Voices
Voices used with text-to-speech (TTS) systems.
The North Sámi female voice Biret is based on the same recording that was previously made in a cooperation between the Norwegian Sámi Parliament and Acapela (ca 2012-2015). The old recordings have been augmented with new recordings with the same speaker.
The voice can read Norwegian, and partially Swedish and English, in addition to North Sámi.
The Lule Sámi male voice package contains two voices: Nihkol and Ábmut. Both recorded speakers are from the Norwegian part of the Lule Sámi area.
Nihkol can read Norwegian texts quite ok (although not without errors). Ábmut is not trained with anything but Lule Sámi input, and will probably not work well on multilingual texts.
The two voices are in the same package, because that way both voices perform better.
The South Sámi female voice Aanna is based on archival recordings of Anna Jacobsen (1924-2004). The recordings are used with acknowledgement and acceptance of her descendants. The archival recordings are taken from the Norwegian national broadcaster NRK, and from audio books, and are digitally restored, cleaned, improved and transcribed by the Divvun group at UiT in cooperation with the Sámi Archives and the National Archives of Norway.
The Lule Sámi female voice Siggá is based on a speaker for Sweden.
The voice can also read Swedish, Norwegian and English, but not without errors.