Get the buckeye corpus here (referred to as "english" and "sample" in the challenge).
Get the NCHLT xitsonga corpus here (referred to as "xitsonga" in the challenge).
The challenge restricted the corpus to the following files:
- sample: sample_files.txt to extract from the Buckeye
- english: english_files.txt to extract from the Buckeye
- xitsonga: xitsonga_files.txt to extract from the NCHLT
We provide for each dataset a voice activity detection file (e.g. sample_vad.txt). For The Track 2 evaluation, you will only be evaluated on those intervals.
Track 1 repository: ABXpy.
Track 2 repository: tde.