── .✦ EXPERIMENT 2 ✦.──






2A ꕀ Translating Memories to AI-Generated Music

OBJECTIVE
Create a non-lyrical melody using generative AI, using text prompts to guide the composition. Experiment with the melody's structure, tone, and dynamics to evoke the emotions associated with the recalled memory, exploring how sound can mirror or reinterpret our feelings.

The Tasks:
  • ✶ Recall a memory
  • ✶ Summarise in text
  • ✶ Prompt & Generate music in AI


2.1-Sounded too intense for the memory
2.1-Accurate, Happy -> Panic -> Sad

2.2-Too short, no change of mood
2.2-Accurate, Anxious -> change of mood -> Relax
    • 2.1 'The Lost Plushie'
    • 2.2 'Cycling to School Results day'

Translating Memories to AI-Generated Visuals

The Tasks:
  • ✶ Using the summary, generate an album cover art using AI
  • ✶ Summarise in text
  • ✶ Prompt & Generate music in AI


OBJECTIVE Putting the AI song and AI cover art together to create a soundtrack for each memory.

RESULTS Realistic styles feel more impactful due to the moody colors. After viewing the generated images, I realized I no longer clearly remember the plush toy's face. Interestingly, the one where it appears the most ragged feels closest to my fragmented memory. However, the outcome differed slightly from my recollection, as it wasn’t indoors as I remembered. Yet, upon enlargement, it’s hard to tell that it’s an outdoor car park. Despite many prompts, accurately recreating the memory through this method proved challenging.




an album collection of memory soundscapes
Generative AI Album Cover art using Adobe Firefly