Editing and Final Scene Integration
This week I focused on the post-production stage of the project, where I began assembling my two AI-generated scenes with the opening and ending sequences created by my teammates. The main challenge at this stage was not generating content, but integrating materials produced by different people using different AI tools.
Although we used similar AI systems, each of us had different prompting styles, which resulted in noticeable variations in tone, composition, and visual consistency. This made me realise that AI-generated content is not inherently unified, even when the same tools are used. Instead, coherence depends heavily on human decisions made during both generation and editing.
Visual and Stylistic Unification
To address this issue, I focused on colour grading and visual alignment. I attempted to reduce inconsistencies between scenes by applying a more unified colour tone, emphasising warm desert hues and soft contrast across all sequences. This helped create a more continuous visual language, even when the original outputs differed.
However, I also realised that full consistency is difficult to achieve, because each AI-generated shot carries its own internal logic. Therefore, the role of editing becomes not only technical, but also interpretative, deciding which differences should be minimised and which should be kept.
Sound Design and Immersion
Another important development this week was the introduction of sound. By adding ambient wind, sand movement, and subtle environmental sound effects, the visual sequences became more immersive and emotionally engaging.
Sound helped to unify the fragmented visual outputs, creating a stronger sense of continuity between shots. In this sense, sound design became a critical tool for bridging gaps between AI-generated sequences.
Reflection
This stage of the project made me realise that AI does not remove the need for post-production; instead, it increases its importance. The role of the editor becomes essential in shaping coherence from fragmented AI outputs.
I also began to understand that authorship in this workflow is distributed but still controlled. While AI generates the raw material, meaning and continuity are constructed through human editing decisions. Therefore, editing is not just a technical process, but a key site of creative authorship.