Comparison of fundamental frequency estimators with subharmonic voice signals

2025-01-08

AI Summary

Subharmonic phonation in clinical voice analysis frequently causes erroneous fundamental frequency (F₀) estimation, undermining diagnostic reliability. Method: We constructed the first sustained-vowel subharmonic dataset, proposed the Subharmonic–Harmonic Ratio (SHR) as a quantitative metric for subharmonic intensity, and developed a quality-aware classification framework to identify subharmonic-induced F₀ estimation errors. We systematically evaluated the robustness of leading F₀ estimators—FCN-F₀, CREPE, Harvest, and Praat/YAAPT—under subharmonic conditions. Results: FCN-F₀ achieved the highest overall accuracy and superior subharmonic discrimination; CREPE and Harvest followed closely. SHR effectively exposed inter-algorithm differences in subharmonic sensitivity. This work establishes a novel benchmark and provides practical tools for objective assessment and algorithmic refinement in subharmonic voice analysis.

Technology Category

• Machine Learning: Evaluation and Analysis
• Natural Language Processing: Speech
• Intelligent Robots: State Estimation

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs
• Web Mining and Content Analysis: Web data quality in the era of algorithmically-generated content

Abstract

In clinical voice signal analysis, mishandling of subharmonic voicing may cause an acoustic parameter to signal false negatives. As such, the ability of a fundamental frequency estimator to identify speaking fundamental frequency is critical. This paper presents a sustained-vowel study, which used a quality-of-estimate classification to identify subharmonic errors and subharmonics-to-harmonics ratio (SHR) to measure the strength of subharmonic voicing. Five estimators were studied with a sustained vowel dataset: Praat, YAAPT, Harvest, CREPE, and FCN-F0. FCN-F0, a deep-learning model, performed the best both in overall accuracy and in correctly resolving subharmonic signals. CREPE and Harvest are also highly capable estimators for sustained vowel analysis.

Problem

• Subharmonic Signals
• Fundamental Frequency Measurement
• Sustained Vowel Signals

Innovation

• Subharmonic speech signal analysis
• FCN-F0 performance
• Advanced computational techniques