--- dataset_info: features: - name: speaker_id dtype: string - name: emotion dtype: string - name: emotion_intensity dtype: string - name: transcript dtype: string - name: repetition dtype: string - name: language dtype: string - name: audio dtype: audio - name: gender dtype: string - name: age dtype: string - name: race dtype: string - name: ethnicity dtype: string splits: - name: jlcorpus num_bytes: 448531424.8 num_examples: 2400 - name: ravdess num_bytes: 592549048.8 num_examples: 1440 - name: enterface num_bytes: 656498481.817 num_examples: 1287 - name: mead num_bytes: 2763409133.106 num_examples: 31734 - name: esd num_bytes: 3267036036.0 num_examples: 35000 - name: cremad num_bytes: 610714547.78 num_examples: 7442 - name: savee num_bytes: 59059712.0 num_examples: 480 download_size: 8226860408 dataset_size: 8397798384.3029995 configs: - config_name: default data_files: - split: jlcorpus path: data/jlcorpus-* - split: ravdess path: data/ravdess-* - split: enterface path: data/enterface-* - split: mead path: data/mead-* - split: esd path: data/esd-* - split: cremad path: data/cremad-* - split: savee path: data/savee-* --- # Emotion Datasets This dataset is a collection of several emotion datasets: JL Corpus, RAVDESS, eNTERFACE, MEAD, ESD, and CREMA-D. Example: ```json { "speaker_id": "crema-d-speaker-1067", # Speaker ID "emotion": "angry", # Emotion label "emotion_intensity": "medium", # Emotion intensity "transcript": "It's eleven o'clock.", # Transcript "repetition": "null", # Repetition "language": "English", # Language "audio": "...", # Audio file "gender": "male", # Gender "age": "66", # Age group "race": "Caucasian", # Race "ethnicity": "Non-Hispanic" # Ethnicity } ``` ## jlcorpus **Labels**: speaker_id, emotion, transcript, language, gender **Emotions**: happy, concerned, excited, anxious, assertive, apologetic, angry, neutral, sad, encouraging **Emotion Intensities**: None **Languages**: English **Num unique transcripts: 29** **Multiple emotions for the same transcript**: Yes | Transcript | Emotion | Count | | --- | --- | --- | | Carl leaps into a jeep. | angry | 16 | | Carl leaps into a jeep. | anxious | 16 | | Carl leaps into a jeep. | apologetic | 16 | | Carl leaps into a jeep. | assertive | 16 | | Carl leaps into a jeep. | concerned | 16 | | Carl leaps into a jeep. | encouraging | 16 | | Carl leaps into a jeep. | excited | 16 | | ... | ... | ... | ## ravdess **Labels**: speaker_id, emotion, emotion_intensity, transcript, repetition, language, gender **Emotions**: angry, happy, disgust, fearful, calm, surprised, neutral, sad **Emotion Intensities**: strong, normal **Languages**: English **Num unique transcripts: 2** **Multiple emotions for the same transcript**: Yes | Transcript | Emotion | Emotion Intensity | Count | | --- | --- | --- | --- | | Dogs are sitting by the door | angry | normal | 48 | | Dogs are sitting by the door | angry | strong | 48 | | Dogs are sitting by the door | calm | normal | 48 | | Dogs are sitting by the door | calm | strong | 48 | | Dogs are sitting by the door | disgust | normal | 48 | | Dogs are sitting by the door | disgust | strong | 48 | | Dogs are sitting by the door | fearful | normal | 48 | | ... | ... | ... | ... | ## enterface **Labels**: speaker_id, emotion, transcript, language **Emotions**: anger, disgust, fear, happiness, sadness, surprise **Emotion Intensities**: None **Languages**: English **Num unique transcripts: 30** **Multiple emotions for the same transcript**: No | Transcript | Emotion | Count | | --- | --- | --- | | Aaaaah a cockroach!!! | disgust | 43 | | Eeeek, this is disgusting!!! | disgust | 43 | | Everything was so perfect! I just don't understand! | sadness | 43 | | He (she) was my life. | sadness | 43 | | I can have you fired you know! | anger | 43 | | I didn't expect that! | surprise | 43 | | I don't care about your coffee! Please serve me! | anger | 43 | | ... | ... | ... | ## mead **Labels**: speaker_id, emotion, emotion_intensity, language, gender **Emotions**: contempt, neutral, angry, sad, surprised, fear, happy, disgusted **Emotion Intensities**: low, medium, high **Languages**: English **Multiple emotions for the same transcript**: Unknown | Emotion | Count | | --- | --- | | angry | 4194 | | contempt | 4215 | | disgusted | 4284 | | fear | 4251 | | happy | 4294 | | neutral | 1916 | | sad | 4310 | | ... | ... | ## esd **Labels**: speaker_id, emotion, transcript, language **Emotions**: Happy, Neutral, Surprise, Angry, Sad **Emotion Intensities**: None **Languages**: Chinese, English **Num unique transcripts: 953** **Multiple emotions for the same transcript**: Yes | Transcript | Emotion | Count | | --- | --- | --- | | A boat put out on the bay. | Angry | 10 | | A boat put out on the bay. | Happy | 10 | | A boat put out on the bay. | Neutral | 10 | | A boat put out on the bay. | Sad | 10 | | A boat put out on the bay. | Surprise | 10 | | A deafening chirruping rent the air. | Angry | 10 | | A deafening chirruping rent the air. | Happy | 10 | | ... | ... | ... | ## cremad **Labels**: speaker_id, emotion, emotion_intensity, transcript, language, gender, age, race, ethnicity **Emotions**: fearful, sad, angry, disgust, happy, neutral **Emotion Intensities**: null, high, medium, low **Languages**: English **Num unique transcripts: 12** **Multiple emotions for the same transcript**: Yes | Transcript | Emotion | Emotion Intensity | Count | | --- | --- | --- | --- | | Don't forget a jacket. | angry | null | 91 | | Don't forget a jacket. | disgust | null | 91 | | Don't forget a jacket. | fearful | null | 91 | | Don't forget a jacket. | happy | null | 91 | | Don't forget a jacket. | neutral | null | 91 | | Don't forget a jacket. | sad | null | 91 | | I think I have a doctor's appointment. | angry | null | 90 | | ... | ... | ... | ... | ## savee **Labels**: gender, transcript, emotion **Emotions**: anger, disgust, fear, happiness, neutral, sadness, surprise **Emotion Intensities**: None **Languages**: Unknown **Num unique transcripts: 198** **Multiple emotions for the same transcript**: Yes | Transcript | Emotion | Count | | --- | --- | --- | | A few years later, the dome fell in. | anger | 4 | | A lot of people were roaming the streets in costumes and masks and having a ball. | anger | 1 | | A lot of people will roam the streets in costumes and masks and having a ball. | anger | 2 | | Agricultural products are unevenly distributed. | neutral | 4 | | Allow Levy here, but rationalise all errors. | neutral | 1 | | Allow leeway here, but rationalise all errors. | neutral | 3 | | American newspaper reviewers like to call his plays nihilistic. | sadness | 3 | | ... | ... | ... |