LanceaKing/asvspoof2019
收藏资源简介:
--- annotations_creators: - other language_creators: - other language: - en license: - odc-by multilinguality: - monolingual size_categories: - 100K<n<1M source_datasets: - extended|vctk task_categories: - audio-classification task_ids: [] pretty_name: asvspoof2019 tags: - voice-anti-spoofing --- # Dataset Card for asvspoof2019 ## Table of Contents - [Dataset Description](#dataset-description) - [Dataset Summary](#dataset-summary) - [Supported Tasks](#supported-tasks-and-leaderboards) - [Languages](#languages) - [Dataset Structure](#dataset-structure) - [Data Instances](#data-instances) - [Data Fields](#data-instances) - [Data Splits](#data-instances) - [Dataset Creation](#dataset-creation) - [Curation Rationale](#curation-rationale) - [Source Data](#source-data) - [Annotations](#annotations) - [Personal and Sensitive Information](#personal-and-sensitive-information) - [Considerations for Using the Data](#considerations-for-using-the-data) - [Social Impact of Dataset](#social-impact-of-dataset) - [Discussion of Biases](#discussion-of-biases) - [Other Known Limitations](#other-known-limitations) - [Additional Information](#additional-information) - [Dataset Curators](#dataset-curators) - [Licensing Information](#licensing-information) - [Citation Information](#citation-information) ## Dataset Description - **Homepage:** https://datashare.ed.ac.uk/handle/10283/3336 - **Repository:** [Needs More Information] - **Paper:** https://arxiv.org/abs/1911.01601 - **Leaderboard:** [Needs More Information] - **Point of Contact:** [Needs More Information] ### Dataset Summary This is a database used for the Third Automatic Speaker Verification Spoofing and Countermeasuers Challenge, for short, ASVspoof 2019 (http://www.asvspoof.org) organized by Junichi Yamagishi, Massimiliano Todisco, Md Sahidullah, Héctor Delgado, Xin Wang, Nicholas Evans, Tomi Kinnunen, Kong Aik Lee, Ville Vestman, and Andreas Nautsch in 2019. ### Supported Tasks and Leaderboards [Needs More Information] ### Languages English ## Dataset Structure ### Data Instances ``` {'speaker_id': 'LA_0091', 'audio_file_name': 'LA_T_8529430', 'audio': {'path': 'D:/Users/80304531/.cache/huggingface/datasets/downloads/extracted/8cabb6d5c283b0ed94b2219a8d459fea8e972ce098ef14d8e5a97b181f850502/LA/ASVspoof2019_LA_train/flac/LA_T_8529430.flac', 'array': array([-0.00201416, -0.00234985, -0.0022583 , ..., 0.01309204, 0.01339722, 0.01461792], dtype=float32), 'sampling_rate': 16000}, 'system_id': 'A01', 'key': 1} ``` ### Data Fields Logical access (LA): - `speaker_id`: `LA_****`, a 4-digit speaker ID - `audio_file_name`: name of the audio file - `audio`: A dictionary containing the path to the downloaded audio file, the decoded audio array, and the sampling rate. Note that when accessing the audio column: `dataset[0]["audio"]` the audio file is automatically decoded and resampled to `dataset.features["audio"].sampling_rate`. Decoding and resampling of a large number of audio files might take a significant amount of time. Thus it is important to first query the sample index before the `"audio"` column, *i.e.* `dataset[0]["audio"]` should **always** be preferred over `dataset["audio"][0]`. - `system_id`: ID of the speech spoofing system (A01 - A19), or, for bonafide speech SYSTEM-ID is left blank ('-') - `key`: 'bonafide' for genuine speech, or, 'spoof' for spoofing speech Physical access (PA): - `speaker_id`: `PA_****`, a 4-digit speaker ID - `audio_file_name`: name of the audio file - `audio`: A dictionary containing the path to the downloaded audio file, the decoded audio array, and the sampling rate. Note that when accessing the audio column: `dataset[0]["audio"]` the audio file is automatically decoded and resampled to `dataset.features["audio"].sampling_rate`. Decoding and resampling of a large number of audio files might take a significant amount of time. Thus it is important to first query the sample index before the `"audio"` column, *i.e.* `dataset[0]["audio"]` should **always** be preferred over `dataset["audio"][0]`. - `environment_id`: a triplet (S,R,D_s), which take one letter in the set {a,b,c} as categorical value, defined as | | a | b | c | | -------------------------------- | ------ | ------- | -------- | | S: Room size (square meters) | 2-5 | 5-10 | 10-20 | | R: T60 (ms) | 50-200 | 200-600 | 600-1000 | | D_s: Talker-to-ASV distance (cm) | 10-50 | 50-100 | 100-150 | - `attack_id`: a duple (D_a,Q), which take one letter in the set {A,B,C} as categorical value, defined as | | A | B | C | | ----------------------------------- | ------- | ------ | ----- | | Z: Attacker-to-talker distance (cm) | 10-50 | 50-100 | > 100 | | Q: Replay device quality | perfect | high | low | for bonafide speech, `attack_id` is left blank ('-') - `key`: 'bonafide' for genuine speech, or, 'spoof' for spoofing speech ### Data Splits | | Training set | Development set | Evaluation set | | -------- | ------------ | --------------- | -------------- | | Bonafide | 2580 | 2548 | 7355 | | Spoof | 22800 | 22296 | 63882 | | Total | 25380 | 24844 | 71237 | ## Dataset Creation ### Curation Rationale [Needs More Information] ### Source Data #### Initial Data Collection and Normalization [Needs More Information] #### Who are the source language producers? [Needs More Information] ### Annotations #### Annotation process [Needs More Information] #### Who are the annotators? [Needs More Information] ### Personal and Sensitive Information [Needs More Information] ## Considerations for Using the Data ### Social Impact of Dataset [Needs More Information] ### Discussion of Biases [Needs More Information] ### Other Known Limitations [Needs More Information] ## Additional Information ### Dataset Curators [Needs More Information] ### Licensing Information This ASVspoof 2019 dataset is made available under the Open Data Commons Attribution License: http://opendatacommons.org/licenses/by/1.0/ ### Citation Information ``` @InProceedings{Todisco2019, Title = {{ASV}spoof 2019: {F}uture {H}orizons in {S}poofed and {F}ake {A}udio {D}etection}, Author = {Todisco, Massimiliano and Wang, Xin and Sahidullah, Md and Delgado, H ́ector and Nautsch, Andreas and Yamagishi, Junichi and Evans, Nicholas and Kinnunen, Tomi and Lee, Kong Aik}, booktitle = {Proc. of Interspeech 2019}, Year = {2019} } ```
--- annotations_creators: - 其他 language_creators: - 其他 language: - 英语 license: - 开放数据 Commons 署名许可(odc-by) multilinguality: - 单语言 size_categories: - 100K<n<1M source_datasets: - 扩展|vctk task_categories: - 音频分类 task_ids: [] pretty_name: asvspoof2019 tags: - 语音反欺骗 --- # ASVspoof 2019 数据集卡片 ## 目录 - [数据集描述](#dataset-description) - [数据集摘要](#dataset-summary) - [支持任务与排行榜](#supported-tasks-and-leaderboards) - [语言](#languages) - [数据集结构](#dataset-structure) - [数据实例](#data-instances) - [数据字段](#data-fields) - [数据划分](#data-splits) - [数据集构建](#dataset-creation) - [数据集构建初衷](#curation-rationale) - [源数据](#source-data) - [注释信息](#annotations) - [个人与敏感信息](#personal-and-sensitive-information) - [数据集使用注意事项](#considerations-for-using-the-data) - [数据集的社会影响](#social-impact-of-dataset) - [偏差讨论](#discussion-of-biases) - [其他已知局限](#other-known-limitations) - [附加信息](#additional-information) - [数据集维护者](#dataset-curators) - [许可信息](#licensing-information) - [引用信息](#citation-information) ## 数据集描述 - **主页**: https://datashare.ed.ac.uk/handle/10283/3336 - **代码仓库**: [待补充] - **论文**: https://arxiv.org/abs/1911.01601 - **排行榜**: [待补充] - **联系人**: [待补充] ### 数据集摘要 本数据集用于第三届自动语音验证欺骗与对抗挑战赛事(简称ASVspoof 2019,http://www.asvspoof.org),该赛事由Junichi Yamagishi、Massimiliano Todisco、Md Sahidullah、Héctor Delgado、Xin Wang、Nicholas Evans、Tomi Kinnunen、Kong Aik Lee、Ville Vestman及Andreas Nautsch于2019年组织。 ### 支持任务与排行榜 [待补充] ### 语言 英语 ## 数据集结构 ### 数据实例 {'speaker_id': 'LA_0091', 'audio_file_name': 'LA_T_8529430', 'audio': {'path': 'D:/Users/80304531/.cache/huggingface/datasets/downloads/extracted/8cabb6d5c283b0ed94b2219a8d459fea8e972ce098ef14d8e5a97b181f8529430/LA/ASVspoof2019_LA_train/flac/LA_T_8529430.flac', 'array': array([-0.00201416, -0.00234985, -0.0022583 , ..., 0.01309204, 0.01339722, 0.01461792], dtype=float32), 'sampling_rate': 16000}, 'system_id': 'A01', 'key': 1} ### 数据字段 #### 逻辑访问(LA)子集: - `speaker_id`: 格式为`LA_****`,为4位说话人标识符 - `audio_file_name`: 音频文件名 - `audio`: 包含音频文件路径、解码后的音频数组及采样率的字典。请注意,当访问音频列时:`dataset[0]["audio"]`会自动对音频文件进行解码并重采样至`dataset.features["audio"].sampling_rate`指定的采样率。解码与重采样大量音频文件可能耗费较长时间,因此建议优先通过样本索引访问音频列,即**始终优先使用`dataset[0]["audio"]`而非`dataset["audio"][0]`**。 - `system_id`: 语音欺骗系统的标识符(取值范围为A01至A19);对于真实语音样本,该字段留空(值为`'-'`) - `key`: 真实语音样本标注为`bonafide`,欺骗语音样本标注为`spoof` #### 物理访问(PA)子集: - `speaker_id`: 格式为`PA_****`,为4位说话人标识符 - `audio_file_name`: 音频文件名 - `audio`: 包含音频文件路径、解码后的音频数组及采样率的字典。请注意,当访问音频列时:`dataset[0]["audio"]`会自动对音频文件进行解码并重采样至`dataset.features["audio"].sampling_rate`指定的采样率。解码与重采样大量音频文件可能耗费较长时间,因此建议优先通过样本索引访问音频列,即**始终优先使用`dataset[0]["audio"]`而非`dataset["audio"][0]`**。 - `environment_id`: 三元组`(S,R,D_s)`,取值为集合`{a,b,c}`中的单字符类别值,定义如下: | | a | b | c | | -------------------------------- | ------ | ------- | -------- | | S: 房间面积(平方米) | 2-5 | 5-10 | 10-20 | | R: 混响时间T60(毫秒) | 50-200 | 200-600 | 600-1000 | | D_s: 说话人与自动语音验证系统的距离(厘米) | 10-50 | 50-100 | 100-150 | - `attack_id`: 二元组`(D_a,Q)`,取值为集合`{A,B,C}`中的单字符类别值,定义如下: | | A | B | C | | ----------------------------------- | ------- | ------ | ----- | | Z: 攻击者与说话人的距离(厘米) | 10-50 | 50-100 | > 100 | | Q: 回放设备质量 | 完美 | 高 | 低 | 对于真实语音样本,`attack_id`字段留空(值为`'-'`) - `key`: 真实语音样本标注为`bonafide`,欺骗语音样本标注为`spoof` ### 数据划分 | | 训练集 | 开发集 | 评估集 | | -------- | ------ | ------ | ------ | | 真实语音 | 2580 | 2548 | 7355 | | 欺骗语音 | 22800 | 22296 | 63882 | | 总计 | 25380 | 24844 | 71237 | ## 数据集构建 ### 数据集构建初衷 [待补充] ### 源数据 #### 初始数据收集与归一化 [待补充] #### 源语言生成者是谁? [待补充] ### 注释信息 #### 注释流程 [待补充] #### 注释人员是谁? [待补充] ### 个人与敏感信息 [待补充] ## 数据集使用注意事项 ### 数据集的社会影响 [待补充] ### 偏差讨论 [待补充] ### 其他已知局限 [待补充] ## 附加信息 ### 数据集维护者 [待补充] ### 许可信息 本ASVspoof 2019数据集采用开放数据 Commons 署名许可(odc-by)发布,详情见:http://opendatacommons.org/licenses/by/1.0/ ### 引用信息 @InProceedings{Todisco2019, Title = {{ASV}spoof 2019: {F}uture {H}orizons in {S}poofed and {F}ake {A}udio {D}etection}, Author = {Todisco, Massimiliano and Wang, Xin and Sahidullah, Md and Delgado, H ́ector and Nautsch, Andreas and Yamagishi, Junichi and Evans, Nicholas and Kinnunen, Tomi and Lee, Kong Aik}, booktitle = {Proc. of Interspeech 2019}, Year = {2019} }
数据集概述
数据集摘要
该数据集用于第三届自动说话人验证欺骗与对策挑战赛(ASVspoof 2019),由Junichi Yamagishi等人组织于2019年。
支持的任务和排行榜
[需要更多信息]
语言
英语
数据集结构
数据实例
json { "speaker_id": "LA_0091", "audio_file_name": "LA_T_8529430", "audio": { "path": "D:/Users/80304531/.cache/huggingface/datasets/downloads/extracted/8cabb6d5c283b0ed94b2219a8d459fea8e972ce098ef14d8e5a97b181f850502/LA/ASVspoof2019_LA_train/flac/LA_T_8529430.flac", "array": array([-0.00201416, -0.00234985, -0.0022583 , ..., 0.01309204, 0.01339722, 0.01461792], dtype=float32), "sampling_rate": 16000 }, "system_id": "A01", "key": 1 }
数据字段
逻辑访问(LA):
speaker_id:LA_****,4位说话人IDaudio_file_name: 音频文件名audio: 包含下载音频文件路径、解码音频数组和采样率的字典system_id: 语音欺骗系统ID(A01 - A19),或真实语音的SYSTEM-ID为空(-)key: bonafide表示真实语音,spoof表示欺骗语音
物理访问(PA):
speaker_id:PA_****,4位说话人IDaudio_file_name: 音频文件名audio: 包含下载音频文件路径、解码音频数组和采样率的字典environment_id: 三元组(S,R,D_s),取集合{a,b,c}中的一个字母作为分类值attack_id: 二元组(D_a,Q),取集合{A,B,C}中的一个字母作为分类值key: bonafide表示真实语音,spoof表示欺骗语音
数据分割
| 训练集 | 开发集 | 评估集 | |
|---|---|---|---|
| 真实语音 | 2580 | 2548 | 7355 |
| 欺骗语音 | 22800 | 22296 | 63882 |
| 总计 | 25380 | 24844 | 71237 |
数据集创建
策划理由
[需要更多信息]
源数据
初始数据收集和规范化
[需要更多信息]
源语言生产者
[需要更多信息]
标注
标注过程
[需要更多信息]
标注者
[需要更多信息]
个人和敏感信息
[需要更多信息]
使用数据的注意事项
数据集的社会影响
[需要更多信息]
偏见的讨论
[需要更多信息]
其他已知限制
[需要更多信息]
附加信息
数据集策展人
[需要更多信息]
许可信息
该ASVspoof 2019数据集在Open Data Commons Attribution License下发布:http://opendatacommons.org/licenses/by/1.0/
引用信息
@InProceedings{Todisco2019, Title = {{ASV}spoof 2019: {F}uture {H}orizons in {S}poofed and {F}ake {A}udio {D}etection}, Author = {Todisco, Massimiliano and Wang, Xin and Sahidullah, Md and Delgado, H ́ector and Nautsch, Andreas and Yamagishi, Junichi and Evans, Nicholas and Kinnunen, Tomi and Lee, Kong Aik}, booktitle = {Proc. of Interspeech 2019}, Year = {2019} }




