dreams and shaping a richer sonic narrative.

 

However, the authenticity of the dream records remains uncertain. During dreamless periods, the assistant collecting the records may use AI to generate fabricated dream texts, ensuring the continuity of the composer’s musical development. As a result, the composer cannot distinguish between genuine and AI-generated dream records. These texts do more than replicate each dreamer’s linguistic style—they construct a distinct auditory dream experience, simulating spatial perception, dynamic shifts, timbral layers, and emotional depth. Nearly indistinguishable from real records, they infiltrate the composer’s creative process, making their work both sincere and unknowingly influenced by AI’s interventions.

 

Yet, fabrication is not exclusive to AI—the human brain itself unconsciously reconstructs dream memories. As AI-generated dream records merge with natural memory distortions, reality erodes; the "just noticeable difference" threshold is crossed, making truth and fiction increasingly indistinguishable. The composer operates within this ambiguous space, while the dreamers absorb these auditory inputs before sleep, feeding them back into new dream memories. Dreaming space thus becomes a malleable system, dynamically shaped by the recursive cultivation of sound. More significantly, each dreamer generates unique dreams, which in turn become part of the composer’s music. Dreamers do not passively receive these inputs—they hear each other’s varied sonic responses, tracing subtle divergences in their dreamscapes, forming a network of shared and transmitted sound geography.

 

This project is therefore not merely an experiment on three dreamers but a reflection of how shared perception is imperceptibly shaped. It seeks to reveal the mechanisms through which reality is constructed in this era—where individuals, as part of an interconnected perceptual system, unknowingly internalize and adapt to new frameworks of reality, continuously reinforcing them within the collective consciousness without perceiving the subtle shifts in their origins.

spectrogram_print
Play Video

The Audio Spectrogram Printed Using Chinese Characters from The Original Dream Records

(This project was supported by a 2025 CCRMA Project Award.)

  • SoundCloud
  • Instagram

©2022 by Ningxin Zhang

bottom of page