Same audio file name for different speakers

[unrea1-sama]

Is your feature request related to a problem? Please describe. Right now MFA requires different audios from different speakers have different file names. For example, this requires a corpus directory structure like:

+-- prosodylab_corpus_directory
|   +-- speaker1
|       --- recording1.wav
|       --- recording1.lab
|       --- recording2.wav
|       --- recording2.lab
|   +-- speaker2
|       --- recording3.wav
|       --- recording3.lab
|   --- ...

When different speakers have different audios but with a same file name, these audio will not be correctly aligned. That is if we have speaker1/recording1.wav and speaker2/recording1.wav, MFA will produce wrong alignment for them. And we need to rename audios to avoid this name collision, which is actually not needed because each audio file can be determined by its path. Describe the solution you'd like Could you please add a new feature to support that different speakers have different audios with a same file name? Describe alternatives you've considered No.

Additional context No.

[mmcauliffe]

Are you on the latest version of MFA? I'm not able to replicate this, and each file has a unique integer ID internally now (specifically to get around this and other issues with using the file name as a unique identifier), so it really shouldn't be happening.

[unrea1-sama]

I'm now using mfa version 2.0.1. I will update it and see if this still happens again. Thankyou!