By default, the base class might try to infer PTS from DTS and vice versa.
While this is generally correct for audio data, it may not be otherwise. Sub-classes implementing such formats should disable timestamp inferring.
| this |
a Parse |
| infer_ts |
|