Datasets:
dataset_info:
features:
- name: audio
dtype: audio
- name: raw_text
dtype: string
- name: normalized_text
dtype: string
- name: speaker_id
dtype: string
splits:
- name: train
num_bytes: 72553370
num_examples: 590
download_size: 72159005
dataset_size: 72553370
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
task_categories:
- automatic-speech-recognition
tags:
- messenian
Messenian is a Greek dialect spoken in the Southwest of the Peloponnese and, thus, it belongs to the Peloponnesian varieties. Historically, Peloponnesian has been considered as one of the basic dialects on which Standard Modern Greek was formed, mainly because of the pivotal role of the Peloponnese during the Greek Revolution of 1821 and the subsequent formation of the Greek state, as well as of the significant migratory movements from the Peloponnese to Athens. Recent studies of some documented linguistic material from the Peloponnesian varieties reveal substantial deviations from Standard Modern Greek, and differences from one variety to the other, particularly on the phonological level. For instance, they display palatalization of the lateral /l/ and the nasal /n/. However, detailed empirical studies and an account of the exact properties of Messenian and how it diverges from other Peloponnesian dialects are lacking.
To assemble the corpus, a field linguist (Katerina Mouzou) interviewed residents from the town of Kalamata and five closeby villages (Bounaria, Sotirianika, Petalidi, Filiatra, Altomira), resulting in 39 minutes of narratives obtained from six speakers (2 male, 4 female). This data collection was carried out in 2023-2024, after obtaining written consent from the informants. For the initial transcription of the audio files, Large-v3 was employed. The transcripts were then manually corrected by the field linguist who is also a speaker of this Greek dialect.
This corpus was used in the Interspeech proceedings paper Speech Recognition for Greek Dialects: A Challenging Benchmark. The paper presented a comprehensive study of automatic speech recognition (ASR) for low-resource Greek varieties, focusing on the unique challenges posed by dialects.