Table 1.
Categories of emotional speech databases, their features, and some examples of each category. A black dot (•) means having a feature.
| Simulated | Semi-Natural | Natural | |
|---|---|---|---|
| Description | created by trained speakers reading the same text with different emotions | made by asking people or actors to read a scenario containing various emotions | extracted from TV shows, YouTube videos, call centers, etc. |
| Natural emotions | • | • | |
| Contains contextual information | • | • | |
| Contains situational information | • | • | |
| Discrete and separable emotions | • | ||
| Single emotions at a time | • | • | |
| Widely used | • | ||
| Standardized | • | ||
| Easy to model | • | ||
| Inter corpora results are comparable | • | ||
| Large Number of emotions | • | • | |
| Used in real-world emotion systems modeling | • | ||
| Controlled privacy and copyright | • | • | |
| Examples |