Try Pro FREE for 2 weeksGet Started

video-edl-planner — Traditional AI Model

Video EDL Planner searches every legal cut position and every candidate 9:16 window inside a take, then keeps the one where the subject is genuinely visible — scored at delivery resolution, not guessed.

Details

  • CategoryTraditional
  • Year-1
  • License

Inputs & Outputs

TakesInputsegments

Candidate takes to plan inside — from Video Analysis, or any other source. Which moments to use is decided here, not by the planner.

ConstraintsInputstringoptional

Optional Video Constraint Map output. Without it every shot is centred, because a window position chosen with nothing measured is a guess.

PlanOutputstring

An EDL document for the Video EDL node: the chosen cut, window and camera move per take, with what each scored

Parameters

target_len_secfloatdefault 45range 15–120

How long the finished short may run. The plan comes in under this, and falls short of it when the takes cannot supply the footage.

Upper bound on the finished length in seconds. Together with tempo it sets the cut count: roughly length divided by the tempo's shot length. When the takes hold less footage than that, the plan keeps the tempo and comes out shorter, rather than stretching shots to fill the time.

tempostringdefault A little fasteroptions: Original, A little faster, Much faster

How fast the short cuts. This changes the number of cuts as well as their length.

Cutting rhythm. Measured from two reference shorts at opposite ends: a mukbang one that alternates (17 cuts in 55s, median shot 2.7s) and a pet one that dwells (median 6.7s). 'Original' dwells, 'Much faster' alternates, 'A little faster' (default) sits between. A faster tempo also changes the camera: fast plans accent each cut with a punch, slow ones glide.

emphasisstringdefault

Where the interesting part is, e.g. "0:55~3:20". Takes inside it get more of the running time.

Optional time range weighting, written the way a person would: "0:55~3:20", "55~200", or several separated by commas. Takes falling inside get a larger share of the length; a take straddling the edge lands in between. Anything unreadable is ignored rather than guessed at.

step_secfloatdefault 0.25range 0.1–2advanced

How finely candidate cut positions are searched inside each take. Smaller finds better cuts and costs more arithmetic.

Spacing of candidate start times within a take. Every candidate is scored at every window position, so halving this doubles the search. 0.25s is what the reference optimiser used.

Tags

  • video
  • shorts
  • edl
  • planner
  • segments

Alternatives in Video Editing

  • Audio Edit

    Trim a sound, fit it to a video or a length, dip or mute parts, and fade it.

  • Audio Finish

    Normalize a video's loudness to a broadcast target (2-pass loudnorm + peak limiter).

  • Audio Mix

    Mix narration, music and sound effects into one track, with ducking and loudness.

  • Audio Replace / Mix

    Replace or mix a clip's audio with a new sound, keeping the video as is.

  • Chroma Compose

    Cut a green-screen person out of one video and place them over another.

  • Hook Teaser (Cold Open)

    Prepend a cold-open teaser of each clip's hook moment to the front.

  • Ken Burns (Still to Video)

    Animate a still image into a video with a moving crop window.

  • Narration Mux (Cold Open)

    Prepend a voiceover cold-open (frozen first frame + narration) to a clip.

All AI models