Robustness spectral tilt

Overall scores

	CNN14	w2v2-b	hubert-b	axlstm
Overall Score	25.0% passed tests (2 passed / 6 failed).	87.5% passed tests (7 passed / 1 failed).	87.5% passed tests (7 passed / 1 failed).	50.0% passed tests (4 passed / 4 failed).

Change Ccc Downward Tilt

Threshold: -0.05

Data	Change CCC Downward Tilt
	CNN14	w2v2-b	hubert-b	axlstm
iemocap-2.3.0-emotion.dimensions.test.gold_standard	-0.09	-0.01	-0.00	-0.04
msppodcast-2.6.1-emotion.dimensions.test-1.gold_standard	-0.03	0.00	-0.00	0.01
mean	-0.06	-0.01	0.00	-0.01

Change Ccc Upward Tilt

Threshold: -0.05

Data	Change CCC Upward Tilt
	CNN14	w2v2-b	hubert-b	axlstm
iemocap-2.3.0-emotion.dimensions.test.gold_standard	0.12	-0.02	-0.02	0.02
msppodcast-2.6.1-emotion.dimensions.test-1.gold_standard	-0.12	-0.02	-0.02	-0.10
mean	0.00	-0.02	-0.02	-0.04

Percentage Unchanged Predictions Downward Tilt

Threshold: 0.8

Data	Percentage Unchanged Predictions Downward Tilt
	CNN14	w2v2-b	hubert-b	axlstm
iemocap-2.3.0-emotion.dimensions.test.gold_standard	0.47	0.99	0.98	0.92
msppodcast-2.6.1-emotion.dimensions.test-1.gold_standard	0.63	1.00	1.00	0.79
mean	0.55	0.99	0.99	0.85

Percentage Unchanged Predictions Upward Tilt

Threshold: 0.8

Data	Percentage Unchanged Predictions Upward Tilt
	CNN14	w2v2-b	hubert-b	axlstm
iemocap-2.3.0-emotion.dimensions.test.gold_standard	0.41	0.75	0.77	0.50
msppodcast-2.6.1-emotion.dimensions.test-1.gold_standard	0.42	0.90	0.85	0.39
mean	0.41	0.82	0.81	0.45

Visualization Downward Tilt

Difference of predictions for original audio and audio with a downward spectral tilt. 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.

CNN14	w2v2-b	hubert-b	axlstm