UNGA‑80 Simultaneous Interpreting Dataset: Acoustic‑Prosodic & Linguistic Metrics for the Dual‑Modality Analytical Framework (DMAF)

The corpus is built upon plenary speeches from the 80th session of the United Nations General Assembly and outputs from three professional interpreters. Eight standardized folders are included: source‑speech transcripts, aligned source‑simultaneous‑interpreting text pairs, Praat TextGrid acoustic annotation files, interpreter transcripts with human‑evaluated quality scores, extracted acoustic‑prosodic and linguistic feature tables, hypothesis‑driven statistical files, annotation standard‑operating‑procedure manuals, and raw Excel sources for all manuscript figures.

Inter‑annotator agreement reaches κ=0.89 under unified coding schemes. Original UN audio recordings are not deposited here due to copyright restrictions; only transcribed, annotated and derived tabular data are provided for replication. Dataset structure: 01_SL‑SpeechTranscription: source‑language speech transcripts 02_SL‑SI‑Aligned‑Audio&Text: aligned source‑SI text segments 03_SL‑Speech‑PraatAnnotated: Praat TextGrid annotation files 04_SI‑TextTranscription&QualityScore: SI transcripts and manual quality scores 05_SL‑Features: acoustic‑prosodic and linguistic metric tables 06_StatisticalAnalysis‑FilesHypothesis: mixed‑linear‑model input tables per research hypothesis 07_Protocols‑SOP: annotation guidelines and calculation protocols 08‑Figures: raw spreadsheet sources for manuscript figures