Delivery primer for AI-made episodes
The primer covers reading a resolution claim, choosing a frame rate, what an HDR master contains, and building an episode runtime out of short clips. As of 2026-09-25.
Every primer cites a standards or platform page first and then the vendor figures that bear on it.
Nothing in the primer measures a file; it reads what was printed.
1Pages on this sheet
- Checking a download against the spec — A printed figure is a promise; the downloaded file is the fact.
- Frame rates for broadcast and streaming — A broadcast master runs at the rate of the region it airs in: 25 or 50 in most of Europe, 23.98, 29.97 or 59.94 in North America.
- From clips to an episode runtime — A 22-minute episode at 15 seconds a clip is about ninety clips.
- Keeping a spec sheet per episode — A spec sheet per episode records, for every clip, the tool, model, resolution, frame rate, shape, length and any mark.
- Landscape masters and portrait cut-downs — A landscape master is usually cut down to a portrait version for phone feeds.
- Letterboxing 21:9 inside 16:9 — A 21:9 picture delivered for 16:9 screens is letterboxed: black bars above and below, the full wide frame in the middle.
- Mixed-resolution timelines — An episode built from several generators, or several tiers of one, mixes resolutions: 720p clips beside 1080p and upscaled 4K.
- Reading a resolution claim — A resolution figure on a vendor page answers one of three questions: what size the model generates, what size a later step scales it to, or what size an editor exports.
- SDR-to-HDR conversion and native HDR — Native HDR is generated with a wide dynamic range from the start.
- What a streaming upload spec asks for — YouTube's upload pages are a public example of a platform spec: progressive scan, the recorded frame rate, recommended bitrates by resolution, and for HDR, Rec. 2020 with PQ or HLG at 10 or 12 bits.
- What an HDR deliverable contains — An HDR deliverable names four things: a colour space, usually BT.2020; a transfer function, PQ or HLG; a bit depth of 10 or more; and metadata describing the grade.
2How one pipeline outputs it
For a delivery question like “Delivery primer”, SceneMixer states the following about what it produces:
Voices and ambience are in the clip, and in the composited episode, with the episode’s score unless you turned the Music switch off.
SceneMixer, Dialogue and Sound in One Generation — “Does the downloaded file include the sound?” (read 2026-10-13)
SceneMixer turns a novel into a watchable short-drama series in one pipeline: upload the novel (Chinese or English) → AI extracts characters, scenes, and an episode-by-episode structure → generates consistent character and scene art → writes storyboard scripts → renders video segment by segment (16:9 by default, 9:16 in one setting) → composites full episodes.
SceneMixer, Turn a Novel into a Video Series with AI (read 2026-10-13)