Nodes/Network Bending for ComfyUI/Audio Reference Encoder
ComfyUI Node

Audio Reference Encoder

A ComfyUI node in audio/reference with 12 inputs and 2 outputs.

By DavidPiazza·Created about a year ago·Updated 8 months ago· 0
Audio Reference Encoder
  • vae
  • audio_a
  • audio_b
  • audio_c
  • audio_d
  • combined_latent
  • info
combination_modeattention
weight_a1.00
weight_b1.00
weight_c1.00
weight_d1.00
extract_segmentsfalse
segment_duration2.00
Categoryaudio/reference

Inputs (12)

NameTypeDefaultDescription
vaeVAEStable Audio VAE model
audio_aAUDIOFirst reference audio
combination_modeCOMBOattentionHow to combine multiple references
audio_boptAUDIOSecond reference audio
audio_coptAUDIOThird reference audio
audio_doptAUDIOFourth reference audio
weight_aoptFLOAT1.000–2
weight_boptFLOAT1.000–2
weight_coptFLOAT1.000–2
weight_doptFLOAT1.000–2
extract_segmentsoptBOOLEANfalseExtract most characteristic segments
segment_durationoptFLOAT2.000.5–10Duration of segments to extract (seconds)

Outputs (2)

NameTypeDescription
combined_latentLATENT
infoLATENT_INFO