24fps vs 30fps AI video: which output.fps to set on Sume

Neither 24 nor 30 fps is the right default for AI video. On Sume's Timeline 1.0, the safest choice is to leave output.fps unset, so the render runs at the rate your clips already have. Set 24, 25, 30 or 60 only when a delivery spec demands it, because any source at a different rate is then resampled by repeating or dropping frames.

What does Sume do if I set nothing?

Per the Timeline docs, an omitted output.fps renders at the rate the sources already run at. The longest video sources decide, stills have no rate, and 30 is used only when nothing in the timeline has one. Timeline compose behaves the same way by default, using the video layer's own frame rate.

Which values can I set?

Timeline accepts four output rates. What differs is the reason to pick each one.

What happens if my clips are 24 fps and I set 30?

The Timeline docs say a rate that differs from a source's is met by repeating or dropping a frame every few frames, which shows as judder on motion, and it is reported as output_fps_resamples_sources. As arithmetic, not a docs figure, going from 24 to 30 fps repeats about 6 frames every second, and going from 30 to 24 drops about 6.

So the question is rarely 24 versus 30 in the abstract. It is whether you want to change the rate your clips already have, and a change always costs some smoothness.

Which rate should a mixed set of clips use?

Pick the rate of the clips that carry the most screen time, since by default the longest video sources decide. Then replace or regenerate the odd clip out instead of resampling it. If a destination forces a rate, set it once on the final Timeline render rather than conforming each clip separately, so no clip is resampled twice.

Does the choice change the price?

The published Timeline rate is $0.10 per output minute, rounded up, and the docs state a single per-minute rate. The cost follows the length of the output, not the frame rate.

How do I know what my generated clips run at?

Sume's docs do not state one native frame rate for its video models, so do not assume. Run the free inspect probe on a generated clip (frames: false) and read the rate before you decide; the fields are in get a video's duration, resolution and fps. If two clips report different rates, expect the output_fps_resamples_sources warning described in video judder with mixed frame rates.

For the 23.976 case specifically, read 23.976 vs 24 fps.