Audio Samples: Singing Voice Effect Estimation
Results on seen speakers|
① Autoencoding: graph decoder with another graph encoder. ② Unconditioned: graph decoder with dummy zero latents. ③ Estimation (proposed: token, 2-stage): token-by-token decoding + 2-stage (categorical/continuous) decoding. ④ Node, 2-stage: node-by-node decoding + 2-stage decoding. ⑤ Token, 1-stage: token-by-token decoding + single-stage autoregressive decoding. ⑥ Oracle source: the proposed method ③ with dry source conditioned reference encoder. Dry source: (sum of) dry source(s) without any processing. |
| Sample #1 | ||
|---|---|---|
Ground-truth full |
||
|
① Autoencoding prototype full |
② Unconditioned prototype full |
Dry source (bypass graph) |
③ Estimation (proposed: token, 2-stage) full |
||
|
④ Node, 2-stage prototype full |
⑤ Token, 1-stage prototype full |
⑥ Oracle source prototype full |
| Sample #2 | ||
|---|---|---|
Ground-truth full |
||
|
① Autoencoding prototype full |
② Unconditioned prototype full |
Dry source (bypass graph) |
③ Estimation (proposed: token, 2-stage) full |
||
|
④ Node, 2-stage prototype full |
⑤ Token, 1-stage prototype full |
⑥ Oracle source prototype full |
| Sample #3 | ||
|---|---|---|
Ground-truth full |
||
|
① Autoencoding prototype full |
② Unconditioned prototype full |
Dry source (bypass graph) |
③ Estimation (proposed: token, 2-stage) full |
||
|
④ Node, 2-stage prototype full |
⑤ Token, 1-stage prototype full |
⑥ Oracle source prototype full |
| Sample #4 | ||
|---|---|---|
Ground-truth full |
||
|
① Autoencoding prototype full |
② Unconditioned prototype full |
Dry source (bypass graph) |
③ Estimation (proposed: token, 2-stage) full |
||
|
④ Node, 2-stage prototype full |
⑤ Token, 1-stage prototype full |
⑥ Oracle source prototype full |
| Sample #5 | ||
|---|---|---|
Ground-truth full |
||
|
① Autoencoding prototype full |
② Unconditioned prototype full |
Dry source (bypass graph) |
③ Estimation (proposed: token, 2-stage) full |
||
|
④ Node, 2-stage prototype full |
⑤ Token, 1-stage prototype full |
⑥ Oracle source prototype full |
| Sample #6 | ||
|---|---|---|
Ground-truth full |
||
|
① Autoencoding prototype full |
② Unconditioned prototype full |
Dry source (bypass graph) |
③ Estimation (proposed: token, 2-stage) full |
||
|
④ Node, 2-stage prototype full |
⑤ Token, 1-stage prototype full |
⑥ Oracle source prototype full |
| Sample #7 | ||
|---|---|---|
Ground-truth full |
||
|
① Autoencoding prototype full |
② Unconditioned prototype full |
Dry source (bypass graph) |
③ Estimation (proposed: token, 2-stage) full |
||
|
④ Node, 2-stage prototype full |
⑤ Token, 1-stage prototype full |
⑥ Oracle source prototype full |
| Sample #8 | ||
|---|---|---|
Ground-truth full |
||
|
① Autoencoding prototype full |
② Unconditioned prototype full |
Dry source (bypass graph) |
③ Estimation (proposed: token, 2-stage) full |
||
|
④ Node, 2-stage prototype full |
⑤ Token, 1-stage prototype full |
⑥ Oracle source prototype full |
| Sample #9 | ||
|---|---|---|
Ground-truth full |
||
|
① Autoencoding prototype full |
② Unconditioned prototype full |
Dry source (bypass graph) |
③ Estimation (proposed: token, 2-stage) full |
||
|
④ Node, 2-stage prototype full |
⑤ Token, 1-stage prototype full |
⑥ Oracle source prototype full |
| Sample #10 | ||
|---|---|---|
Ground-truth full |
||
|
① Autoencoding prototype full |
② Unconditioned prototype full |
Dry source (bypass graph) |
③ Estimation (proposed: token, 2-stage) full |
||
|
④ Node, 2-stage prototype full |
⑤ Token, 1-stage prototype full |
⑥ Oracle source prototype full |