Support audio duration mismatch tolerance in MixedCut.load_audio()
#1054
Conversation
| return mixture | ||
|
|
||
|
|
||
| def _create_mixture( |
There was a problem hiding this comment.
Is there a reason why this function is moved outside the class? The algorithm is specific to this particular simulator.
There was a problem hiding this comment.
Makes sense, I changed this back and used the same "trick" as in the simulator.
There was a problem hiding this comment.
Thanks. Could you do the same for _simulate_worker (i.e. make it a function of the specific class)? The logic for those is also tied to that algorithm.
There was a problem hiding this comment.
That would complicate the workarounds for parallelization; anyway each simulation method has a separate file, so I think it should be OK and actually consistent with other functions in lhotse that are being parallelized this way, e.g. _X_single() functions at the end of lhotse/cut/set.py.
No description provided.