Samrómur Queries Icelandic Speech 1.0
| Item Name: | Samrómur Queries Icelandic Speech 1.0 |
| Author(s): | Staffan Hedström, Judy Fong, Ragnheiður Þórhallsdóttir, David Mollberg, Smári Freyr Guðmundsson, Ólafur Helgi Jónsson, Sunneva Þorsteinsdóttir, Eydis Huld Magnusdottir, Jon Gudnason |
| LDC Catalog No.: | LDC2023S05 |
| ISLRN: | 363-728-488-848-1 |
| DOI: | https://doi.org/10.35111/aq18-1540 |
| Release Date: | August 15, 2023 |
| Member Year(s): | 2023 |
| DCMI Type(s): | Sound, Text |
| Sample Type: | flac |
| Sample Rate: | 16000 |
| Data Source(s): | web collection |
| Application(s): | speaker identification, speaker verification, speech recognition |
| Language(s): | Icelandic |
| Language ID(s): | isl |
| License(s): |
Samrómur Queries Icelandic Speech 1.0 Agreement (For-Profit Member) Samrómur Queries Icelandic Speech 1.0 Agreement (Non-Member) Samrómur Queries Icelandic Speech 1.0 Agreement (Not-for-Profit) |
| Online Documentation: | LDC2023S05 Documents |
| Licensing Instructions: | Subscription & Standard Members, and Non-Members |
| Citation: | Hedström, Staffan, et al. Samrómur Queries Icelandic Speech 1.0 LDC2023S05. Web Download. Philadelphia: Linguistic Data Consortium, 2023. |
| Related Works: | View |
Introduction
Samrómur Queries Icelandic Speech 1.0 was developed by the Language and Voice Lab, Reykjavik University in cooperation with Almannarómur, Center for Language Technology. The corpus contains 20 hours of Icelandic prompted queries from 3,809 speakers representing 17,475 utterances.
This version 1.0 is equivalent to "Samrómur Queries Icelandic Speech 21.12" as used by the Language Technology Programme for Icelandic 2019-2023.
Data
Speech data was collected between October 2019 and December 2021 using the Samrómur website which displayed prompts to participants. The prompts were mainly from The Icelandic Gigaword Corpus, which includes text from novels, news, plays, and from a list of location names in Iceland. Additional prompts were taken from the Icelandic Web of Science and others were created by combining a name followed by a question. Prompts and speaker metadata are included in the corpus.
The audio data is divided into train, dev, and test sets and is presented as flac compressed, single channel, 16 kHz, 16-bit linear PCM.
Samples
Please view this audio sample (FLAC).
Updates
None at this time.