One-line prompt vs structured prompt — same subject, tested — HyperFrames (the finale: a vague brief films vague)

<direction> THE ONE-LINER (first half, rendered literally): "Make a showreel." That is the whole prompt. The half is deliberately plain: the line fades in, a small caption admits that is all there was — NO DIRECTION · NO STRUCTURE · NO BEATS — and nothing else happens. THE STRUCTURED PROMPT (second half): an 8-second landscape showreel for a showreel. Kicker: ONE PROMPT — ONE SHOWREEL. Headline MAKE IT MOVE. slams in with a hard settle. Three chips land in stagger: HOOK / BUILD / PAYOFF. End line: STRUCTURE IS THE DIFFERENCE, in accent #ff5c1a. House style: near-black #0a0a0a background, white type, ONE orange accent #ff5c1a, Arial/Helvetica bold, no webfonts. Fictional brands/data only. Banned (structured half): glows, gradients, particle bursts, bouncy cartoon easing, a headline that does not land on a beat. </direction> <structure> One-liner half (8s): label ONE-LINE PROMPT, the line, the caption, hold. Structured half (8s): label STRUCTURED PROMPT, headline, chips, end line. Concatenated one-liner first, structured second, with ffmpeg. </structure> <build> Two HyperFrames + GSAP compositions, 1280x720 landscape, 30fps. The one-liner half uses only fade-ins, on purpose — dressing it up would falsify the test. The structured half uses a seek-driven timeline with a headline slam and staggered chips. </build> <gotchas> - Do not "improve" the one-liner half. Its plainness IS the result. - Burn the labels INTO each half (ONE-LINE PROMPT / STRUCTURED PROMPT); without them the pair reads as one confused film. </gotchas> <methods> Methods note: two HyperFrames compositions from the two prompts, concatenated one-liner first, structured second, each labelled. Final render times: one-liner 26.7s + structured 10.2s. Iterations: 1 per half, 2 total. </methods> Before you finish: step through the timeline at five points and fix anything that overlaps, clips the frame, or drifts off the beat grid.