NST Pronunciation Lexicon for Norwegian Bokmål

This fullform pronunciation lexicon for Norwegian Bokmål was originally produced by Nordic Language Technology AS (NST), and contains approximately 785,000 entries. The lexicon was made to support the development of speech technology, and is based on the 100,000 most frequent word forms in NST's Norwegian text corpus. The lexicon is stored as one large csv-file. Each entry (line) contains 51 fields, separated by a semicolon. Not all fields are equally relevant, but given the format, it should be easy to extract relevant information. The phonetic transcription was in part done manually, but to a large extent it was done automatically using an inflector. Parts of the output of this process was manually checked afterwards. The inflector and other lexical tools that can be used in processing the lexicon, can be downloaded as a separate file. The words are transcribed using SAMPA (Speech Assessment Methods Phonetic Alphabet). See http://www.phon.ucl.ac.uk/home/sampa/index.html for more information on this transcription format. A script for converting the SAMPA transcriptons in the lexicon to IPA can be found on GitHub (https://github.com/peresolb/sampa_to_ipa).