VeriSpoof
Collection
Beyond real or fake: one unified audio LLM for speech deepfake forensics. • 4 items • Updated
Datasets:
item_id string | subset string | task string | corpus string | corpus_license string | source_split string | variant string | n_audio int8 | audio_path_a string | audio_path_b string | prompt string | options list | candidates list | label string | label_index int16 | labels string | group string | meta string | reference string | ref_kind string | ref_spans list | ref_token_labels list | ref_measurements dict | ref_condition dict | ref_source_profile dict | audio list |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
authenticity__asvspoof2021df_eval:00000 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2000048.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2000048.flac",
"utt_id": "asvspoof2021df/DF_E_2000048",
"md5": "f0e48ae326844da63934ce9a6c2de011",
"bytes": 48335,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:SPO-N10",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00001 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2000388.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2000388.flac",
"utt_id": "asvspoof2021df/DF_E_2000388",
"md5": "253c91d11117fc0dbff7cf82b3e220b2",
"bytes": 41881,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team31",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00002 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2000475.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2000475.flac",
"utt_id": "asvspoof2021df/DF_E_2000475",
"md5": "52af2c38385fb075896a2119393bd0fb",
"bytes": 39889,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task1-team28",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00003 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2000539.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2000539.flac",
"utt_id": "asvspoof2021df/DF_E_2000539",
"md5": "3ee61394b9dd7cb45c6fbd0d080e2f1b",
"bytes": 33307,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team20",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00004 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2000594.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2000594.flac",
"utt_id": "asvspoof2021df/DF_E_2000594",
"md5": "ccdc410b304859050650f83e42586333",
"bytes": 114169,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-N10",
"acoustic_model": ... |
authenticity__asvspoof2021df_eval:00005 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2000604.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2000604.flac",
"utt_id": "asvspoof2021df/DF_E_2000604",
"md5": "12e53bf107b42a915e8ad6fed022e52f",
"bytes": 63051,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:A11",
"acoustic_model": null,... |
authenticity__asvspoof2021df_eval:00006 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2000714.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2000714.flac",
"utt_id": "asvspoof2021df/DF_E_2000714",
"md5": "3bdd7bf5e6cf1cf6f9134ca990eda50c",
"bytes": 80800,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task1-team23",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00007 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2000767.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2000767.flac",
"utt_id": "asvspoof2021df/DF_E_2000767",
"md5": "07ef55860a80813de5daa786b9d927f7",
"bytes": 30248,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:SPO-N06",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00008 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2000877.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2000877.flac",
"utt_id": "asvspoof2021df/DF_E_2000877",
"md5": "c9a9a3b578cb9b2d476513e5bbf372de",
"bytes": 27241,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-N15",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00009 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2000898.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2000898.flac",
"utt_id": "asvspoof2021df/DF_E_2000898",
"md5": "401e00bfeedc30889dd83281af3b6c30",
"bytes": 70329,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:A11",
"acoustic_model": null,... |
authenticity__asvspoof2021df_eval:00010 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2000994.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2000994.flac",
"utt_id": "asvspoof2021df/DF_E_2000994",
"md5": "8736d478ebe693f9074eb3f7c79035e4",
"bytes": 40857,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-D02",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00011 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2001122.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2001122.flac",
"utt_id": "asvspoof2021df/DF_E_2001122",
"md5": "d69078e19a3d031212bf9ac850b8bb94",
"bytes": 79018,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task1-team13",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00012 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2001423.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2001423.flac",
"utt_id": "asvspoof2021df/DF_E_2001423",
"md5": "b6d5840f094f4d84d893a9c6de632c26",
"bytes": 76887,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team11",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00013 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2001670.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2001670.flac",
"utt_id": "asvspoof2021df/DF_E_2001670",
"md5": "dae64722a86179beee93a9c214ba7879",
"bytes": 89763,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task1-team04",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00014 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2001677.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2001677.flac",
"utt_id": "asvspoof2021df/DF_E_2001677",
"md5": "c9f15ca86d81a288c4fde7e8354b56b9",
"bytes": 32114,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-N07",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00015 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2001984.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2001984.flac",
"utt_id": "asvspoof2021df/DF_E_2001984",
"md5": "b43d7ec3b522c2ccf10a392cab3e3886",
"bytes": 108462,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:SPO-N16",
"acoustic_model": ... |
authenticity__asvspoof2021df_eval:00016 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2002092.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2002092.flac",
"utt_id": "asvspoof2021df/DF_E_2002092",
"md5": "f71c34ed78dbbd775ce3e15fd6eff3b9",
"bytes": 43556,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task1-team04",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00017 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2002884.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | bona fide | 0 | null | null | null | This recording is bona fide: it is genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2002884.flac",
"utt_id": "asvspoof2021df/DF_E_2002884",
"md5": "e7d277036e46da168711e5898f831f8a",
"bytes": 49100,
"authenticity": "bonafide",
"generator_id": null,
"acoustic_model": null,
"vocoder... |
authenticity__asvspoof2021df_eval:00018 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2003442.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2003442.flac",
"utt_id": "asvspoof2021df/DF_E_2003442",
"md5": "2ee7c5bb2c965d30dd3c77bd76407b78",
"bytes": 49119,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team02",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00019 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2003999.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2003999.flac",
"utt_id": "asvspoof2021df/DF_E_2003999",
"md5": "9552b8c04696ca0166119bf849fe518d",
"bytes": 67736,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team03",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00020 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2004240.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2004240.flac",
"utt_id": "asvspoof2021df/DF_E_2004240",
"md5": "194a51334499a58b73752260ee86d37c",
"bytes": 63918,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task1-team12",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00021 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2004466.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2004466.flac",
"utt_id": "asvspoof2021df/DF_E_2004466",
"md5": "98cfbbeea284470b9a12384b4215e7ce",
"bytes": 36053,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:A15",
"acoustic_model": null,... |
authenticity__asvspoof2021df_eval:00022 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2004525.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2004525.flac",
"utt_id": "asvspoof2021df/DF_E_2004525",
"md5": "9f08017d9da5b58bf1b80aab1beb87c6",
"bytes": 37190,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:A13",
"acoustic_model": null,... |
authenticity__asvspoof2021df_eval:00023 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2004894.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2004894.flac",
"utt_id": "asvspoof2021df/DF_E_2004894",
"md5": "ec79db33477935d158339d14bf0b1e77",
"bytes": 44477,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team20",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00024 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2005708.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2005708.flac",
"utt_id": "asvspoof2021df/DF_E_2005708",
"md5": "e70013c3b7df7cfde570461dd49da8fb",
"bytes": 88736,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-N08",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00025 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2005753.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2005753.flac",
"utt_id": "asvspoof2021df/DF_E_2005753",
"md5": "468792a89ea7d420a33f1673de36cff1",
"bytes": 81101,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:SPO-N12",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00026 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2005817.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2005817.flac",
"utt_id": "asvspoof2021df/DF_E_2005817",
"md5": "f54b46143dc133a9bdc74914002a4ce4",
"bytes": 12894,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-N08",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00027 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2007002.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2007002.flac",
"utt_id": "asvspoof2021df/DF_E_2007002",
"md5": "5e26964b2df4685639dd5e616c445e65",
"bytes": 27271,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-N17",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00028 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2007050.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2007050.flac",
"utt_id": "asvspoof2021df/DF_E_2007050",
"md5": "92afe5968d80211ab5b39910bdea2260",
"bytes": 71464,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-N10",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00029 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2007236.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2007236.flac",
"utt_id": "asvspoof2021df/DF_E_2007236",
"md5": "86953cf6c5b77db675c361774490c96d",
"bytes": 60613,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:A17",
"acoustic_model": null,... |
authenticity__asvspoof2021df_eval:00030 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2007501.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2007501.flac",
"utt_id": "asvspoof2021df/DF_E_2007501",
"md5": "5ce8f2c7a52c5b36f3293da4f162b165",
"bytes": 39967,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team28",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00031 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2007735.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2007735.flac",
"utt_id": "asvspoof2021df/DF_E_2007735",
"md5": "f30616ed017badb768e4344a2266b549",
"bytes": 49584,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:A12",
"acoustic_model": null,... |
authenticity__asvspoof2021df_eval:00032 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2007931.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2007931.flac",
"utt_id": "asvspoof2021df/DF_E_2007931",
"md5": "18cd94b5f9137f9fea6b7073c2d88714",
"bytes": 87443,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:SPO-N06",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00033 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2007965.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2007965.flac",
"utt_id": "asvspoof2021df/DF_E_2007965",
"md5": "d9c22b031aec27cb84b3891d4e8b5d25",
"bytes": 61228,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:SPO-N04",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00034 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2008125.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2008125.flac",
"utt_id": "asvspoof2021df/DF_E_2008125",
"md5": "3ed4cdd0a055d4c7bd6d8db15e2220a9",
"bytes": 75970,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:A16",
"acoustic_model": null,... |
authenticity__asvspoof2021df_eval:00035 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2008631.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2008631.flac",
"utt_id": "asvspoof2021df/DF_E_2008631",
"md5": "a43fba419b11e9752066ee7a406d1671",
"bytes": 75047,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team25",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00036 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2008941.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2008941.flac",
"utt_id": "asvspoof2021df/DF_E_2008941",
"md5": "59abee115e2e6a8fcb6316eb813cfd8f",
"bytes": 35963,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:SPO-N05",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00037 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2009527.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2009527.flac",
"utt_id": "asvspoof2021df/DF_E_2009527",
"md5": "53466258e872c505bd6adce687ece6f9",
"bytes": 53665,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team03",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00038 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2009581.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2009581.flac",
"utt_id": "asvspoof2021df/DF_E_2009581",
"md5": "3c262d0185c36c3529affa3aa3dca282",
"bytes": 98649,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-D03",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00039 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2010562.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2010562.flac",
"utt_id": "asvspoof2021df/DF_E_2010562",
"md5": "f7942fc7e176ed11a97923c66ac889f1",
"bytes": 66036,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team31",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00040 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2010611.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2010611.flac",
"utt_id": "asvspoof2021df/DF_E_2010611",
"md5": "1dde265faced5910d490e431b2aa4802",
"bytes": 57096,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-N05",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00041 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2011287.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2011287.flac",
"utt_id": "asvspoof2021df/DF_E_2011287",
"md5": "ad7c171bb7b82d67f98d1028e1062d1c",
"bytes": 73148,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task1-team29",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00042 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2011289.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2011289.flac",
"utt_id": "asvspoof2021df/DF_E_2011289",
"md5": "72cff7128999ed2aa09d282d2a13c186",
"bytes": 51411,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:A07",
"acoustic_model": null,... |
authenticity__asvspoof2021df_eval:00043 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2011396.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2011396.flac",
"utt_id": "asvspoof2021df/DF_E_2011396",
"md5": "8a939a1ab0bd60b91234b0efef0d40b9",
"bytes": 50017,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:SPO-N06",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00044 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2011441.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2011441.flac",
"utt_id": "asvspoof2021df/DF_E_2011441",
"md5": "59da743c71d6c14f7ff99d1ec1edd940",
"bytes": 72251,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team30",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00045 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2011482.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2011482.flac",
"utt_id": "asvspoof2021df/DF_E_2011482",
"md5": "c3b24d561cc3353be6aeb4bd56e68934",
"bytes": 54115,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team22",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00046 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2011539.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2011539.flac",
"utt_id": "asvspoof2021df/DF_E_2011539",
"md5": "62f65a9fb781e9a775aafad9b84ae2f4",
"bytes": 109989,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-D01",
"acoustic_model": ... |
authenticity__asvspoof2021df_eval:00047 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2012161.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2012161.flac",
"utt_id": "asvspoof2021df/DF_E_2012161",
"md5": "af950f7c4bcd7c88e229fb8843361f85",
"bytes": 40300,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team02",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00048 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2012518.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2012518.flac",
"utt_id": "asvspoof2021df/DF_E_2012518",
"md5": "f84a674b0c26dd9098d73c313cfbaca1",
"bytes": 26965,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:SPO-N11",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00049 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2012821.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2012821.flac",
"utt_id": "asvspoof2021df/DF_E_2012821",
"md5": "330b2178c1bf33c78bb4339722414ec0",
"bytes": 71668,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team02",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00050 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2013004.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2013004.flac",
"utt_id": "asvspoof2021df/DF_E_2013004",
"md5": "61c202b84a0b4243753d471132cdb516",
"bytes": 80611,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task1-team14",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00051 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2013306.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2013306.flac",
"utt_id": "asvspoof2021df/DF_E_2013306",
"md5": "98e83dbbe515a645abc47a08e69b79ef",
"bytes": 80314,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:SPO-N11",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00052 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2013561.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | bona fide | 0 | null | null | null | This recording is bona fide: it is genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2013561.flac",
"utt_id": "asvspoof2021df/DF_E_2013561",
"md5": "f0a3594343ccb40201f08858469998f1",
"bytes": 35238,
"authenticity": "bonafide",
"generator_id": null,
"acoustic_model": null,
"vocoder... |
authenticity__asvspoof2021df_eval:00053 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2014166.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2014166.flac",
"utt_id": "asvspoof2021df/DF_E_2014166",
"md5": "da355793e15f31ff70a1b52c149cf340",
"bytes": 59708,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team27",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00054 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2014330.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2014330.flac",
"utt_id": "asvspoof2021df/DF_E_2014330",
"md5": "14d9ec75a237a925f74988831386c682",
"bytes": 93520,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:SPO-N04",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00055 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2014680.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2014680.flac",
"utt_id": "asvspoof2021df/DF_E_2014680",
"md5": "9422f011a5251dbb204dcafab50dde9e",
"bytes": 81571,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task1-team27",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00056 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2014830.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2014830.flac",
"utt_id": "asvspoof2021df/DF_E_2014830",
"md5": "27b375718f8e8eee2c688e770c22e0a0",
"bytes": 21728,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-N11",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00057 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2015029.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2015029.flac",
"utt_id": "asvspoof2021df/DF_E_2015029",
"md5": "2b1b0403d413845e4a3825e1d01ab07e",
"bytes": 36083,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:A07",
"acoustic_model": null,... |
authenticity__asvspoof2021df_eval:00058 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2015846.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2015846.flac",
"utt_id": "asvspoof2021df/DF_E_2015846",
"md5": "9f48b9bdf16e24631b38274fdaeba24f",
"bytes": 89634,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team07",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00059 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2016009.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2016009.flac",
"utt_id": "asvspoof2021df/DF_E_2016009",
"md5": "9d37d8fbed8592c9ce3f720bd13a0a90",
"bytes": 88229,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task1-team33",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00060 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2016120.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2016120.flac",
"utt_id": "asvspoof2021df/DF_E_2016120",
"md5": "14bcb0e8e56c26c326c5e6b839223fc9",
"bytes": 23434,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-N07",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00061 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2016151.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2016151.flac",
"utt_id": "asvspoof2021df/DF_E_2016151",
"md5": "f57d5af1864abd3582512f431ddb6969",
"bytes": 89825,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:A18",
"acoustic_model": null,... |
authenticity__asvspoof2021df_eval:00062 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2017240.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2017240.flac",
"utt_id": "asvspoof2021df/DF_E_2017240",
"md5": "6ac63e74d32fd510261fa9c47071f069",
"bytes": 51014,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team12",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00063 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2017552.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2017552.flac",
"utt_id": "asvspoof2021df/DF_E_2017552",
"md5": "19b4d5548047f294ef81d7b599cc2603",
"bytes": 66035,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-D02",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00064 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2017785.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2017785.flac",
"utt_id": "asvspoof2021df/DF_E_2017785",
"md5": "b19da65c2d2b654d6ee558721227fb41",
"bytes": 28275,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-N09",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00065 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2018147.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2018147.flac",
"utt_id": "asvspoof2021df/DF_E_2018147",
"md5": "1a55f6a878c5ba4b6640c78516bc02d2",
"bytes": 23010,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-N15",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00066 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2018571.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2018571.flac",
"utt_id": "asvspoof2021df/DF_E_2018571",
"md5": "ecf77c94085c72f03b7ec8c2c5f03e69",
"bytes": 55101,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team15",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00067 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2019043.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2019043.flac",
"utt_id": "asvspoof2021df/DF_E_2019043",
"md5": "f98d8a5df84f48d08510b6290b7b38c2",
"bytes": 78790,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team06",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00068 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2019162.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2019162.flac",
"utt_id": "asvspoof2021df/DF_E_2019162",
"md5": "72faee08070c3194a0d1b5b985c30269",
"bytes": 68743,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task1-team20",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00069 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2019163.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2019163.flac",
"utt_id": "asvspoof2021df/DF_E_2019163",
"md5": "2662ef29b0b59d442044465e409d52b9",
"bytes": 41097,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task1-team14",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00070 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2019402.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2019402.flac",
"utt_id": "asvspoof2021df/DF_E_2019402",
"md5": "3947e836239a72007d731190ad3e1b8a",
"bytes": 78554,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task1-team20",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00071 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2019723.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2019723.flac",
"utt_id": "asvspoof2021df/DF_E_2019723",
"md5": "dbb86563afa656a3af71cb049d9860ab",
"bytes": 62122,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team15",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00072 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2020056.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2020056.flac",
"utt_id": "asvspoof2021df/DF_E_2020056",
"md5": "1703fdfeade09de2468f7a8275183ecb",
"bytes": 75163,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:SPO-N06",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00073 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2020361.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2020361.flac",
"utt_id": "asvspoof2021df/DF_E_2020361",
"md5": "6648d6d03982d38d2de14aa49289667c",
"bytes": 70677,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team18",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00074 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2020651.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2020651.flac",
"utt_id": "asvspoof2021df/DF_E_2020651",
"md5": "6519fde16e1deab548b02d66e51f3d95",
"bytes": 38570,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team02",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00075 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2021070.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2021070.flac",
"utt_id": "asvspoof2021df/DF_E_2021070",
"md5": "b4c64d5810d1272aae4c8c1a7d2af861",
"bytes": 47583,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-D02",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00076 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2021310.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2021310.flac",
"utt_id": "asvspoof2021df/DF_E_2021310",
"md5": "289998532947b680b014d024d7dc72db",
"bytes": 85771,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team13",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00077 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2021724.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2021724.flac",
"utt_id": "asvspoof2021df/DF_E_2021724",
"md5": "f0d6b9ea64d80192a6896718d08ae1ec",
"bytes": 26843,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-N14",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00078 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2021771.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2021771.flac",
"utt_id": "asvspoof2021df/DF_E_2021771",
"md5": "6c85b588fe89f806e54489eb93466b00",
"bytes": 50688,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:SPO-B01",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00079 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2021914.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2021914.flac",
"utt_id": "asvspoof2021df/DF_E_2021914",
"md5": "ce21733377f788024453d0eb175e2e72",
"bytes": 39478,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task1-team26",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00080 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2022484.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2022484.flac",
"utt_id": "asvspoof2021df/DF_E_2022484",
"md5": "316d9252360020abefde9565eb63181b",
"bytes": 67730,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team09",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00081 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2023312.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2023312.flac",
"utt_id": "asvspoof2021df/DF_E_2023312",
"md5": "a6d755bec2d2a5b5cd0df9165d8d8233",
"bytes": 77358,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task1-team13",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00082 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2023953.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2023953.flac",
"utt_id": "asvspoof2021df/DF_E_2023953",
"md5": "cb382689abc32cb1eeeb57a32d8e4f75",
"bytes": 18777,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-N18",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00083 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2023990.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2023990.flac",
"utt_id": "asvspoof2021df/DF_E_2023990",
"md5": "b04f15e171153f8d033c3a99c27072a7",
"bytes": 45610,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:A10",
"acoustic_model": null,... |
authenticity__asvspoof2021df_eval:00084 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2024110.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2024110.flac",
"utt_id": "asvspoof2021df/DF_E_2024110",
"md5": "7ff84db7a3a5856848bc4221bb7468df",
"bytes": 51375,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team11",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00085 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2024153.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2024153.flac",
"utt_id": "asvspoof2021df/DF_E_2024153",
"md5": "5652d9d469b5d57aa3efba6d28c31b98",
"bytes": 107886,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:A17",
"acoustic_model": null... |
authenticity__asvspoof2021df_eval:00086 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2024199.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2024199.flac",
"utt_id": "asvspoof2021df/DF_E_2024199",
"md5": "1a194130d202df910812413f4644fe3b",
"bytes": 65321,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team16",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00087 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2024582.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2024582.flac",
"utt_id": "asvspoof2021df/DF_E_2024582",
"md5": "b1698744bb67df32ace657200c18b918",
"bytes": 72650,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-N04",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00088 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2024623.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2024623.flac",
"utt_id": "asvspoof2021df/DF_E_2024623",
"md5": "d3b2cfe9b83c7f7fe958a083ec9f093e",
"bytes": 25887,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-N08",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00089 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2025050.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2025050.flac",
"utt_id": "asvspoof2021df/DF_E_2025050",
"md5": "461041243bb1ca04ec80f19f0f5b13cc",
"bytes": 92004,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:SPO-N17",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00090 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2025191.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2025191.flac",
"utt_id": "asvspoof2021df/DF_E_2025191",
"md5": "2157214a3b6de081c91343d174dd847a",
"bytes": 28854,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:A08",
"acoustic_model": null,... |
authenticity__asvspoof2021df_eval:00091 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2025440.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2025440.flac",
"utt_id": "asvspoof2021df/DF_E_2025440",
"md5": "361af7fbbc45224e3dc0571e0cae569c",
"bytes": 65075,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:Task2-team27",
"acoustic_mode... |
authenticity__asvspoof2021df_eval:00092 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2025791.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2025791.flac",
"utt_id": "asvspoof2021df/DF_E_2025791",
"md5": "bfabcead04587afbf38c7ae3ed1a73af",
"bytes": 59706,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-N18",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00093 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2025972.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2025972.flac",
"utt_id": "asvspoof2021df/DF_E_2025972",
"md5": "ada364913a4f84e7a872b6719973a37d",
"bytes": 44823,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-D04",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00094 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2026228.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2026228.flac",
"utt_id": "asvspoof2021df/DF_E_2026228",
"md5": "647440d2548f46f30ca437cd0623fcfc",
"bytes": 42139,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-N07",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00095 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2026430.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2026430.flac",
"utt_id": "asvspoof2021df/DF_E_2026430",
"md5": "6bf5c432981940b4d5fd6f958c5bfecb",
"bytes": 118892,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-N09",
"acoustic_model": ... |
authenticity__asvspoof2021df_eval:00096 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2026812.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2026812.flac",
"utt_id": "asvspoof2021df/DF_E_2026812",
"md5": "44cfba007e62f340abcf1658f80f8d05",
"bytes": 46510,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:SPO-N11",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00097 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2026832.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2026832.flac",
"utt_id": "asvspoof2021df/DF_E_2026832",
"md5": "c43ce9e7dd3cd8ac32258b413449f021",
"bytes": 57004,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:SPO-N12",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00098 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2027379.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2027379.flac",
"utt_id": "asvspoof2021df/DF_E_2027379",
"md5": "9e81b7c935c5a25ddd24f27bcf2dc800",
"bytes": 66831,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-N16",
"acoustic_model": n... |
authenticity__asvspoof2021df_eval:00099 | authenticity__asvspoof2021df_eval | authenticity | ASVspoof 2021 DF | ODbL 1.0 (Zenodo 4835108) or ODC-By (asvspoof.org): conflicting; the stricter ODbL (share-alike) is followed | eval | clean | 1 | ASVspoof2021/ASVspoof2021_DF_eval/flac/DF_E_2027490.flac | null | <audio>
Is this recording genuine human speech or spoofed speech? Options: bona fide; spoof. | [
"bona fide",
"spoof"
] | [
"This recording is bona fide: it is genuine human speech.",
"This recording is a spoof: it is synthetic or converted speech, not genuine human speech."
] | spoof | 1 | null | null | null | This recording is a spoof: it is synthetic or converted speech, not genuine human speech. | null | null | null | null | null | null | [
{
"role": "A",
"root": "ASVspoof2021",
"path": "ASVspoof2021_DF_eval/flac/DF_E_2027490.flac",
"utt_id": "asvspoof2021df/DF_E_2027490",
"md5": "3dc80ef050a405220af273c14b3ee9a1",
"bytes": 42105,
"authenticity": "spoof",
"generator_id": "asvspoof2021df:HUB-N05",
"acoustic_model": n... |
End of preview. Expand in Data Studio
VeriSpoof-Bench
The evaluation suite of VeriSpoof: speech-deepfake forensics beyond a real/fake answer. It covers detection under channel and post-processing conditions, generator attribution, source verification (do two spoofs come from the same system?), recording-condition recognition, and grounded reasoning whose lines can each be checked against the signal and the labels.
No audio is included. The repository holds item lists, labels, prompts, reference answers, the parameters that regenerate every derived item, scoring scripts and reference results. Get each corpus from its distributor; the scripts find the files in your copy and rebuild the derived items exactly.
| Tech report | VeriSpoof: Beyond Real or Fake with One Audio LLM, https://arxiv.org/abs/2611.04634 |
| Models | https://huggingface.co/AustinXiao/VeriSpoof-Plus, https://huggingface.co/AustinXiao/VeriSpoof-Base |
| Code · Demo · Project page | https://github.com/swagshaw/verispoof · https://huggingface.co/spaces/AustinXiao/VeriSpoof-demo · https://swagshaw.github.io/verispoof/ |
At a glance
54 subsets, 177,852 items, in six groups (each group and each subset is a config):
| Group | Task | Subsets | Items |
|---|---|---|---|
detection |
bona fide vs. spoof, clean and under simulated channel and post-processing conditions | 16 | 95,852 |
attribution |
generator identity (closed set), acoustic model under an unseen vocoder | 3 | 10,000 |
source_verification |
same generator / acoustic model / vocoder for two spoofed recordings | 13 | 43,000 |
recording_condition |
noise level, reverberation, codec class | 6 | 16,000 |
grounded_reasoning |
detection reasoning and faithfulness interventions | 12 | 6,000 |
sv_reasoning |
source-verification reasoning over pairs | 4 | 7,000 |
Sources: ASVspoof 2019 LA / 2021 LA and DF / 5, In-the-Wild, SpoofCeleb, MLAAD, STOPA, DFADD, LibriSeVoc, CodecFake+, PartialSpoof, HAD, PartialEdit and LlamaPartialSpoof. Subset construction and metrics are described in the tech report.
from datasets import load_dataset
det = load_dataset("AustinXiao/VeriSpoof-Bench", "authenticity__la19_eval", split="test")
grp = load_dataset("AustinXiao/VeriSpoof-Bench", "source_verification", split="test") # a whole group
print(det[0]["prompt"], det[0]["options"], det[0]["label"], det[0]["audio_path_a"])
Key fields: item_id, subset, task, prompt (the question; <audio> marks each recording), options /
candidates / label, audio_path_a / audio_path_b (paths in the distributor's layout), reference (reference
reasoning text) and audio (per-recording labels, md5, and for derived items the full realization).
Quick start
- Get the audio. Download the corpora you need and keep their layout:
Derived subsets also need MUSAN (https://www.openslr.org/17/) and RIRS_NOISES (https://www.openslr.org/28/).
- Prepare, run and score:
pip install numpy scipy pandas pyarrow soundfile librosa "git+https://github.com/swagshaw/verispoof"
python scripts/prepare_audio.py --roots LA19=/corpora/ASVspoof2019 InTheWild=/corpora/itw ... \
--musan /corpora/musan --rirs /corpora/RIRS_NOISES --out bench_audio # add --check --md5 to verify your copy
python scripts/run_verispoof.py --model AustinXiao/VeriSpoof-Plus --manifest bench_audio/manifest.csv --out preds/plus
python scripts/score.py --pred preds/plus --out preds/plus/metrics.json
Your own model: write one file per subset (<pred_dir>/<subset>.parquet|jsonl|csv, keyed by item_id) with
scores (one per option) or score for closed tasks and output (reasoning text in the format of reference) for
reasoning tasks, then run scripts/score.py --pred <pred_dir>.
Reference results
| VeriSpoof-Base | VeriSpoof-Plus | |
|---|---|---|
| Detection, mean EER over 16 subsets | 3.17% | 2.36% |
| Source verification, mean EER over 13 subsets | 23.84% | 19.21% |
| Recording condition, mean accuracy over 6 subsets | 75.9% | 81.3% |
Per-item outputs of both models are in reference_results/, all metrics in reference_results/*/reported_metrics.json;
python scripts/check_reference.py re-scores them. Per-subset results are in the tech report.
Licences
No audio is redistributed. Each subset carries the terms of its source corpus (column corpus_license, see
LICENSE.md): MLAAD subsets are non-commercial, LibriSeVoc, In-the-Wild and ASVspoof 2021 DF subsets share-alike,
the rest attribution or MIT. Scripts: CC BY-NC 4.0.
Citation
@misc{verispoof2026,
title = {{VeriSpoof}: Beyond Real or Fake with One Audio {LLM}},
author = {Xiao, Yang and Dang, Ting},
year = {2026},
eprint = {2611.04634},
archivePrefix = {arXiv},
primaryClass = {eess.AS},
url = {https://arxiv.org/abs/2611.04634}
}
Please also cite the source corpora of the subsets you use.
- Downloads last month
- -