The files in the dataset correspond to results that have been generated
for the Multimedia Tools and Applications (Springer ISSN: 1380-7501 / 1573-7721) article: "Lightly supervised alignment of subtitles on multigenre broadcasts".
The files in the zip file are of three types:
- .ctm, which correspond to the output of the automatic speech
recognition system or lightly supervised alignment system. - .rttm, which correspond to the output of the speech segmentation system.
- .sys, which correspond to scoring of the speech segmentation, automatic speech
recognition or lightly supervised alignment system.
The following is a description about the naming convention of the files:
TableX-LineY-[ser|wer|f1]: This is the output and scoring results corresponding to Line Y of Table X in the article in terms of SER, WER or F1 score.
All three file types are standard outputs that are recognised by the speech technology community and can be opened using any text
EPSRC Programme Grant EP/I031022/1 (Natural Speech Technology)
There is no personal data or any that requires ethical approval
The data complies with the institution and funders' policies on access and sharing