Video & audio

Knowledge short video

Landscape 1920×1080 knowledge video, from topic check to voiceover, cards and cover.

Paid $15 list, $12 until 2026-12-31 · one skill, yours for good. The outline, the steps and the file open after purchase.

Background

A film's quality is composed of topic 60% · footage 30% · technique 10%. The process is fixed, the metrics are locked, and what remains is judgment.

Why it is built this way

The stages are ordered so nothing gets authored before its evidence exists: footage, script, voiceover, cards, composite. The failure this blocks is writing the script first and filling every cell with animation at the end, which yields a film made entirely of self-drawn illustration, with nothing a viewer can pause and verify. Every claim is listed with the real image it needs, sourced where possible and deleted where not, and the real-footage share is counted by shot against a 60 percent floor.

Speech rate is written as an outcome, not a parameter, because a TTS rate setting and the measured characters per second are different numbers: generate, measure with ffprobe, compute, adjust and re-run, since guessing from the formula has drifted 30 percent. Acceptance follows the same logic. Every technical metric can be green while the film is empty, so each shot is asked whether it holds something the editor did not draw. The topic gate keeps two vetoes, participatory and clickable, that the six-item score cannot override.

Where it bites

• Animation filling every cell Twelve of thirteen shots procedural, every technical metric green, and the delivery judged empty. Question 0 exists for this: list the real image each claim needs first, report the procedural share in the notes file, and delete a claim rather than cover it with an invented illustration. Changing the background colour is not changing the content.
• Guessing the speaking rate A TTS rate parameter and the actual characters per second can differ by 30 percent, so a script timed from the formula lands with the voiceover either rushing or short. The order is fixed: generate the audio, measure with ffprobe, compute characters per second, adjust and re-run, with 4.9 as the target.
• One 16:9 cover A single 16:9 cover misses the delivery spec, which asks for a 4:3 landscape cover at 1920×1440 and a 3:4 portrait cover at 1080×1440 built from one visual system. Stretching a landscape frame into the portrait version goes soft, so the portrait cover needs portrait footage.
• An even run of long shots Six long shots across 60 seconds read slower than twelve, and adjacent shots sharing one motion flatten the cut. The storyboard rules ask for two-shot beats of 3 to 6 seconds, never over 7, motion families alternating by shot, and static frames held back for the payoff or key line.

When to use it

Complete pipeline for a landscape (1920×1080) knowledge short video: topic decision → footage sourcing → talking-head script → Doubao voiceover → graphic cards → composite → dual-aspect-ratio covers + publish copy. Use when the user wants a "knowledge short video", "a one-minute landscape explainer", "turn a hot topic into a 1-minute video", "a Douyin knowledge short", or asks for the level of the pipeline's existing reference cuts. Not for pure editing, pure voiceover, vertical talking-head video (that is koubo-video), or WeChat Official Account / Xiaohongshu image posts. Question 0 before work (hard gate set by the user): for every claim, list the real footage first and go find each item; if you cannot find it, delete the sentence or change the claim — illustrations and pure procedural cards are forbidden as substitutes; write the script only after the footage is locked; animation does transitions and number emphasis only, never the lead; besides the technical metrics, acceptance must answer "are these few seconds worth watching".