{
  "@context": {
    "@vocab": "https://schema.org/",
    "sc": "https://schema.org/",
    "cr": "http://mlcommons.org/croissant/",
    "rai": "http://mlcommons.org/croissant/RAI/",
    "prov": "http://www.w3.org/ns/prov#"
  },
  "@type": "sc:Dataset",
  "name": "AnimalSpeak Pseudovox Train-Unseen",
  "description": "Train-unseen split of AnimalSpeak Pseudovox: short, silence-trimmed, single-vocalization clips with per-clip metadata.",
  "url": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox",
  "version": "1.0.0",
  "datePublished": "2026-05-06",
  "license": "CC-BY-NC-4.0",
  "isAccessibleForFree": false,
  "isLiveDataset": false,
  "keywords": [
    "bioacoustics",
    "animal vocalizations",
    "single vocalization",
    "AnimalSpeak",
    "Pseudovox",
    "train unseen"
  ],
  "creator": [
    {
      "@type": "Organization",
      "name": "Earth Species Project"
    }
  ],
  "publisher": {
    "@type": "Organization",
    "name": "Earth Species Project",
    "url": "https://www.earthspecies.org/"
  },
  "distribution": [
    {
      "@type": "cr:FileObject",
      "@id": "metadata-parquet",
      "name": "Per-example metadata",
      "description": "Parquet table with one row per audio clip (346,907 rows).",
      "contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/metadata.parquet",
      "encodingFormat": "application/vnd.apache.parquet",
      "prov:wasGeneratedBy": {
        "@id": "activity-hf-export"
      }
    },
    {
      "@type": "cr:FileObject",
      "@id": "shard-000000",
      "name": "WebDataset shard 000000",
      "description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
      "contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000000.tar",
      "encodingFormat": "application/x-tar",
      "prov:wasGeneratedBy": {
        "@id": "activity-hf-export"
      }
    },
    {
      "@type": "cr:FileObject",
      "@id": "shard-000001",
      "name": "WebDataset shard 000001",
      "description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
      "contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000001.tar",
      "encodingFormat": "application/x-tar",
      "prov:wasGeneratedBy": {
        "@id": "activity-hf-export"
      }
    },
    {
      "@type": "cr:FileObject",
      "@id": "shard-000002",
      "name": "WebDataset shard 000002",
      "description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
      "contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000002.tar",
      "encodingFormat": "application/x-tar",
      "prov:wasGeneratedBy": {
        "@id": "activity-hf-export"
      }
    },
    {
      "@type": "cr:FileObject",
      "@id": "shard-000003",
      "name": "WebDataset shard 000003",
      "description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
      "contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000003.tar",
      "encodingFormat": "application/x-tar",
      "prov:wasGeneratedBy": {
        "@id": "activity-hf-export"
      }
    },
    {
      "@type": "cr:FileObject",
      "@id": "shard-000004",
      "name": "WebDataset shard 000004",
      "description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
      "contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000004.tar",
      "encodingFormat": "application/x-tar",
      "prov:wasGeneratedBy": {
        "@id": "activity-hf-export"
      }
    },
    {
      "@type": "cr:FileObject",
      "@id": "shard-000005",
      "name": "WebDataset shard 000005",
      "description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
      "contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000005.tar",
      "encodingFormat": "application/x-tar",
      "prov:wasGeneratedBy": {
        "@id": "activity-hf-export"
      }
    },
    {
      "@type": "cr:FileObject",
      "@id": "shard-000006",
      "name": "WebDataset shard 000006",
      "description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
      "contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000006.tar",
      "encodingFormat": "application/x-tar",
      "prov:wasGeneratedBy": {
        "@id": "activity-hf-export"
      }
    },
    {
      "@type": "cr:FileObject",
      "@id": "shard-000007",
      "name": "WebDataset shard 000007",
      "description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
      "contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000007.tar",
      "encodingFormat": "application/x-tar",
      "prov:wasGeneratedBy": {
        "@id": "activity-hf-export"
      }
    },
    {
      "@type": "cr:FileObject",
      "@id": "shard-000008",
      "name": "WebDataset shard 000008",
      "description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
      "contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000008.tar",
      "encodingFormat": "application/x-tar",
      "prov:wasGeneratedBy": {
        "@id": "activity-hf-export"
      }
    },
    {
      "@type": "cr:FileObject",
      "@id": "shard-000009",
      "name": "WebDataset shard 000009",
      "description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
      "contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000009.tar",
      "encodingFormat": "application/x-tar",
      "prov:wasGeneratedBy": {
        "@id": "activity-hf-export"
      }
    },
    {
      "@type": "cr:FileObject",
      "@id": "shard-000010",
      "name": "WebDataset shard 000010",
      "description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
      "contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000010.tar",
      "encodingFormat": "application/x-tar",
      "prov:wasGeneratedBy": {
        "@id": "activity-hf-export"
      }
    },
    {
      "@type": "cr:FileObject",
      "@id": "shard-000011",
      "name": "WebDataset shard 000011",
      "description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
      "contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000011.tar",
      "encodingFormat": "application/x-tar",
      "prov:wasGeneratedBy": {
        "@id": "activity-hf-export"
      }
    },
    {
      "@type": "cr:FileObject",
      "@id": "shard-000012",
      "name": "WebDataset shard 000012",
      "description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
      "contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000012.tar",
      "encodingFormat": "application/x-tar",
      "prov:wasGeneratedBy": {
        "@id": "activity-hf-export"
      }
    },
    {
      "@type": "cr:FileObject",
      "@id": "shard-000013",
      "name": "WebDataset shard 000013",
      "description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
      "contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000013.tar",
      "encodingFormat": "application/x-tar",
      "prov:wasGeneratedBy": {
        "@id": "activity-hf-export"
      }
    },
    {
      "@type": "cr:FileObject",
      "@id": "shard-000014",
      "name": "WebDataset shard 000014",
      "description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
      "contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000014.tar",
      "encodingFormat": "application/x-tar",
      "prov:wasGeneratedBy": {
        "@id": "activity-hf-export"
      }
    },
    {
      "@type": "cr:FileObject",
      "@id": "shard-000015",
      "name": "WebDataset shard 000015",
      "description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
      "contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000015.tar",
      "encodingFormat": "application/x-tar",
      "prov:wasGeneratedBy": {
        "@id": "activity-hf-export"
      }
    },
    {
      "@type": "cr:FileObject",
      "@id": "shard-000016",
      "name": "WebDataset shard 000016",
      "description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
      "contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000016.tar",
      "encodingFormat": "application/x-tar",
      "prov:wasGeneratedBy": {
        "@id": "activity-hf-export"
      }
    },
    {
      "@type": "cr:FileObject",
      "@id": "shard-000017",
      "name": "WebDataset shard 000017",
      "description": "Tar shard containing AnimalSpeak Pseudovox WAV clips under audio/.",
      "contentUrl": "https://huggingface.co/datasets/EarthSpeciesProject/animalspeak-pseudovox/resolve/main/data-20k/train-000017.tar",
      "encodingFormat": "application/x-tar",
      "prov:wasGeneratedBy": {
        "@id": "activity-hf-export"
      }
    },
    {
      "@type": "cr:FileSet",
      "@id": "audio-file-set",
      "name": "WAV audio files inside WebDataset shards",
      "description": "16 kHz WAV clips stored inside tar shards under audio/.",
      "containedIn": [
        {
          "@id": "shard-000000"
        },
        {
          "@id": "shard-000001"
        },
        {
          "@id": "shard-000002"
        },
        {
          "@id": "shard-000003"
        },
        {
          "@id": "shard-000004"
        },
        {
          "@id": "shard-000005"
        },
        {
          "@id": "shard-000006"
        },
        {
          "@id": "shard-000007"
        },
        {
          "@id": "shard-000008"
        },
        {
          "@id": "shard-000009"
        },
        {
          "@id": "shard-000010"
        },
        {
          "@id": "shard-000011"
        },
        {
          "@id": "shard-000012"
        },
        {
          "@id": "shard-000013"
        },
        {
          "@id": "shard-000014"
        },
        {
          "@id": "shard-000015"
        },
        {
          "@id": "shard-000016"
        },
        {
          "@id": "shard-000017"
        }
      ],
      "includes": "audio/*.wav",
      "encodingFormat": "audio/wav",
      "prov:wasGeneratedBy": [
        {
          "@id": "activity-source-separation"
        },
        {
          "@id": "activity-window-extraction"
        },
        {
          "@id": "activity-hf-export"
        }
      ]
    }
  ],
  "recordSet": [
    {
      "@type": "cr:RecordSet",
      "@id": "metadata",
      "name": "Metadata rows",
      "description": "One row per Pseudovox clip. The audio_name column refers to the WAV entry stored under audio/ inside a tar shard.",
      "key": [
        {
          "@id": "metadata/id"
        }
      ],
      "field": [
        {
          "@type": "cr:Field",
          "@id": "metadata/id",
          "name": "id",
          "description": "Stable clip identifier, derived from the WAV filename stem.",
          "dataType": "sc:Text",
          "source": {
            "fileObject": {
              "@id": "metadata-parquet"
            },
            "extract": {
              "column": "id"
            }
          }
        },
        {
          "@type": "cr:Field",
          "@id": "metadata/audio_name",
          "name": "audio_name",
          "description": "WAV filename inside shard audio/ directories.",
          "dataType": "sc:Text",
          "source": {
            "fileObject": {
              "@id": "metadata-parquet"
            },
            "extract": {
              "column": "audio_name"
            }
          }
        },
        {
          "@type": "cr:Field",
          "@id": "metadata/duration_s",
          "name": "duration_s",
          "description": "Clip duration in seconds.",
          "dataType": "sc:Float",
          "source": {
            "fileObject": {
              "@id": "metadata-parquet"
            },
            "extract": {
              "column": "duration_s"
            }
          },
          "prov:wasGeneratedBy": {
            "@id": "activity-window-extraction"
          }
        },
        {
          "@type": "cr:Field",
          "@id": "metadata/species_scientific_gbif",
          "name": "species_scientific_gbif",
          "description": "Scientific species name when available.",
          "dataType": "sc:Text",
          "source": {
            "fileObject": {
              "@id": "metadata-parquet"
            },
            "extract": {
              "column": "species_scientific_gbif"
            }
          },
          "prov:wasGeneratedBy": {
            "@id": "activity-species-labeling"
          }
        },
        {
          "@type": "cr:Field",
          "@id": "metadata/custom_classifier_species",
          "name": "custom_classifier_species",
          "description": "Top species predicted by the custom classifier when available.",
          "dataType": "sc:Text",
          "source": {
            "fileObject": {
              "@id": "metadata-parquet"
            },
            "extract": {
              "column": "custom_classifier_species"
            }
          },
          "prov:wasGeneratedBy": {
            "@id": "activity-species-labeling"
          }
        },
        {
          "@type": "cr:Field",
          "@id": "metadata/custom_classifier_max_p",
          "name": "custom_classifier_max_p",
          "description": "Custom classifier maximum probability for the predicted species.",
          "dataType": "sc:Float",
          "source": {
            "fileObject": {
              "@id": "metadata-parquet"
            },
            "extract": {
              "column": "custom_classifier_max_p"
            }
          },
          "prov:wasGeneratedBy": {
            "@id": "activity-species-labeling"
          }
        },
        {
          "@type": "cr:Field",
          "@id": "metadata/birdnet_confidence",
          "name": "birdnet_confidence",
          "description": "BirdNET confidence score.",
          "dataType": "sc:Float",
          "source": {
            "fileObject": {
              "@id": "metadata-parquet"
            },
            "extract": {
              "column": "birdnet_confidence"
            }
          },
          "prov:wasGeneratedBy": {
            "@id": "activity-birdnet-scoring"
          }
        }
      ],
      "examples": [
        {
          "id": "Amazona_rhodocorytha_SSS_XC3_Jandaira_source1_clip1",
          "audio_name": "Amazona_rhodocorytha_SSS_XC3_Jandaira_source1_clip1.wav",
          "duration_s": 0.37615625
        }
      ]
    },
    {
      "@type": "cr:RecordSet",
      "@id": "audio-files",
      "name": "Audio files",
      "description": "Audio objects extracted from the tar shards.",
      "field": [
        {
          "@type": "cr:Field",
          "@id": "audio-files/path",
          "name": "path",
          "dataType": "sc:Text",
          "source": {
            "fileSet": {
              "@id": "audio-file-set"
            },
            "extract": {
              "fileProperty": "fullpath"
            }
          }
        },
        {
          "@type": "cr:Field",
          "@id": "audio-files/content",
          "name": "content",
          "dataType": "sc:AudioObject",
          "source": {
            "fileSet": {
              "@id": "audio-file-set"
            },
            "extract": {
              "fileProperty": "content"
            }
          }
        }
      ]
    }
  ],
  "measurementTechnique": [
    "BirdMixIT source separation",
    "DRASDIC-style short activity-window extraction",
    "sl-efficientnet-bio species prediction and metadata-match filtering",
    "BirdNET confidence scoring"
  ],
  "variableMeasured": [
    {
      "@type": "PropertyValue",
      "name": "duration_s",
      "description": "Clip duration in seconds."
    },
    {
      "@type": "PropertyValue",
      "name": "species_scientific_gbif",
      "description": "Scientific species name when available."
    },
    {
      "@type": "PropertyValue",
      "name": "custom_classifier_max_p",
      "description": "Confidence of the custom species classifier."
    },
    {
      "@type": "PropertyValue",
      "name": "birdnet_confidence",
      "description": "BirdNET confidence score."
    }
  ],
  "usageInfo": {
    "downstreamUse": "Research on bioacoustic representation learning, vocalization-level audio-language training data, and evaluation of held-out species generalization.",
    "outOfScopeUse": "Do not use as the sole basis for ecological monitoring, sensitive species localization, or unsupported claims about real-world species presence.",
    "biasRisksLimitations": {
      "bias": "Coverage reflects the source recording collection, train-unseen split construction, and classifier confidence filters.",
      "risks": "Models trained on this data can inherit taxonomic, geographic, and recording-quality biases.",
      "limitations": "Clips are short isolated vocalizations and may not represent full soundscape context, overlapping callers, or field deployment conditions."
    },
    "safetyMitigations": "Validate on real target-domain recordings, document target taxa and recording conditions, and consult bioacoustics expertise for high-stakes uses."
  },
  "rai:dataLimitations": "Short isolated Pseudovox clips do not capture all ecological and acoustic context present in the original recordings.",
  "rai:dataBiases": "Taxonomic and acoustic coverage is inherited from source recordings, split criteria, and classifier confidence thresholds.",
  "rai:personalSensitiveInformation": "No human personal data is intentionally included. Source recordings may contain incidental metadata or background sounds inherited from the source collection.",
  "rai:dataUseCases": "Bioacoustic model research, training data construction, and held-out species experiments.",
  "rai:dataSocialImpact": "The dataset may support better animal-sound AI systems, but ecological deployment requires expert validation and safeguards for sensitive species and sites.",
  "rai:hasSyntheticData": false,
  "prov:wasDerivedFrom": [
    {
      "@type": "sc:Dataset",
      "name": "AnimalSpeak"
    }
  ],
  "prov:wasGeneratedBy": [
    {
      "@type": "prov:Activity",
      "@id": "activity-source-separation",
      "name": "Source separation",
      "description": "Original field recordings were source-separated with BirdMixIT to isolate candidate animal vocalization sources.",
      "prov:used": [
        {
          "@type": "prov:SoftwareAgent",
          "name": "BirdMixIT"
        }
      ]
    },
    {
      "@type": "prov:Activity",
      "@id": "activity-window-extraction",
      "name": "Short activity-window extraction",
      "description": "Short vocal activity windows were extracted from the separated sources following the DRASDIC pseudovox procedure.",
      "prov:wasInformedBy": {
        "@type": "sc:CreativeWork",
        "name": "DRASDIC"
      }
    },
    {
      "@type": "prov:Activity",
      "@id": "activity-species-labeling",
      "name": "Species prediction and metadata-match filtering",
      "description": "Each candidate clip was tagged with a species label predicted by sl-efficientnet-bio, then filtered to retain clips whose predicted species matched the recording metadata.",
      "prov:used": [
        {
          "@type": "prov:SoftwareAgent",
          "name": "sl-efficientnet-bio"
        }
      ]
    },
    {
      "@type": "prov:Activity",
      "@id": "activity-birdnet-scoring",
      "name": "BirdNET confidence scoring",
      "description": "BirdNET confidence scores were computed for released clips.",
      "prov:used": [
        {
          "@type": "prov:SoftwareAgent",
          "name": "BirdNET"
        }
      ]
    },
    {
      "@type": "prov:Activity",
      "@id": "activity-hf-export",
      "name": "Hugging Face sharded export",
      "description": "The filtered train-unseen split was packaged as WebDataset tar shards, metadata.parquet, and Croissant JSON-LD for this Hugging Face release."
    }
  ],
  "conformsTo": "http://mlcommons.org/croissant/1.1"
}