A character-based analysis of impacts of dialects on end-to-end Norwegian ASR

Phoebe Parsons, Knut Kvale, Torbjørn Svendsen, Giampiero Salvi

Keywords: ASR, Norwegian, dialect, character-based errors, wav2vec2

TL;DR: We demonstrate, through character-based analysis of system-generated transcriptions, that cues from acoustic dialectal features can influence the output of an end-to-end ASR systems in a dialectally predictable manner.

Abstract: We present a method for analyzing character errors for use with character-based, end-to-end ASR systems, as used herein for investigating dialectal speech. As end-to-end systems are able to produce novel spellings, there exists a possibility that the spelling variants produced by these systems can capture phonological information beyond the intended target word. We therefore first introduce a way of guaranteeing that similar words and characters are paired during alignment, thus ensuring that any resulting analysis of character errors is founded on sound substitutions. Then, from such a careful character alignment, we find trends in system-generated spellings that align with known phonological features of Norwegian dialects, in particular, “r” and “l” confusability and voiceless stop lenition. Through this analysis, we demonstrate that cues from acoustic dialectal features can influence the output of an end-to-end ASR systems.

Student Paper: Yes, the first author is a student