Is there a functional reason for SOFA to read ".lab" transcriptions instead of ".txt"? #10
-
|
I wonder if it would be worth changing, as most users won't use .lab for transcriptions, just .txt. Other forced aligners function similarly using .txt. |
Beta Was this translation helpful? Give feedback.
Replies: 1 comment 5 replies
-
|
For Chinese users, .lab files are the preferred format for transcriptions instead of .txt files. This is because the Montreal Forced Aligner (MFA) utilizes the .lab format, and consequently, nearly all tools in the DiffSinger dataset ecosystem, including LyricFA and minlabel, are designed to work with the .lab extension. |
Beta Was this translation helpful? Give feedback.
-
|
Okay, that makes sense! Though, I've always used ".txt" with MFA. Maybe that can be a choice when inferring? |
Beta Was this translation helpful? Give feedback.
-
|
I've committed this feature, but it hasn't been tested. Could you help me test it? |
Beta Was this translation helpful? Give feedback.
-
|
Yes! I just saw the commit, I'll give it a test right now :) |
Beta Was this translation helpful? Give feedback.
-
|
There was one small error but I fixed it and submitted a pull request :) |
Beta Was this translation helpful? Give feedback.
-
|
Thank you very much. |
Beta Was this translation helpful? Give feedback.
For Chinese users, .lab files are the preferred format for transcriptions instead of .txt files. This is because the Montreal Forced Aligner (MFA) utilizes the .lab format, and consequently, nearly all tools in the DiffSinger dataset ecosystem, including LyricFA and minlabel, are designed to work with the .lab extension.