Skip to content

Commit

Permalink
Merge pull request #61 from SuperKogito/add-more-datasets
Browse files Browse the repository at this point in the history 
Add more datasets
  • Loading branch information
@SuperKogito
SuperKogito authored Jun 23, 2024
2 parents d636434 + 7459c59 commit 54706b7
Show file tree
Hide file tree
Showing 2 changed files with 255 additions and 47 deletions.
8 changes: 3 additions & 5 deletions CONTRIBUTING.md
View file
Original file line number Diff line number Diff line change
Expand Up @@ -8,22 +8,20 @@ please feel free to add it.

* Read the criteria below
* Add one dataset per Pull Request.
* Add the dataset and its information as a table row.
* Make sure to maintain the table chronological order.
* Add the dataset and its information as a json entry.
* Make sure to maintain the json chronological order.
* Keep descriptions in the content and emotion cells concise.
* Check your spelling and grammar.
* Keep the table correctly aligned.
* Send a Pull Request with a short explanation of why does it belong to this list.

## Criteria

* The dataset must be related to the field of *Spoken Emotion Recognition*.
* The dataset must be related to the field of *Speech Emotion Recognition*.
* The dataset should not be a duplicate.
* The dataset should not be provided in an active PR.
* The dataset should be available for researchers for free.
* The information about the dataset must be accessible for verification.

## How to contribute
First go to `src/` using `cd src`. Then add a the dictionary / part json data of the contributed dataset to `src/ser-datasets`.
Make sure the json is valid, then run the `python generate_files.py` to update the restructured text file, csv file and the README.
That's it, Congrats! and thank you for your contribution. Now open a PR with your changes. I will review it and then publish the results :))
294 changes: 252 additions & 42 deletions src/ser-datasets.json
View file
Original file line number Diff line number Diff line change
@@ -1,4 +1,60 @@
{
"nEmo": {
"Year": 2024,
"Content": "3 hours of samples recorded with the participation of nine actors.",
"Emotions": "6 emotions: anger, fear, happiness, sadness, surprised, and neutral.",
"Format": "Audio",
"Size": "0.434 GB",
"Language": "Polish",
"Paper": "nEMO: Dataset of Emotional Speech in Polish",
"Access": "Open",
"License": "CC BY 4.0",
"Dataset-link": "https://github.com/amu-cai/nEMO",
"Paper-link": "https://arxiv.org/abs/2404.06292",
"License-link": "https://creativecommons.org/licenses/by/4.0/"
},
"EMOVOME": {
"Year": 2024,
"Content": "999 spontaneous voice messages from 100 Spanish speakers, collected from real conversations on a messaging app.",
"Emotions": "Valence & arrousal dimensions and 7 emotions: happiness, disgust, anger, surprise, fear, sadness, and neutral.",
"Format": "Audio",
"Size": "--",
"Language": "Spanish",
"Paper": "EMOVOME Database: Advancing Emotion Recognition in Speech Beyond Staged Scenarios",
"Access": "Partially open",
"License": "CC BY 4.0",
"Dataset-link": "https://zenodo.org/records/10694370",
"Paper-link": "https://arxiv.org/abs/2403.02167",
"License-link": "https://creativecommons.org/licenses/by/4.0/"
},
"EMNS": {
"Year": 2023,
"Content": "1206 high quality labeled utterances by one female speaker (2-3 hours).",
"Emotions": "Anger, excitement, disgust, happiness, surprise, sadness, and neutral (plus sarcasm)",
"Format": "Audio",
"Size": "0.042 GB",
"Language": "English (British)",
"Paper": "EMNS /Imz/ Corpus: An emotive single-speaker dataset for narrative storytelling in games, television and graphic novels",
"Access": "Open",
"License": "Apache 2.0",
"Dataset-link": "http://www.openslr.org/136/",
"Paper-link": "https://arxiv.org/abs/2305.13137",
"License-link": "https://apache.org/licenses/LICENSE-2.0"
},
"CAVES": {
"Year": 2023,
"Content": "Full hd visual recordings of 10 native cantonese speakers uttering 50 sentences.",
"Emotions": "Anger, happiness, sadness, surprise, fear, disgust and neutral",
"Format": "Audio",
"Size": "47 GB",
"Language": "Chinese (cantonese)",
"Paper": "A Cantonese Audio-Visual Emotional Speech (CAVES) dataset",
"Access": "Open",
"License": "Available for research purposes only",
"Dataset-link": "https://rds.westernsydney.edu.au/Institutes/MARCS/2024/Christopher_Davis/",
"Paper-link": "https://link.springer.com/article/10.3758/s13428-023-02270-7",
"License-link": "--"
},
"BANSpEmo": {
"Year": 2023,
"Content": "792 utterance recordings from 22 unprofessional speakers (11 males and 11 females) of six basic emotional reactions of two sets of sentences.",
Expand Down Expand Up @@ -68,6 +124,20 @@
"Dataset-link": "https://data.mendeley.com/datasets/t9h6p943xy/5",
"Paper-link": "https://www.sciencedirect.com/science/article/pii/S235234092200302X",
"License-link": "https://creativecommons.org/licenses/by/4.0/"
},
"B-SER": {
"Year": 2022,
"Content": "1224 speech-audio recordings by 34 non-professional participating actors (17 male and 17 female) from diverse age groups between 19 and 47 years.",
"Emotions": "angry, happy, sad and surprise",
"Format": "Audio",
"Size": "0.363 GB",
"Language": "Bangla",
"Paper": "--",
"Access": "Open",
"License": "CC BY 4.0",
"Dataset-link": "https://data.mendeley.com/datasets/t9h6p943xy/3",
"Paper-link": "--",
"License-link": "https://creativecommons.org/licenses/by/4.0/"
},
"Kannada": {
"Year": 2022,
Expand All @@ -83,6 +153,104 @@
"Paper-link": "--",
"License-link": "https://creativecommons.org/licenses/by/4.0/"
},
"Quechua-SER": {
"Year": 2022,
"Content": "12420 audio recordings (~15 hours) and their transcriptions by 7 native speakers.",
"Emotions": "Emotional labels using dimensions: valence, arousal, and dominance.",
"Format": "Audio",
"Size": "3.53 GB",
"Language": "Quechua Collao",
"Paper": "A speech corpus of Quechua Collao for automatic dimensional emotion recognition",
"Access": "Open",
"License": "CC BY 4.0",
"Dataset-link": "https://figshare.com/articles/media/Quechua_Collao_for_Speech_Emotion_Recognition/20292516",
"Paper-link": "https://www.nature.com/articles/s41597-022-01855-9",
"License-link": "https://creativecommons.org/licenses/by/4.0/"
},
"MESD": {
"Year": 2022,
"Content": "864 audio files of single-word emotional utterances with Mexican cultural shaping.",
"Emotions": "6 emotions provides single-word utterances for anger, disgust, fear, happiness, neutral, and sadness.",
"Format": "Audio",
"Size": "0.097 GB",
"Language": "Spanish (Mexican)",
"Paper": "The Mexican Emotional Speech Database (MESD): elaboration and assessment based on machine learning",
"Access": "Open",
"License": "CC BY 4.0",
"Dataset-link": "https://data.mendeley.com/datasets/cy34mh68j9/5",
"Paper-link": "https://pubmed.ncbi.nlm.nih.gov/34891601/",
"License-link": "https://creativecommons.org/licenses/by/4.0/"
},
"SyntAct": {
"Year": 2022,
"Content": "Synthesized database with 997 utterances of three basic emotions and neutral expression based on rule-based manipulation for a diphone synthesizer which we release to the public ",
"Emotions": "6 emotions: angry, bored, happy, neutral, sad and scared",
"Format": "Audio",
"Size": "0.941 GB",
"Language": "German",
"Paper": "SyntAct: A Synthesized Database of Basic Emotions",
"Access": "Open",
"License": "CC BY-SA 4.0",
"Dataset-link": "https://zenodo.org/record/6573016#.ZAjy_9LMJpj",
"Paper-link": "http://felix.syntheticspeech.de/publications/synthetic_database.pdf",
"License-link": "https://creativecommons.org/licenses/by/4.0"
},
"BEAT": {
"Year": 2022,
"Content": "76-Hour and 30-Speaker of 4 different languages: English (60h), Chinese (12h), Spanish (2h) and Japanese (2h).",
"Emotions": "8 emotions: happiness, anger, disgust, sadness, contempt, surprise, fear, and neutral",
"Format": "Audio, Video",
"Size": "42 GB",
"Language": "English, Chinese, Spanish, Japanese",
"Paper": "A Large-Scale Semantic and Emotional Multi-Modal Dataset for Conversational Gestures Synthesis",
"Access": "Open",
"License": "Non-commercial license",
"Dataset-link": "https://drive.google.com/drive/folders/1EKuWH8q178QOtFUYaNohdkZbBHQYAmhL",
"Paper-link": "https://www.ecva.net/papers/eccv_2022/papers_ECCV/papers/136670605.pdf",
"License-link": "--"
},
"Dusha": {
"Year": 2022,
"Content": " 300 000 audio recordings (~350 hours) of Russian speech, their transcripts and emotiomal labels. The dataset has two subsets: acted and real-life",
"Emotions": "4 emotions: angry, happy, sad and neutral. Arousal and valence metrics are also available.",
"Format": "Audio",
"Size": "58 GB",
"Language": "Russian",
"Paper": "Large Raw Emotional Dataset with Aggregation Mechanism",
"Access": "Open",
"License": "Public license with attribution and conditions reserved",
"Dataset-link": "https://github.com/salute-developers/golos/tree/master/dusha",
"Paper-link": "https://arxiv.org/abs/2212.12266",
"License-link": "https://github.com/salute-developers/golos/blob/master/license/en_us.pdf"
},
"MAFW": {
"Year": 2022,
"Content": "10045 video-audio clips in the wild.",
"Emotions": "11 single-label emotion categories (anger, disgust, fear, happiness, neutral, sadness, surprise, contempt, anxiety, helplessness, and disappointment) and 32 multi-label emotion categories.",
"Format": "Audio, Video",
"Size": "--",
"Language": "--",
"Paper": "MAFW: A Large-scale, Multi-modal, Compound Affective Database for Dynamic Facial Expression Recognition in the Wild",
"Access": "Restricted",
"License": "Non-commercial research purposes",
"Dataset-link": "https://mafw-database.github.io/MAFW/",
"Paper-link": "https://arxiv.org/abs/2208.00847",
"License-link": "--"
},
"EMOVIE": {
"Year": 2021,
"Content": "9724 samples with audio files and its emotion human-labeled annotation.",
"Emotions": "Polarity metrics (positive:+1, negative:-1)",
"Format": "Audio",
"Size": "0.572 GB",
"Language": "Chinese (Mandarin)",
"Paper": "EMOVIE: A Mandarin Emotion Speech Dataset with a Simple Emotional Text-to-Speech Model",
"Access": "Open",
"License": "CC BY-NC-SA 2.0",
"Dataset-link": "https://viem-ccy.github.io/EMOVIE/dataset_release.html",
"Paper-link": "https://arxiv.org/abs/2106.09317",
"License-link": "https://creativecommons.org/licenses/by-nc-sa/2.0/legalcode"
},
"emoUERJ": {
"Year": 2021,
"Content": "Ten sentences from eight actors, equally divided between genders, and they were free to choose the phrases for record audios in four emotions (377 audios). ",
Expand All @@ -97,6 +265,34 @@
"Paper-link": "--",
"License-link": "https://creativecommons.org/licenses/by/4.0/"
},
"Thorsten-Voice Dataset 2021.06 emotional": {
"Year": 2021,
"Content": "2400 normalized mono recordings by one person (Thorsten Müller) representing 300 sentences. ",
"Emotions": "Amusement, Disgust Anger, Suprise and Neutral (plus drunk, whispering and sleepy states)",
"Format": "Audio",
"Size": "0.399 GB",
"Language": "German",
"Paper": "--",
"Access": "Open",
"License": "CC0: Public Domain",
"Dataset-link": "https://zenodo.org/records/5525023",
"Paper-link": "--",
"License-link": "https://creativecommons.org/publicdomain/zero/1.0/"
},
"ASED": {
"Year": 2021,
"Content": "2474 recordings by 65 participants (25 females and 40 males)). Recordings were judged and rejected according to the opionion of eight judges.",
"Emotions": "Five emotions: anger, happiness, fear, sadness and neutral",
"Format": "Audio",
"Size": "0.135 GB",
"Language": "Amharic",
"Paper": "A New Amharic Speech Emotion Dataset and Classification Benchmark",
"Access": "Open",
"License": "--",
"Dataset-link": "https://github.com/Ethio2021/ASED_V1",
"Paper-link": "https://dl.acm.org/doi/10.1145/3529759",
"License-link": "--"
},
"ESCorpus-PE": {
"Year": 2021,
"Content": "Spanish peruvian speech gathered from Spanish interviews, TV reports, political debate and testimonials. It contains 3749 utterances, 80 speakers (44 male and 36 female), created from Youtube audios",
Expand Down Expand Up @@ -139,48 +335,6 @@
"Paper-link": "--",
"License-link": "https://creativecommons.org/publicdomain/zero/1.0/"
},
"Quechua-SER": {
"Year": 2022,
"Content": "12420 audio recordings (~15 hours) and their transcriptions by 7 native speakers.",
"Emotions": "Emotional labels using dimensions: valence, arousal, and dominance.",
"Format": "Audio",
"Size": "3.53 GB",
"Language": "Quechua Collao",
"Paper": "A speech corpus of Quechua Collao for automatic dimensional emotion recognition",
"Access": "Open",
"License": "CC BY 4.0",
"Dataset-link": "https://figshare.com/articles/media/Quechua_Collao_for_Speech_Emotion_Recognition/20292516",
"Paper-link": "https://www.nature.com/articles/s41597-022-01855-9",
"License-link": "https://creativecommons.org/licenses/by/4.0/"
},
"MESD": {
"Year": 2022,
"Content": "864 audio files of single-word emotional utterances with Mexican cultural shaping.",
"Emotions": "6 emotions provides single-word utterances for anger, disgust, fear, happiness, neutral, and sadness.",
"Format": "Audio",
"Size": "0.097 GB",
"Language": "Spanish (Mexican)",
"Paper": "The Mexican Emotional Speech Database (MESD): elaboration and assessment based on machine learning",
"Access": "Open",
"License": "CC BY 4.0",
"Dataset-link": "https://data.mendeley.com/datasets/cy34mh68j9/5",
"Paper-link": "https://pubmed.ncbi.nlm.nih.gov/34891601/",
"License-link": "https://creativecommons.org/licenses/by/4.0/"
},
"SyntAct": {
"Year": 2022,
"Content": "Synthesized database of three basic emotions and neutral expression based on rule-based manipulation for a diphone synthesizer which we release to the public ",
"Emotions": "997 utterances including 6 emotions: angry, bored, happy, neutral, sad and scared",
"Format": "Audio",
"Size": "0.941 GB",
"Language": "German",
"Paper": "SyntAct: A Synthesized Database of Basic Emotions",
"Access": "Open",
"License": "CC BY-SA 4.0",
"Dataset-link": "https://zenodo.org/record/6573016#.ZAjy_9LMJpj",
"Paper-link": "http://felix.syntheticspeech.de/publications/synthetic_database.pdf",
"License-link": "https://creativecommons.org/licenses/by/4.0"
},
"LSSED": {
"Year": 2021,
"Content": "LSSED: A Large-Scale Dataset and Benchmark for Speech Emotion Recognition",
Expand Down Expand Up @@ -307,6 +461,20 @@
"Paper-link": "http://www.interspeech2020.org/index.php?m=content&c=index&a=show&catid=290&id=684",
"License-link": "Academic License & Commercial License"
},
"AISHELL-3": {
"Year": 2020,
"Content": "Roughly 85 hours of emotion-neutral recordings spoken by 218 native Chinese mandarin speakers and total 88035 utterances.",
"Emotions": "Neutral",
"Format": "Audio",
"Size": "19 GB",
"Language": "Chinese (Mandarin)",
"Paper": "AISHELL-3: A Multi-speaker Mandarin TTS Corpus and the Baselines",
"Access": "Open",
"License": "Apache 2.0",
"Dataset-link": "https://www.openslr.org/93/",
"Paper-link": "https://arxiv.org/abs/2010.11567",
"License-link": "https://apache.org/licenses/LICENSE-2.0"
},
"BEASC": {
"Year": 2020,
"Content": "Bangla Emotional Audio-Speech Corpus",
Expand Down Expand Up @@ -377,6 +545,34 @@
"Paper-link": "http://dx.doi.org/10.1037/emo0001048",
"License-link": "https://creativecommons.org/licenses/by-nc-sa/4.0/"
},
"VESUS": {
"Year": 2019,
"Content": "252 distinct phrases, each read by 10 actors totalling 6 hours of speech.",
"Emotions": "5 emotions: anger, happiness, sadness, fear and neutral.",
"Format": "Audio",
"Size": "--",
"Language": "English",
"Paper": "VESUS: A Crowd-Annotated Database to Study Emotion Production and Perception in Spoken English",
"Access": "Restricted",
"License": "Academic EULA",
"Dataset-link": "https://engineering.jhu.edu/nsa/vesus/",
"Paper-link": "https://engineering.jhu.edu/nsa/wp-content/uploads/2019/10/IS191413.pdf",
"License-link": "Academic EULA"
},
"Morgan Emotional Speech Set": {
"Year": 2019,
"Content": "999 spontaneous voice messages from 100 Spanish speakers, collected from real conversations on a messaging app.",
"Emotions": "Valence & arrousal dimensions and 4 emotions: happiness, anger, sadness, and calmness.",
"Format": "Audio",
"Size": "0.192 GB",
"Language": "English",
"Paper": "Categorical and Dimensional Ratings of Emotional Speech: Behavioral Findings From the Morgan Emotional Speech Set",
"Access": "Open",
"License": "CC BY 4.0",
"Dataset-link": "https://arxiv.org/abs/2403.02167",
"Paper-link": "https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7203525/",
"License-link": "https://creativecommons.org/licenses/by/4.0/"
},
"PMEmo": {
"Year": 2019,
"Content": "Dataset containing emotion annotations of 794 songs as well as the simultaneous electrodermal activity (EDA) signals. A Music Emotion Experiment was well-designed for collecting the affective-annotated music corpus of high quality, which recruited 457 subjects.",
Expand Down Expand Up @@ -475,6 +671,20 @@
"Paper-link": "https://arxiv.org/pdf/1806.09514.pdf",
"License-link": "--"
},
"OMG Emotion": {
"Year": 2018,
"Content": "420 relatively long emotion videos with an average length of 1 minute, collected from a variety of Youtube channels.",
"Emotions": "7 emotions:anger, disgust, fear, happy, sad, surprise and neutral. Plus valence, arousal.",
"Format": "Audio, Video",
"Size": "--",
"Language": "English",
"Paper": "The OMG-Emotion Behavior Dataset",
"Access": "Open",
"License": "CC BY-NC-SA 3.0",
"Dataset-link": "https://www2.informatik.uni-hamburg.de/wtm/OMG-EmotionChallenge/",
"Paper-link": "https://arxiv.org/abs/1803.05434",
"License-link": "https://creativecommons.org/licenses/by-nc-sa/3.0/"
},
"RAVDESS": {
"Year": 2018,
"Content": "7356 recordings by 24 actors.",
Expand Down

0 comments on commit 54706b7

Please sign in to comment.