Prompt driven and audio driven, told apart
Some tools generate a scene from a written description. Others animate a person from a supplied audio track. Both publish lengths in seconds, and the figures describe completely different quantities, so ranking them together is a category error. As of 2026-09-12.
| Kind of tool | What sets the length | What the ceiling describes |
|---|---|---|
| Prompt driven | A duration parameter | How long a scene holds together |
| Audio driven | How long the supplied track runs | A bound on file handling |
| Beat driven | A grain repeated inside one request | How many beats one job returns |
Inclusion rule. Kinds of generation tool these sheets read, separated by what decides the length of the output. No ordering is implied between the resulting figures. Order. The most common kind first.
1Minutes and seconds are not two grades of the same thing
A prompt-driven ceiling says how long a generated scene can be held together before it falls apart. An audio-driven ceiling says how long a file the service will handle. Sorting them into one order puts a handling bound above a coherence bound and calls it capability.
That is why a length column on these sheets is never ranked. The mechanism is printed beside the figure so a reader can see which quantity is being measured.
2The choice is made in the brief, not in the column
An audio-driven tool is the right answer when the advertisement is a person speaking. A prompt-driven tool is the right answer when it is a scene playing. Neither is a better version of the other.
Once that is settled, the length column becomes useful again, because the comparison is then between figures that mean the same thing. Before it is settled the column is actively misleading.
3Beat-driven output is a third case
Where one request returns a long piece already cut into fixed beats, the length is neither a coherence bound nor a handling bound. It is a count of beats, and the cuts inside it were placed by the model.
That makes it the most favourable-looking figure on a length column and one of the most constrained in practice, because the beat structure is fixed and the cut points cannot be moved.
A vocabulary entry: the definition, plus how the specification pages on these sheets usually publish it. Table checked 2026-09-12. Nearby entries: IAA and IAP, First and last frame.
4Making the ad itself
For ad work around “Prompt and audio driven”, SceneMixer's pages explain how its ad projects are produced:
Measured across every job that finished successfully on scenemixer.com in the thirty days to 20 September 2026: analysis of a whole book runs about 2 minutes 49 seconds at the median, one episode's storyboard 1 minute 54 seconds, one video segment 3 minutes 20 seconds, and compositing an episode 3 minutes 52 seconds.
SceneMixer, Turn a Novel into a Video Series with AI — “How long does it take to produce one episode?” (read 2026-10-13)
The final cut ends on an end card: your main image plus four lines — Brand, Product name, Tagline, Call to action. The first two are read from your photo and your brief; the tagline and call to action are written with the script.
SceneMixer, The End Card (read 2026-10-13)