OSynth is released under CC BY 4.0. The people, persona names, organizations, meetings, events, and scenarios depicted in OSynth are fictional. Any resemblance or overlap with real people (living or deceased), organizations, or events is coincidental. Synthetic persona names do not identify the speakers whose recordings conditioned voice synthesis.

OSynth — text

Synthetic multi-party meetings embedded in simulated organizations ("workspaces"), with the surrounding workspace artifacts and the latent ground-truth state that generated them. 160 workspaces, 2,126 meetings. The matching audio (570 h) is in OSynth — audio; join on workspace_id + meeting_id.

Splits

test is the held-out evaluation set (15 workspaces, 238 meetings); every result in the paper is computed on it. train is everything else. Splits are by workspace, so no workspace appears in both.

cohort	split	workspaces	meetings	hours	description
eval	test	15	238	19.8	Held-out evaluation set (1-day and 14-day workspaces); all paper results
month	train	20	530	40.1	25-day workspaces with many short meetings
week	train	27	690	207.5	1-week workspaces
day	train	39	245	53.1	1-day workspaces
scenario	train	12	116	8.8	Product-design scenario workspaces with a four-person team
long_meeting	train	47	307	240.8	2-3 day workspaces with few, long (~47 min) meetings

Configs

config	one row per	key columns
`meetings`	meeting	spec, attendance, join mode, transcripts, RTTM, timing, summary, outline, notes
`workspaces`	workspace	profile, org chart and personas, projects, calendar, narrative, week plans
`artifacts`	email thread or document	content, spec, jargon
`latent_state`	workspace	event log, reconciled state, artifact alignment
`word_alignments`	meeting	partial word-level timing (~10% of turns, model-annotated for turn-overlap calculations; not reference)

Nested objects (org chart, specs, timing, event logs, ...) are JSON strings; use json.loads. The three transcripts are: transcript (as spoken, with disfluencies), transcript_canonical (clean), transcript_semantic_verbatim. RTTM speaker labels and speaker_ids match the persona ids in workspaces.org.

from datasets import load_dataset
meetings = load_dataset("iporada-otterai/osynth-text", "meetings", split="test")

Planned improvements for v2

We plan to address the following in v2:

Bracketed sound descriptions. Some turns contain annotations such as [car noise] that the synthesized speaker reads aloud. These words are actually spoken, so their inclusion in the reference transcript is appropriate for WER scoring, but the result is unrealistic. We plan to handle these annotations without having speakers read them aloud.
Speaker consistency across turns. Improve the consistency of each speaker's voice from one turn to the next.
Abrupt turn endings. Address turns that are cut off abruptly by voice activity detection (VAD).

Feedback, examples of affected turns, and suggestions are welcome in the Community discussions.

Attribution

Credit the OSynth contributors, link to this dataset and the CC BY 4.0 license, and indicate any modifications. The dataset URL is the attribution reference for this release.

License and access acknowledgment

This dataset is released under CC BY 4.0; see LICENSE. Reuse, including commercial reuse, is allowed subject to the license, including attribution and indicating modifications.

The people, persona names, organizations, meetings, events, and scenarios depicted in OSynth are fictional. Any resemblance or overlap with real people (living or deceased), organizations, or events is coincidental. Synthetic persona names do not identify the speakers whose recordings conditioned voice synthesis.

Downloaders must accept the license and acknowledge the fictional nature of the dataset through the Hugging Face access form. See ACCESS_AGREEMENT.md for the acknowledgment text.

Downloads last month: -

Total file size:

145 MB