Robustness low quality phone¶
w2v2-L |
hubert-L |
wavlm |
data2vec |
|
|---|---|---|---|---|
Overall Score |
25.0% passed tests (1 passed / 3 failed). |
100.0% passed tests (4 passed / 0 failed). |
100.0% passed tests (4 passed / 0 failed). |
50.0% passed tests (2 passed / 2 failed). |
Change Ccc Low Quality Phone¶
Data |
Change CCC Low Quality Phone |
|||
|---|---|---|---|---|
w2v2-L |
hubert-L |
wavlm |
data2vec |
|
iemocap-2.3.0-emotion.dimensions.test.gold_standard |
-0.06 |
-0.04 |
0.01 |
-0.02 |
msppodcast-2.6.1-emotion.dimensions.test-1.gold_standard |
-0.02 |
-0.03 |
-0.01 |
-0.05 |
mean |
-0.04 |
-0.04 |
0.00 |
-0.04 |
Percentage Unchanged Predictions Low Quality Phone¶
Data |
Percentage Unchanged Predictions Low Quality Phone |
|||
|---|---|---|---|---|
w2v2-L |
hubert-L |
wavlm |
data2vec |
|
iemocap-2.3.0-emotion.dimensions.test.gold_standard |
0.41 |
0.71 |
0.72 |
0.54 |
msppodcast-2.6.1-emotion.dimensions.test-1.gold_standard |
0.38 |
0.72 |
0.82 |
0.43 |
mean |
0.40 |
0.71 |
0.77 |
0.48 |
Visualization¶
Difference of predictions for original audio and low quality phone audio. The allowed prediction difference \(\delta < 0.05\) is highlighted in green in the upper plot. The lower plot shows the distributions of the two predictions.
w2v2-L |
hubert-L |
wavlm |
data2vec |
|---|---|---|---|
|
|
|
|
|
|
|
|