This project aims to examine the possibility of utilising the tone-transfer application from the "Differentiable Digital Signal Processing (DDSP)" function as a sound effect design tool for music, visual contents and many other art forms.

The process of this work:

1. Choose a video clip convenient for the sound design including dynamic movements of an object, streams and bubbles of water and liquid.
2. Train tone models out of numerous sound effects from commercial sound libraries such as "whoosh", "water stream, "chimes", etc.
3. Record "voice performance" while watching the video clip.
4. Cross-apply those models and sound sources through the tone-transfer application.
5. Build sound effects that match the video clip using a Digital Audio Workstation. The sound design was only carried out by volume control and editing, without applying any plug-ins or special audio effects.

свернуть