Audiobook Corpus for Acoustic and Comprehension Studies, 20 Literary Pieces

A 9.7 GB collection of 60 audio files derived from 20 English audiobooks. Alireza Malekmohammadi created this dataset, which includes normal, word-shuffled, and time-reversed versions of each literary piece, last updated on May 11, 2026. Each audiobook is sourced from a unique literary piece and narrated by a different voice.

Use Cases

• Train models for speech intelligibility assessment based on manipulated audio signals.
• Study the impact of syntactic structure on comprehension using word-shuffled narratives.
• Benchmark audio processing algorithms against time-reversed speech as a noise control.
• Analyze the effect of narrator voice variation on listener engagement across literary pieces.

Strengths

• Includes 60 total audio files across three distinct manipulation conditions.
• Derived from 20 unique literary pieces with different narrators, providing voice diversity.
• Released under a permissive CC-BY-4.0 license.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.
• The description does not specify the original source or selection criteria for the 20 literary pieces.

Provenance

Source

figshare

Collection Method

Audiobooks were sourced from literary pieces and manipulated to create shuffled and time-reversed versions.

Freshness

Last updated 2026-05-11 11:18:19; freshness should be verified.