MIR-1K

dataset·singing voice separation dataset·active

Dataset of 1,000 Chinese karaoke song clips with separated accompaniment/vocal channels and manual annotations.

Recorded facts

Official site	https://zenodo.org/records/3532216 ↗
Geography	Taiwan
zenodo	3532216
access note	Zenodo record mirrors an older MIR Lab download; rights provenance for source karaoke songs needs review.
dedupe risk	low
discovery route	Simplified Chinese, Traditional Chinese and English source discovery
documented size	clips: 1000 · duration minutes: 133
evidence strength	strong
recommended prompt	06-datasets-and-benchmarks
classification rationale	Primary evidence shows material singing voice separation dataset relevance to AI music or audio AI.

Sources & changes

Checked 10d ago · highhow verification works

Field-level evidence

VERIFIEDdedupe riskcited source ↗

VERIFIEDexternal ids.zenodocited source ↗

VERIFIEDdocumented sizecited source ↗

VERIFIEDentity typecited source ↗

VERIFIEDsubtypecited source ↗

VERIFIEDcanonical namecited source ↗

VERIFIEDStatuscited source ↗

VERIFIEDshort descriptioncited source ↗

VERIFIEDOfficial URLcited source ↗

VERIFIEDprimary geographycited source ↗

VERIFIEDclassification rationalecited source ↗

VERIFIEDdiscovery routecited source ↗

VERIFIEDevidence strengthcited source ↗

VERIFIEDrecommended promptcited source ↗

VERIFIEDaccess notecited source ↗

Public change history

19 Jul 2026

Status unknown → active
Official URL unknown → zenodo.org/records/3532216
Record maintenance · 12 fields updated

Is this yours? Claim this record →·See something wrong? Report a correction →