What methods are used to analyze pitch accents in Japanese dialects

Methods used to analyze pitch accents in Japanese dialects include a range of phonetic, acoustic, and computational techniques:

1. Acoustic Analysis: Researchers use acoustic phonetics tools to measure fundamental frequency (F0) contours in speech to identify pitch accent patterns. This typically involves recording native speakers and analyzing their speech with software like Praat to visualize pitch changes over time. Acoustic cues are crucial as pitch accent in Japanese is primarily realized through pitch height variations rather than stress or intensity. ^{1, 2, 3}

2. Minimal Pair Identification: Experimental studies may test listener perception by presenting minimal pairs of words differing only in pitch accent to determine how pitch differences are perceived in dialects. Noise-vocoded speech has been used to isolate the contribution of pitch and other acoustic dimensions to understand pitch accent perception. ^{4, 5}

3. Morphological and Linguistic Analysis: Linguistic methods involve annotating pitch accent patterns in corpora and comparing phonological rules across dialects. This includes comparing tonal patterns and investigating extrametricality or other formal aspects of pitch accent systems, often in dialects distinct from standard Tokyo Japanese. ^{6, 7}

4. Computational Approaches: Modern approaches utilize machine learning models and text-to-speech systems that incorporate accent prediction components. These use pre-trained language models like BERT embeddings combined with morphological features to predict pitch accent automatically for various dialects. ^{8, 9, 10}

5. Perception and Production Studies: Experimental phonetics research also assesses how both native and second-language speakers produce and perceive pitch accent. This might include cross-linguistic investigations to understand how pitch accent is learned and processed. ^{2, 11}

These methods are complementary and provide a comprehensive understanding of pitch accent variation and representation in Japanese dialects. They range from detailed acoustic measurements to computational modeling to perceptual experiments. ^{3, 9, 11, 1, 2, 4, 6, 8}

References

• Pitch Comparison of Aizu Dialects and Standard Japanese

• Accuracy and Stability in English Speakers’ Production of Japanese Pitch Accent

• Japanese Accent Pronunciation Error by Japanese Learners in Elementary and Intermediate Level

• Identification of Minimal Pairs of Japanese Pitch Accent in Noise-Vocoded Speech

• Identification of Minimal Pairs of Japanese Pitch Accent in Noise-Vocoded Speech

• Tonal Patterns and Extrametricality of Japanese 2-Pattern Accent Systems

• Some Issues In Japanese Accent

• Polyphone disambiguation and accent prediction using pre-trained language models in Japanese TTS front-end

• Cross-Dialect Text-To-Speech in Pitch-Accent Language Incorporating Multi-Dialect Phoneme-Level BERT

• Accent Estimation of Japanese Words from Their Surfaces and Romanizations for Building Large Vocabulary Accent Dictionaries

• Formation of the skill of perception of pitch accent at the initial stage of learning the Japanese language

• Applying Syntax$\unicode{x2013}$Prosody Mapping Hypothesis and Prosodic Well-Formedness Constraints to Neural Sequence-to-Sequence Speech Synthesis

• Focus perception in Japanese: Effects of lexical accent and focus location

• Focus perception in Japanese: Effects of lexical accent and focus location

• The Effects of Lexical Pitch Accent on Infant Word Recognition in Japanese

• Voice onset time and beyond: Exploring laryngeal contrast in 19 languages

• What conditions tone paradigms in Yukuna: Phonological and machine learning approaches