ComfyUI Node

H3 Extras

The canvas from an h3_inputs wire, as plain numbers: width, height, the requested length and the frame count it snaps up to, the video and audio latent lengths, and the real duration in seconds.

By Nynxz·Created 14 days ago·Updated 5 days ago· 4
H3 Extras
  • h3_inputs
  • width
  • height
  • length
  • frame_count
  • latent_t
  • audio_t
  • seconds
  • task
  • is_keyframes
  • is_references
  • summary
modeauto
CategoryNynxz/H3

Inputs (2)

NameTypeDefaultDescription
h3_inputsNYNXZH3_INPUTFrom H3 Inputs, or anywhere along the chain.
modeCOMBOautoMatch whatever H3 Encode is set to, so the task reported here is the task that will actually run.

Outputs (11)

NameTypeDescription
widthINT
heightINT
lengthINTThe length the canvas asked for.
frame_countINTWhat it snaps up to on the model's 17k+5 grid — the frames you actually get.
latent_tINTVideo latent frames.
audio_tINTAudio latent frames (40 Hz).
secondsFLOATframe_count / 24.
taskSTRINGt2va, fl2va or ref2va — what the same list will make H3 Encode do.
is_keyframesBOOLEANTrue for fl2va — a first and/or last frame is driving the clip. Wire into a switch to branch a graph on it.
is_referencesBOOLEANTrue for ref2va — <Picture i> / <Video k> / <Audio j> references.
summarySTRINGOne line naming all of it, for a text preview.