Saved in:
Bibliographic Details
Main Authors: Ljubešić, Nikola, Porupski, Ivan, Rupnik, Peter
Format: Preprint
Published: 2025
Subjects:
Online Access:https://arxiv.org/abs/2505.24571
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866916769111736320
author Ljubešić, Nikola
Porupski, Ivan
Rupnik, Peter
author_facet Ljubešić, Nikola
Porupski, Ivan
Rupnik, Peter
contents Automating primary stress identification has been an active research field due to the role of stress in encoding meaning and aiding speech comprehension. Previous studies relied mainly on traditional acoustic features and English datasets. In this paper, we investigate the approach of fine-tuning a pre-trained transformer model with an audio frame classification head. Our experiments use a new Croatian training dataset, with test sets in Croatian, Serbian, the Chakavian dialect, and Slovenian. By comparing an SVM classifier using traditional acoustic features with the fine-tuned speech transformer, we demonstrate the transformer's superiority across the board, achieving near-perfect results for Croatian and Serbian, with a 10-point performance drop for the more distant Chakavian and Slovenian. Finally, we show that only a few hundred multi-syllabic training words suffice for strong performance. We release our datasets and model under permissive licenses.
format Preprint
id arxiv_https___arxiv_org_abs_2505_24571
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Identifying Primary Stress Across Related Languages and Dialects with Transformer-based Speech Encoder Models
Ljubešić, Nikola
Porupski, Ivan
Rupnik, Peter
Audio and Speech Processing
Computation and Language
Sound
Automating primary stress identification has been an active research field due to the role of stress in encoding meaning and aiding speech comprehension. Previous studies relied mainly on traditional acoustic features and English datasets. In this paper, we investigate the approach of fine-tuning a pre-trained transformer model with an audio frame classification head. Our experiments use a new Croatian training dataset, with test sets in Croatian, Serbian, the Chakavian dialect, and Slovenian. By comparing an SVM classifier using traditional acoustic features with the fine-tuned speech transformer, we demonstrate the transformer's superiority across the board, achieving near-perfect results for Croatian and Serbian, with a 10-point performance drop for the more distant Chakavian and Slovenian. Finally, we show that only a few hundred multi-syllabic training words suffice for strong performance. We release our datasets and model under permissive licenses.
title Identifying Primary Stress Across Related Languages and Dialects with Transformer-based Speech Encoder Models
topic Audio and Speech Processing
Computation and Language
Sound
url https://arxiv.org/abs/2505.24571