Katechon ArcHow board-to-video works
- Approve a frame for each shot.
- Generate the shot's clip with the frame as its first image plus the shot's action, line and length.
- Optionally give the next shot's frame as the last image, so consecutive clips cut together cleanly.
- Review the clip; if the motion is wrong, regenerate the clip, not the board.
What we learned on BROODFALL
BROODFALL was made in storyboard mode with grid storyboards: each group of four shots is one 2x2 grid where a cell is the first frame of one shot and the last frame of the one before. We rendered 972 clips on Seedance 2.0 Fast at 720p and cut 364 of them into a 10:09 film.
- 3x3 grids drifted off their cut lines, so cells came out cropped wrongly; 2x2 grids gave clean cells.
- Board problems carry into video. A face that read like a famous actor on the board got about 5% of that character's shots blocked at the video stage.
- Grids speed up review: four shots on one image show whether a sequence flows.
Board-to-video versus reference-to-video
Starting from a board fixes composition exactly but needs a good frame per shot. Reference-to-video attaches character and location sheets and lets the model compose the shot, which is faster when composition is less critical. Most of our dialogue-led films used references; our action film used boards. See which mode to use.
Cost per shot in Arc: the frame (13 credits, or a quarter of a grid) plus the clip, for example 181 credits for 8 seconds on Seedance 2.0 Fast at 720p.
Where Katechon Arc fits
Arc's storyboard mode sends approved frames, including chained grid cells, straight to video generation in the same project.
If your storyboard is a hand-drawn sketch, a model will not turn it into realistic footage directly; redraw it as a generated frame with your references first.