Nine caption families across 45 presets
Versely's 45 caption presets are nine families of five. Map them by backing, position and word timing to shortlist three candidates in a couple of minutes.
Versely's 45 caption presets are nine families of five. Map them by backing, position and word timing to shortlist three candidates in a couple of minutes.
Text a video model draws inside the frame has to be right 200 times, not once. The size threshold where letterforms break, and the clean-plate rule.
A ten-prompt test set and a four-gate proofing loop for text-dense image renders, so garbled type gets caught before a client ever sees the file.
A guessed font name doesn't error — it just doesn't do what you asked. What each font category does to legibility at caption size on moving footage.
Prose can't tell a model where to put a headline — 'top-left' is a suggestion, not a coordinate. Structured JSON prompts can, and here's when that helps.
Generated graphics fail contrast because the model optimized for looking good, not for a ratio. The actual thresholds, the exemptions, and a two-minute check.
One variable font file replaces a shelf of static weight cuts — and its optical size axis does something no weight dropdown can ever reproduce.