Table 2.
Statistics of the reviewed databases and their types of emotion. Availability of an emotion in a database is represented with a black dot (•) in the corresponding cell. The light gray section in the table shows the emotions shared between all the databases, and the darker gray section at the end of the table shows the 3 dimensions of continuous dimensional emotion representation.
| Berlin Database of Emotional Speech (EMO-DB) | Danish Emotional Speech Database (DES) | The Ryerson Audio-Visual Database of Emotional Speech and Song (RAVDESS) | Toronto Emotional Speech Set (TESS) | Crowd-sourced Emotional Multimodal Actors Dataset (CREMA-D) | Interactive Emotional Dyadic Motion Capture Database (IEMOCAP) | Vera am Mittag Database (VAM) | |
|---|---|---|---|---|---|---|---|
| Number of emotions | 7 | 5 | 8 | 7 | 6 | 9 emotions, 3 dimensions | 3 dimensions |
| Number of samples | 700 | 210 | 2496 | 2800 | 7442 | 1150 | 1018 |
| Number of Speakers | 10 | 4 | 24 | 2 | 91 | 10 | 47 |
| Average Length | 2.8 s | 2.7 s | 3.7 s | 2.1 s | 2.5 s | 5 m | 3.0 s |
| Anger | • | • | • | • | • | • | |
| Happiness | • | • | • | • | • | • | |
| Sadness | • | • | • | • | • | • | |
| Neutral | • | • | • | • | • | • | |
| Surprise | • | • | • | • | |||
| Fear | • | • | • | • | • | ||
| Disgust | • | • | • | • | • | ||
| Boredom | • | ||||||
| Calm | • | ||||||
| Frustration | • | ||||||
| Excited | • | ||||||
| Valence | • | • | |||||
| Activation | • | • | |||||
| Dominance | • | • |