Sample output of our model (STFT-SEANet) on real historical music recordings

Input

STFT-SEANet: Reconstruction loss only (better)

STFT-SEANet: Adversarial + recon. loss (worse)

0 0

1 1

2 2

3 3

4 4

5 5

6 6

7 7

8 8

9 9

10 10

11 11

12 12

13 13

14 14

15 15

16 16

17 17

18 18

19 19