LibriSpeech Segment: English Read Speech with Phone-Level Alignments

LibriSpeech Segment is an English read-speech corpus with phone-level time alignments generated by the Montreal Forced Aligner. The dataset is derived from the LibriSpeech corpus and is suitable for training and evaluating phone recognition and phonetic segmentation models. It was created by changelinglab and last updated on the platform in April 2026.

Use Cases

• Training phone recognition models based on the provided phone-level time alignments.
• Evaluating phonetic segmentation models using the forced alignment data.
• Researching speech acoustics and phonetics with a time-aligned English speech corpus.

Strengths

• Includes phone-level time alignments, a feature mentioned as suitable for model training.
• Derived from the established LibriSpeech corpus, suggesting a known baseline quality.
• Contains at least 28,538 utterances in the 'train.clean.100' split, as stated in the description.

Limitations

• Column-level documentation is absent; field semantics must be inferred after download.
• Row count for the full dataset is unknown, which may limit suitability assessment.
• Data may reflect bias inherent to the source LibriSpeech corpus, such as read speech from audiobooks.

Provenance

Source

Audio sourced from LibriSpeech (OpenSLR 12) by Vassil Panayotov et al. (2015). Phone alignments from anyspeech/librispeech_MFA_alignments.

Collection Method

Phone alignments were generated using the Montreal Forced Aligner (MFA).

Freshness

Last updated 2026-04-12 17:09:10; freshness should be verified.