Donate Speech Corpus: Test data (10h)
View resource name in all available languages
Lahjoita puhetta -aineisto: Testidata (10h)
Persistent Identifier of this resource:
This resource is available for download in Kielipankki - The Language Bank of Finland as part of "Donate Speech: Selected dataset", http://urn.fi/urn:nbn:fi:lb-2022060127.
The resource contains a subset of 10 hours of transcribed speech that was selected from the Donate Speech Corpus and used for testing an ASR system at Aalto University.
The test data includes at least ten minutes of speech for each metadata class in each of the five metadata domains (age, dialect, gender, native/non-native and theme). The set contains speech from 103 different speakers (according to the metadata accompanying the original recordings). The gender ratio has been debiased, so that the set includes over 40% male speakers (similarly to the puhelahjat-dev set, while the puhelahjat-train set has just over 20% of male speakers).
For speech technology development purposes, the test dataset can be used together with the puhelahjat-dev and puhelahjat-train datasets. There is no overlap of speakers between these three sets.
People who looked at this resource also viewed the following: