The Mumospee dataset supports the Meetween project's mission of enabling inclusive, language-neutral collaboration across virtual environments. The release provides metadata and download URLs for a curated collection of speech audio sourced from publicly available datasets, optimized for processing on high-performance computing clusters.

Mumospee Overview

Mumospee is a comprehensive multilingual speech-metadata corpus featuring:

140,084 hours of speech metadata across 53,983,241 samples
Coverage of 25 EU languages plus a long tail of additional languages
Collections drawn from existing datasets in different speaking styles and content genres:

_TAGS = ["CoVoST", "GigaSpeech", "PeopleSpeech", "Librispeech", "LibriTTS", "Emilia", "MOSEL"]

A smaller version with fewer than 1000 rows is also available as mumospee_small for testing purposes.

Dataset Statistics

Overview (All Splits Combined)

Metric	Value
Total samples	53,983,241
Total audio duration	140,084h 10m 12.4s (140,084.2 hours)
Average duration per sample	9.34s
Avg transcript length	16.5 words
Total parquet shards	29

Per-Split Overview

Split	# Samples	Duration	Avg Duration	Avg Words	Shards
`train`	53,319,102	139,005h 44m 08.1s (139,005.7h)	9.39s	16.5	27
`test`	341,118	547h 25m 08.3s (547.4h)	5.78s	10.3	1
`validation`	323,021	531h 00m 56.0s (531.0h)	5.92s	10.4	1

Language Distribution

Value	train samples	train %	test samples	test %	validation samples	validation %	Total samples	Total %	Total Duration	Total Dur %
`en`	28,643,120	53.72%	266,720	78.19%	248,203	76.84%	29,158,043	54.01%	67,241h 32m 58.9s	48.00%
`zh`	19,969,319	37.45%	0	0.00%	0	0.00%	19,969,319	36.99%	49,922h 40m 38.9s	35.64%
`ja`	869,665	1.63%	0	0.00%	0	0.00%	869,665	1.61%	1,715h 27m 58.6s	1.22%
`de`	841,219	1.58%	13,511	3.96%	13,511	4.18%	868,241	1.61%	2,682h 25m 04.8s	1.91%
`fr`	777,904	1.46%	14,760	4.33%	14,760	4.57%	807,424	1.50%	2,428h 51m 30.3s	1.73%
`es`	165,080	0.31%	13,221	3.88%	13,221	4.09%	191,522	0.35%	939h 03m 30.5s	0.67%
`it`	123,812	0.23%	8,183	2.40%	8,940	2.77%	140,935	0.26%	886h 02m 54.3s	0.63%
`cs`	106,037	0.20%	0	0.00%	0	0.00%	106,037	0.20%	842h 13m 51.9s	0.60%
`et`	100,734	0.19%	1,571	0.46%	1,576	0.49%	103,881	0.19%	783h 51m 52.6s	0.56%
`pl`	102,193	0.19%	0	0.00%	0	0.00%	102,193	0.19%	811h 23m 20.4s	0.58%
`sl`	100,710	0.19%	360	0.11%	509	0.16%	101,579	0.19%	779h 12m 25.6s	0.56%
`fi`	100,236	0.19%	0	0.00%	0	0.00%	100,236	0.19%	788h 31m 35.8s	0.56%
`sv`	99,891	0.19%	0	0.00%	0	0.00%	99,891	0.19%	787h 41m 55.6s	0.56%
`el`	99,761	0.19%	0	0.00%	0	0.00%	99,761	0.18%	787h 42m 47.6s	0.56%
`pt`	99,487	0.19%	0	0.00%	0	0.00%	99,487	0.18%	781h 43m 50.3s	0.56%
`ro`	99,411	0.19%	0	0.00%	0	0.00%	99,411	0.18%	788h 28m 42.2s	0.56%
`nl`	99,400	0.19%	0	0.00%	0	0.00%	99,400	0.18%	787h 51m 08.8s	0.56%
`hu`	99,143	0.19%	0	0.00%	0	0.00%	99,143	0.18%	776h 56m 01.5s	0.55%
`lt`	99,078	0.19%	0	0.00%	0	0.00%	99,078	0.18%	771h 23m 15.8s	0.55%
`da`	98,868	0.19%	0	0.00%	0	0.00%	98,868	0.18%	768h 47m 46.0s	0.55%
`hr`	97,028	0.18%	0	0.00%	0	0.00%	97,028	0.18%	767h 17m 38.4s	0.55%
`lv`	92,504	0.17%	1,629	0.48%	1,125	0.35%	95,258	0.18%	707h 04m 04.2s	0.50%
`mt`	94,360	0.18%	0	0.00%	0	0.00%	94,360	0.17%	735h 44m 26.9s	0.53%
`sk`	92,345	0.17%	0	0.00%	0	0.00%	92,345	0.17%	731h 14m 15.9s	0.52%
`ko`	92,184	0.17%	0	0.00%	0	0.00%	92,184	0.17%	217h 10m 58.0s	0.16%
`bg`	89,209	0.17%	0	0.00%	0	0.00%	89,209	0.17%	684h 51m 35.5s	0.49%
`ca`	54,255	0.10%	12,730	3.73%	12,730	3.94%	79,715	0.15%	120h 28m 33.5s	0.09%
`fa`	4,348	0.01%	3,445	1.01%	3,445	1.07%	11,238	0.02%	14h 20m 48.6s	0.01%
`ar`	2,776	0.01%	1,695	0.50%	1,758	0.54%	6,229	0.01%	11h 28m 01.2s	0.01%
`mn`	2,018	0.00%	1,759	0.52%	1,761	0.55%	5,538	0.01%	8h 21m 37.1s	0.01%
`id`	1,243	0.00%	844	0.25%	792	0.25%	2,879	0.01%	2h 58m 58.9s	0.00%
`cy`	763	0.00%	690	0.20%	690	0.21%	2,143	0.00%	3h 20m 59.3s	0.00%
`nn`	426	0.00%	0	0.00%	0	0.00%	426	0.00%	3h 24m 46.5s	0.00%
`la`	289	0.00%	0	0.00%	0	0.00%	289	0.00%	2h 21m 50.7s	0.00%
`ru`	113	0.00%	0	0.00%	0	0.00%	113	0.00%	0h 53m 07.0s	0.00%
`he`	66	0.00%	0	0.00%	0	0.00%	66	0.00%	0h 28m 39.7s	0.00%
`sq`	40	0.00%	0	0.00%	0	0.00%	40	0.00%	0h 14m 25.8s	0.00%
`tr`	35	0.00%	0	0.00%	0	0.00%	35	0.00%	0h 17m 23.8s	0.00%
`gl`	15	0.00%	0	0.00%	0	0.00%	15	0.00%	0h 07m 15.9s	0.00%
`uk`	10	0.00%	0	0.00%	0	0.00%	10	0.00%	0h 04m 41.9s	0.00%
`af`	2	0.00%	0	0.00%	0	0.00%	2	0.00%	0h 00m 36.8s	0.00%
`jw`	1	0.00%	0	0.00%	0	0.00%	1	0.00%	0h 00m 30.0s	0.00%
`ur`	1	0.00%	0	0.00%	0	0.00%	1	0.00%	0h 00m 30.0s	0.00%
`sr`	1	0.00%	0	0.00%	0	0.00%	1	0.00%	0h 00m 30.0s	0.00%
`hy`	1	0.00%	0	0.00%	0	0.00%	1	0.00%	0h 00m 16.7s	0.00%
`no`	1	0.00%	0	0.00%	0	0.00%	1	0.00%	0h 00m 30.0s	0.00%

Tag / Source Distribution

Value	train samples	train %	test samples	test %	validation samples	validation %	Total samples	Total %	Total Duration	Total Dur %
`Emilia`	40,237,834	75.47%	0	0.00%	0	0.00%	40,237,834	74.54%	101,585h 04m 02.8s	72.52%
`GigaSpeech`	5,053,116	9.48%	0	0.00%	0	0.00%	5,053,116	9.36%	6,297h 24m 07.6s	4.50%
`CoVoST`	3,591,777	6.74%	290,706	85.22%	288,492	89.31%	4,170,975	7.73%	6,519h 01m 42.7s	4.65%
`MOSEL`	2,300,046	4.31%	0	0.00%	0	0.00%	2,300,046	4.26%	18,127h 01m 37.8s	12.94%
`PeopleSpeech`	1,501,271	2.82%	34,898	10.23%	18,622	5.76%	1,554,791	2.88%	5,987h 42m 22.5s	4.27%
`LibriTTS`	353,817	0.66%	9,955	2.92%	10,340	3.20%	374,112	0.69%	585h 37m 48.6s	0.42%
`Librispeech`	281,241	0.53%	5,559	1.63%	5,567	1.72%	292,367	0.54%	982h 18m 30.3s	0.70%

License Distribution

Value	train samples	train %	test samples	test %	validation samples	validation %	Total samples	Total %
`CC-BY-NC-4.0`	40,237,834	75.47%	0	0.00%	0	0.00%	40,237,834	74.54%
`unknown`	5,053,116	9.48%	0	0.00%	0	0.00%	5,053,116	9.36%
`CC0`	3,591,777	6.74%	290,706	85.22%	288,492	89.31%	4,170,975	7.73%
`CC-BY-4.0`	2,935,104	5.50%	15,514	4.55%	15,907	4.92%	2,966,525	5.50%
`CC-BY;CC-BY-SA`	1,501,271	2.82%	34,898	10.23%	18,622	5.76%	1,554,791	2.88%

unknown rows correspond to the GigaSpeech subset. GigaSpeech's code and manifest are licensed under Apache 2.0, but the audio is governed by the GigaSpeech Data User Agreement and the platform terms (YouTube, podcasts, audiobooks) of the underlying clips. Use is limited to non-commercial academic research; raw-audio redistribution is prohibited upstream.

Data quality notes

MOSEL durations

Per-segment durations for all 2,300,046 MOSEL rows are sourced from VoxPopuli's authoritative segment manifest (unlabelled_v2.tsv), keyed by (event_id, segment_no) parsed from the row's path. Coverage is 100% — no MOSEL row has a missing or zero duration. Median segment is 30.0 s; minimum is 14.95 s.

MOSEL audio and language-label caveats

MOSEL audio is hosted in a separate repository, meetween/mumospee_mosel. Two known caveats apply to MOSEL-tagged rows and are documented on that repository's data card:

Audio coverage is incomplete for some languages (voxpopuli.download_audios did not finish for the full language set). Per-segment transcripts and durations in this dataset remain valid for the affected rows; the missing audio can be fetched from FBK-MT/mosel or VoxPopuli upstream.
Per-segment language labels are inherited from VoxPopuli's session-level metadata and are not human-verified at the utterance level. A fastText lid.176 sanity check disagrees with the declared language on ~1.5% of MOSEL rows, mostly on close-language pairs and short transcripts; the original declared labels are kept as-is here.

Subsets other than MOSEL (CoVoST, GigaSpeech, Emilia, PeoplesSpeech, LibriSpeech, LibriTTS) carry the language labels supplied by their original curators and are not independently verified.

Notes

train: 0 rows with unparseable duration
test: 0 rows with unparseable duration
validation: 0 rows with unparseable duration

Mumospee dataset structure

Each row in the metadata represents one audio sample with the following fields:

path: the relative path of the audio file
url: the link to download the parquet shard containing the audio
type: the sample type (audio or video)
duration: duration in seconds
language: language of the audio
transcript: transcript text
tag: origin dataset (one of _TAGS above)
split: train, test, or validation
license: license governing this sample

Example row:

{
  "path": "3660-172183-0000.flac",
  "url": "https://huggingface.co/datasets/meetween/mumospee_librispeech/resolve/main/librispeech-parquet/dev-other.parquet",
  "type": "audio",
  "duration": 5.405,
  "language": "en",
  "transcript": "GERAINT AS HE HAD BEEN USED TO DO WHEN HE WAS AT ARTHUR'S COURT FREQUENTED TOURNAMENTS",
  "tag": "Librispeech",
  "split": "validation",
  "license": "CC-BY-4.0"
}

Intended Uses

This dataset is designed to enable SpeechLLM and other large language models to support language-neutral virtual meeting applications.

Data Sources

The release includes metadata and download URLs for the following publicly available datasets:

Example usage

# pip install datasets

from datasets import load_dataset

# ── Load all splits at once ───────────────────────────────────────────────────

dataset = load_dataset("meetween/mumospee")
print(dataset)
# DatasetDict({
#     train:      Dataset({features: [...], num_rows: ...})
#     test:       Dataset({features: [...], num_rows: ...})
#     validation: Dataset({features: [...], num_rows: ...})
# })

# ── Load a specific split ─────────────────────────────────────────────────────

train_data      = load_dataset("meetween/mumospee", split="train")
test_data       = load_dataset("meetween/mumospee", split="test")
validation_data = load_dataset("meetween/mumospee", split="validation")

License

The metadata is published under CC-BY-4.0. Each individual sample is governed by its own license, recorded per-row in the license column. Users must comply with the licensing terms of each underlying dataset.

Changelog

Latest update (2026-06-08)

Repackaged from a single 43 GB dataset.csv into 29 sharded Parquet files (27 train, 1 test, 1 validation). Streaming, columnar reads, and load_dataset(...) are substantially faster, and CSV parsing edge cases are eliminated.
Recomputed all headline statistics directly from the released data. Earlier versions of this card cited duration totals from upstream documentation; the figures above are now computed from the actual rows in the parquet shards.
Backfilled MOSEL duration for all 2,300,046 rows from VoxPopuli's segment manifest. Previously these rows carried no real per-segment duration (the ingest hard-coded "n/a", which a downstream cleanup step had converted to 0.0, silently zeroing out ~18,000 hours of MOSEL).
Row repair pass on the metadata:
- rows whose transcript contained unescaped commas (and so were split into >9 fields) were re-joined into the correct 9-column shape;
- rows whose duration carried a unit suffix (e.g. "3.5 s") were stripped to a numeric value;
- rows whose language was empty or longer than 5 characters were dropped;
- exact duplicate rows were removed.
Corrected license attribution. The license value for tag == Emilia rows is now reported as CC-BY-NC-4.0 to match the upstream Emilia license; an earlier version of this card had bucketed those rows under CC-BY-4.0 in the distribution table even though the per-row values were already correct.
Corrected MOSEL language for 55,387 rows by overwriting with the VoxPopuli interpretation-channel code embedded in path (e.g. ..._bg_3 → bg).
duration set to NaN for 252 CoVoST rows whose source MP3 bytes are unreadable upstream (mutagen and ffprobe both fail to parse them). These rows still carry valid path, transcript, and language; users requiring duration should filter with df["duration"].notna().
Schema verification: 0 rows with empty language, 0 rows with len(language) > 5, 0 rows with duration == 0 (252 CoVoST rows are null/NaN — see above — and excluded from duration statistics).

The row schema (path, url, type, duration, language, transcript, tag, split, license), the _TAGS set, and the load_dataset(...) filters are unchanged.

Downloads last month: 66

Total file size:

4.61 GB