Human and AI Speech Samples for Intonation and Emotional Content Comparison

Two audio samples in WAV format, one of human speech and one of AI-generated speech, speaking the same words. A short video provides a phrase-by-phrase comparison of the samples. The dataset is associated with the research paper 'Continuous human meaning cannot be represented by discrete AI' by William Stewart, last updated in June 2026.

Use Cases

• Compare intonation patterns between human and AI speech based on the provided WAV samples.
• Analyze emotional content differences in synthetic speech based on the audio comparison.
• Benchmark AI speech generation models for prosodic and affective qualities based on the human reference sample.
• Study multimodal presentation of audio data for perceptual analysis based on the accompanying comparison video.

Strengths

• Provides direct, high-fidelity WAV format audio samples for acoustic analysis.
• Includes a video for convenient side-by-side and phrase-by-phrase comparison of the samples.
• Is associated with a specific research paper providing a deep analysis of the samples.

Limitations

• Dataset contains only two audio samples, limiting statistical analysis.
• Row count and column-level documentation are unknown, which may limit suitability assessment.
• Description metadata is limited; actual data quality requires manual inspection after download.

Provenance

Source

Harvard Dataverse, author William Stewart.

Collection Method

Likely created for a specific research project comparing human and AI speech.

Freshness

Last updated 2026-06-08 14:58:55; freshness should be verified.