Open paul-krug opened 4 years ago
I agree, negative values make sense. Especially if there is no silence before the recording (i.e., the sample was taken from continuous speech).
Ok, so I will fix the NaN issue and will allow negative values.
Conceptually they make sense. But negative values will cause issues with the data export to gestural score files, which would be incompatible with VocalTractLab as of now. I'd say do not allow negative values for now, but keep it on the list of future improvements once we figure out how to deal with the related downstream issues.
Ok it is fixed. For now only positive numbers are allowed in the table. I will leave the issue open but changed the label from bug to feature request.
Boundaries can be set to negative values or nan, when manipulating the boundary table. Nan is a bug for sure, but negative values might be a feature.