SoundMorpher: Perceptually-Uniform Sound Morphing with Diffusion Model

Return index page

Enviromental Sound Morphing (AudioCaps)

This section provide additional demonstration of SoundMorpher on complex environmental sounds which sourced from AudioCaps dataset, each example pair are randomly chosen from AudioCaps dataset. In this experiment, we set N=10. The AudioCaps dataset are sourced from AudioCaps-git.


Example 1

Source audio

Spectrogram α=0.0

Target audio

Spectrogram α=0.25






α=0.0

Spectrogram α=0.0


Spectrogram α=0.25


Spectrogram α=0.5


Spectrogram α=0.75


Spectrogram α=1.0


Spectrogram α=1.0


Spectrogram α=1.0


Spectrogram α=1.0


Spectrogram α=1.0

α=1.0

Spectrogram α=1.0

Example 2

Source audio

Spectrogram α=0.0

Target audio

Spectrogram α=0.25






α=0.0

Spectrogram α=0.0


Spectrogram α=0.25


Spectrogram α=0.5


Spectrogram α=0.75


Spectrogram α=1.0


Spectrogram α=1.0


Spectrogram α=1.0


Spectrogram α=1.0


Spectrogram α=1.0

α=1.0

Spectrogram α=1.0

Example 3

Source audio

Spectrogram α=0.0

Target audio

Spectrogram α=0.25






α=0.0

Spectrogram α=0.0


Spectrogram α=0.25


Spectrogram α=0.5


Spectrogram α=0.75


Spectrogram α=1.0


Spectrogram α=1.0


Spectrogram α=1.0


Spectrogram α=1.0


Spectrogram α=1.0

α=1.0

Spectrogram α=1.0
Return index page