Update README.md

README.md

@@ -24,3 +24,40 @@ configs:
   - split: test
     path: data/test-*
 ---
+
+
+# Talrómur Dataset (ROSA Speaker Subset)
+
+This Huggingface dataset card describes a subset of the Talrómur speech corpus, focusing exclusively on the "ROSA" speaker data. This is not my original dataset, but rather a subset of the publicly available Talrómur corpus.
+
+## Dataset Description
+
+Talrómur is a public domain speech corpus designed for text-to-speech research and development. The complete corpus consists of 122,417 short audio clips from eight different speakers reading short sentences. The audio was recorded in 2020 by Reykjavík University and The Icelandic National Broadcasting Service as part of The Icelandic Language Technology Program.
+
+The original dataset is publicly accessible from the CLARIN-IS repository at: https://repository.clarin.is/repository/xmlui/handle/20.500.12537/330
+
+This subset contains only the recordings from the speaker identified as "ROSA".
+
+## Features
+
+- Loudness-normalized recordings using ffmpeg's implementation of EBU R128
+- Alignments in both JSON and Textgrid format obtained using the Montreal Forced Aligner (MFA)
+- Trimmed recordings with at most 100ms silence at the beginning and end of each clip
+
+## Citation
+
+If you use this dataset in your research, please cite:
+
+```bibtex
+@misc{20.500.12537/330,
+ title = {Talr{\'o}mur 1 (24.04)},
+ author = {Sigurgeirsson, Atli {\TH}{\'o}r and Gunnarsson, {\TH}orsteinn Da{\dh}i and {\"O}rn{\'o}lfsson, Gunnar Thor and {\TH}{\'o}rhallsd{\'o}ttir, Ragnhei{\dh}ur and Magn{\'u}sd{\'o}ttir, Eyd{\'{\i}}s Huld and Gu{\dh}nason, J{\'o}n},
+ url = {http://hdl.handle.net/20.500.12537/330},
+ note = {{CLARIN}-{IS}},
+ copyright = {Creative Commons - Attribution 4.0 International ({CC} {BY} 4.0)},
+ year = {2024} }
+```
+
+## Acknowledgments
+
+The original Talrómur dataset was created as part of The Icelandic Language Technology Program by Reykjavík University and The Icelandic National Broadcasting Service. This Huggingface dataset card is merely a description of a subset of their work and is intended to provide proper attribution and information about the "ROSA" speaker subset.