Nhan Tran’s filmmaking experience includes pulling focus, lighting sets, working in the art department and editing. On his personal website, the Cornell computer science doctoral student describes writing and directing three short films. He also studies film and video production. The problems behind his research, he explains, include difficulties he has encountered while making films himself.[2]
That combination gives his work an unusually concrete starting point: the adjustments still needed after a scene has been assembled. Micro-Edits, a research project with Adobe, uses AI and optimisation to explore those small changes. Its tools adjust pauses, shift attention between characters and find reaction shots.[1]
Imagine an announcement of a departure. Cutting immediately to the listener reveals their first response; holding on the speaker leaves the declaration occupying the frame. The words can stay identical while the emphasis changes.[1]
The prototype works within windows that preserve necessary sound–picture synchronisation. Editors can browse non-speaking reactions by character, emotion and shot type. It keeps the editor’s question at the scale of a glance or a pause.[1]
Sources & further reading
- Micro-Edits Cornell/Adobe · Accessed 29 Oct 2026
- Nhan Tran Author biography · Accessed 29 Oct 2026