How to make the next episode of a cinematic mystery series
Give Vizard Agent the previous episode and the script for the next one. Vizard Agent studies episode one's frames, voice, structure and sound design, matches the narrator, re-records with written pauses so the pacing carries suspense, and generates the new scenes in the established look.
What do you type into Vizard Agent?
Say it continues directly from the last one. Vizard Agent treats that as an instruction to go and analyse the previous episode rather than to start fresh, and everything downstream, from the voice search and the grade to the caption style and the sound design, is chosen to match what it finds there.
Prompt
What does Vizard Agent actually do?
A serialised mystery only works if episode two looks and sounds like episode one: the same voice, the same grade, the same pacing, and a cliffhanger that lands the way the last one did. So most of the early work is study rather than generation. In the session behind this article it:
- Probed episode one's specs, reviewed a montage of its frames for the visual style and characters, and analysed its audio for the voice, tone and sound design.
- Analysed episode one's structure, styling and transcript, then searched for a voice matching the original narrator and generated the new voiceover.
- Found the read too flat, rewrote the script with dramatic pauses and recorded it again, then transcribed it for precise timings.
- Generated twelve scenes in batches, including the hallway, the warning letter, the secret room and the cliffhanger, and reviewed them all as a grid.
- Measured every audio stem in LUFS, built the badge and closing card inside the vertical safe zones, and rebuilt the master after refining the badges.
- Delivered 51.5 seconds at 1080x1920, H.264, 30fps, against a 45 to 55 second brief, with the picture timed to the read.
When does this not work well?
Serialised fiction is the hardest kind of video to keep consistent, because every new episode is another chance for the look and the voice to drift. Vizard Agent matches deliberately against the previous episode, so drift is managed rather than eliminated. Voice matching is approximate, and fifty seconds holds one revelation, not three.
Common questions
Do I need to send the previous episode? Yes, and it does most of the work. Vizard Agent reads the style, voice and structure from it.
Will the narrator sound the same? Vizard Agent searches for the closest match against episode one's audio. It is close rather than identical.
Why write pauses into the script? Because a voice engine reads evenly by default. The pauses are what make it sound like suspense rather than narration.