A Deep Learning Approach to Intelligent Drum Mixing with the Wave-U-Net

Audio examples for the paper:

Martínez Ramírez M. A., Stoller, D. and Moffat, D., “A Deep Learning Approach to Intelligent Drum Mixing with the Wave-U-Net” submitted to the Journal of the Audio Engineering Society, May 2020.

View the source code.

Phrase - afro: simple slow sticks

- Reference Wet
- Wave-U-Net Wet
- Random Forest Wet
- Reference Dry
- Wave-U-Net Dry
- Random Forest Dry
- Anchor

 

Phrase - blues: simple slow brushes

- Reference Wet
- Wave-U-Net Wet
- Random Forest Wet
- Reference Dry
- Wave-U-Net Dry
- Random Forest Dry
- Anchor

 

Phrase - reggae: simple slow sticks

- Reference Wet
- Wave-U-Net Wet
- Random Forest Wet
- Reference Dry
- Wave-U-Net Dry
- Random Forest Dry
- Anchor

 

Phrase - oriental: simple slow sticks

- Reference Wet
- Wave-U-Net Wet
- Random Forest Wet
- Reference Dry
- Wave-U-Net Dry
- Random Forest Dry
- Anchor

 

Phrase - salsa: complex medium sticks

- Reference Wet
- Wave-U-Net Wet
- Random Forest Wet
- Reference Dry
- Wave-U-Net Dry
- Random Forest Dry
- Anchor

 

Phrase - waltz: complex slow brushes

- Reference Wet
- Wave-U-Net Wet
- Random Forest Wet
- Reference Dry
- Wave-U-Net Dry
- Random Forest Dry
- Anchor

 

Accompaniment - celtic rock: brushes

- Reference Wet
- Wave-U-Net Wet
- Random Forest Wet
- Reference Dry
- Wave-U-Net Dry
- Random Forest Dry
- Anchor

 

Accompaniment - rock: sticks

- Reference Wet
- Wave-U-Net Wet
- Random Forest Wet
- Reference Dry
- Wave-U-Net Dry
- Random Forest Dry
- Anchor

 

Solo - toms: mallets

- Reference Wet
- Wave-U-Net Wet
- Random Forest Wet
- Reference Dry
- Wave-U-Net Dry
- Random Forest Dry
- Anchor