Experiments

AudioX

2025
FINISHED - PASSEDyesAudio Gen

AI audio-to-video and text-to-audio - found a viable replacement for my ComfyUI audio workflow.

Audio GenVideo GenComfyUI
GitHub

Aim

Goal is to test AudioX.

Plan

Just test this out and see where all it can be applied. Given that I'm running AI Influencers, it might be useful for generating audio for those images.

Implementation

Features -

  1. text to audio
  2. add audio to given video
  3. add music to given video

text to audio →

  • can be a one-stop destination for creating audios for your videos and shorts.
  • generates good and clear sounds of people clapping, laughing, crying, etc.
  • not good for voice
  • bg music is good (but almost always messes up when voice is included in the prompt)

add audio to video →

  • good for adding sounds to your animations.
  • can add audio of - typewriter typing, car drifting, waves on beach, wind blowing in forest, etc.

add music to video →

  • good for adding bg music. since the audio it makes is good (sometimes a little crackly) and it merges the audio with the video, a good replacement for my ComfyUI workflow. (need to test more)
  • it can understand the mood of the video and then give bg song, so -
    • for a beach video it was smooth and calm
    • for a car racing video, it had more progressive beats
    • the below video is NatGeo documentary styled.

Next experiment

N

Nari AI

FINISHED - PASSEDAudio Gen