search

BUTSpeechFIT / EEND

72 stars 9 forks source link

How to generate the utt2spk file and utt2spk file？ #14

Open skxooo opened 3 hours ago

skxooo commented 3 hours ago

I understand the general purpose of these two files, but I’m still a bit uncertain about how to generate them correctly. For instance, when creating an utt2spk file, does uttid correspond to the filename? And what should the corresponding spkid be? For different speakers, would it be appropriate to use identifiers like spk00, spk01, and spk02? For the second utterance, can I continue using spk00, spk01, and spk02, or would it be better to use new identifiers, starting from spk03?

Additionally, I have a question regarding the spk2utt file. Should speaker IDs be unique within each utterance, or do they need to be unique across the entire dataset?

Thank you very much for your time and assistance—it’s greatly appreciated!

skxooo commented 3 hours ago

May I ask an additional question: does each utterance correspond to a separate WAV file? Thank you, and I look forward to your response.