KuralHubStar

KuralHubDatasets

Speech emotion recognition datasets by language

We searched for emotional speech datasets in the 70 most widely spoken languages (by Ethnologue). 45 of them have at least one, and for the other 25 we found none. In all, 90 datasets are documented here.

Each language has its own page with the language family, ISO code and, for every dataset: emotion labels, size, speakers, recording setup, licence, access link, original paper and a BibTeX entry. The pages are generated from the Markdown files in the repository's datasets/ folder. To add a dataset, see CONTRIBUTING.md.

Download: datasets.json · languages.json

Missing a dataset? Suggest it with a short form or open a pull request using the template. 25 languages still have none.

Languages

70 languages
All 70 surveyed languages. Benchmarked marks the 29 languages used in the paper's experiments.
LanguageFamilyISODatasetsDataset namesBenchmarked
AfrikaansIndo-Europeanaf1AfrikaansSEyes
AmharicAfro-Asiaticam1Amharic Speech Emotion Dataset (ASED)yes
ArabicAfro-Asiaticar2ANAD, BAVEDyes
AssameseIndo-Europeanas2Assamese Emotional Speech Corpus, Assamese Emotional Speech Dataset
AzerbaijaniTurkicaz0none found
BambaraNiger-Congobm0none found
BengaliIndo-Europeanbn4BANSpEmo, BanglaSER, KBES, SUBESCOyes
BhojpuriIndo-Europeanbho0none found
BurmeseSino-Tibetanmya3BMISEC, ESMBS, New et al. (2001)
CebuanoAustronesianceb0none found
ChineseSino-Tibetanzh8CASIA, CHEAVD, CHEAVD 2.0, EESDB, EMOVIE, ESD, M3ED, SECEPDByes
CzechIndo-Europeancs2CDSE, Slavic Call Center Emotional Speech Corpus
DutchIndo-Europeannl1PICKA-NH
EnglishIndo-Europeanen3IEMOCAP, IESC, RAVDESSyes
FilipinoAustronesianfil1Filipino Multimodal Emotion Database (FilMED)
FrenchIndo-Europeanfr2CaFE, OréauFRyes
FulfuldeNiger-Congoff0none found
GermanIndo-Europeande1EmoDByes
GreekIndo-Europeanel3AESDD, GreThE, SEWAyes
GujaratiIndo-Europeangu1Tank et al. (2020)
HausaAfro-Asiaticha0none found
HindiIndo-Europeanhi3Bansal And Dev et al., IITKGP-SEHSC, VishalByes
HungarianUralichu2HESD, SEWAyes
IgboNiger-Congoig0none found
IndonesianAustronesianid2Fahmi et al (2020), IndoWaveSentimentyes
ItalianIndo-Europeanit3DEMoS, EMOVO, Emozionalmenteyes
JapaneseJaponicja2JVNV, Keio-ESDyes
JavaneseAustronesianjv1Java-SED
KannadaDravidiankn2Agrawal et al (2022), KESyes
KazakhTurkickk2KazEmoTTS, Kazakh-Russian Emotion Corpusyes
KhmerAustroasiatickm0none found
KoreanKoreanicko1KESDy18yes
KurdishIndo-Europeanku2Kurdish Emotional Speech Database, Nematullah et al. (2024)
MagahiIndo-Europeanmag0none found
MaithiliIndo-Europeanmai0none found
MalayAustronesianms1Emolah
MalayalamDravidianml4IIITSaint-EmoMDB, Kannur Malayalam Speech Emotion Recognition (KMSER) Dataset, MAVES DB, Malayalam Speech Emotion Dataset
NepaliIndo-Europeanne0none found
Nigerian PidginEnglish-based creolepcm0none found
Northern PashtoIndo-Europeanps0none found
Northern UzbekTurkicuzn0none found
OdiaIndo-Europeanor1SITB-OSEDyes
PersianIndo-Europeanfa1ShEMOyes
PolishIndo-Europeanpl3Cichosz and Slot (2005), Sapinski et al. (2019), nEMOyes
PortugueseIndo-Europeanpt2VERBO, emoUERJyes
PunjabiIndo-Europeanpa1PEMO
QuechuaQuechuanqu1Quechua Collao Corpusyes
RomanianIndo-Europeanro1SRoL
RussianIndo-Europeanru2Dusha, RESDyes
SaraikiIndo-Europeanskr0none found
SindhiIndo-Europeansd1Urdu-Sindhi Speech Emotion Corpus
SinhalaIndo-Europeansi0none found
SomaliAfro-Asiaticso0none found
SothoNiger-Congost0none found
SpanishIndo-Europeanes3EMOVOME, ESCorpus-PE, MESDyes
SundaneseAustronesiansun0none found
SwahiliNiger-Congosw1Kenyan Swahili SERyes
SwedishIndo-Europeansv2EU-Emotion Voice Database, Swedish Voice Provider Emotional Speech Corpus
TagalogAustronesiantl0none found
TamilDravidianta3EmoTa, Ram, C. Sunitha, et al., TaMaR-EmoDByes
TeluguDravidiante2IITKGP-SESC, Sowmith et al (2024)yes
ThaiKra-Daith1Thai Speech Emotion Dataset
TswanaNiger-Congotn0none found
TurkishTurkictr2Oflazoglu et al (2011), TurEV-DByes
UkrainianIndo-Europeanuk0none found
UrduIndo-Europeanur2URDU-Dataset, UrduSERyes
VietnameseAustroasiaticvi1VNEMOS
XhosaNiger-Congoxho0none found
YorubaNiger-Congoyo0none found
ZuluNiger-Congozul0none found
Tree diagram grouping the surveyed languages by family: Indo-European (with Indo-Aryan, Iranian, Slavic, Italic, Germanic and Hellenic branches), Niger-Congo, Afro-Asiatic, Austronesian, Austroasiatic, Dravidian, Turkic, Sino-Tibetan, Uralic, Kra-Dai, Japonic, Koreanic and Quechuan.
Language families of the surveyed languages. Click to enlarge.

All datasets

90 datasets
All 90 datasets. Values come from the dataset pages in the repository. A blank cell means the source did not say.
DatasetLanguageTypeEmotion labelsSamplesRecordingAccessLicence
AfrikaansSEAfrikaansAudioAnticipation, Joy, Trust, Fear, Surprise, Sadness, Disgust and Anger800Studio (Local Youtube Channel)On requestCreative Commons Attribution license
Amharic Speech Emotion Dataset (ASED)AmharicAudioNeutral, Fearful, Happy, Sad, Angry2,474ActedOpen
ANADArabicAudioHappy, Angry, Surprised1384SpontaneousOpenCC BY-NC-SA 4.0
BAVEDArabicAudioLow, Neutral, High1935ActedOpen
Assamese Emotional Speech CorpusAssameseAudioAngry, Happy, Sad, Neutral800Acted (controlled settings)On requestCC BY 4.0
Assamese Emotional Speech DatasetAssameseAudioCalm, Neutral, Anger, Sad, Happy, Surprise, Disgust1750Acted (controlled settings)On request
BanglaSERBengaliAudioAngry, Happy, Neutral, Sad, and Surprise1467ActedOpenCC-BY 4.0
BANSpEmoBengaliAudioDisgust, Happy, Sad, Surprised, Anger, Fear792ActedOpenCC-BY 4.0
KBESBengaliAudioNeutral, Happy (Low), Happy (High), Sad (Low), Sad (High), Angry (Low), Angry (High), Disgust (Low), Disgust (High)900ActedOpenCC-BY 4.0
SUBESCOBengaliAudioAnger, Disgust, Fear, Happiness, Neutral, Sadness, and Surprise7000ActedOpenCC-BY 4.0
BMISECBurmeseAudioAngry, Happy, Disgust, Fear, Sad, Surprise, Neutral1,332Spontaneous (Myanmar movies, interviews)Open
ESMBSBurmeseAudioAnger, happiness, sadness, disgust, fear, surprise720StudioOn request
New et al. (2001)BurmeseAudioAnger, Dislike, Fear, Happiness, Sadness and SurpriseSpontaneousOn request
CASIAChineseAudioAnger, Fear, Happy, Neutral, Sad, Surprise7,200ActedOpenApache 2.0
CHEAVDChineseAudio-VisualAnger, Anxious, Sadness, Surprise, Disgust, Fear, Happiness, Worried2,322ActedOpen
CHEAVD 2.0ChineseAudio-VisualAnger, Anxious, Sadness, Surprise, Disgust, Happiness, Worried, Neutral7,030SpontaneousOpen
EESDBChineseAudioAnger, Disgust, Fear, Happiness, Neutral, Sadness, and Surprise400SpontaneousOn request
EMOVIEChineseAudio / Text5-point polarity scale (Strong Negative to Strong Positive)9,724Movie speechOpen
ESDChineseAudioNeutral, Anger, Happiness, Sadness, Surprise17,500SpontaneousOn request
M3EDChineseMultimodalHappy, Neutral, Sad, Disgust, Anger, Fear, Surprise24,449Acted (TV series)Open
SECEPDBChineseMultimodalAnxiety, boredom, Anger, Disgust, Happiness, Neutral, Sadness810SpontaneousOn request
CDSECzechAudioHappiness, Anger, Sadness, NeutralSpontaneous (extracted from radio and TV broadcasts)On request
Slavic Call Center Emotional Speech CorpusCzechAudioAnger, Happiness, Sadness, Surprise, Neutral600Spontaneous (real call center interactions)On request
PICKA-NHDutchAudioHappiness, Anger, SadnessStudioOpenMIT, CC-BY
IEMOCAPEnglishAudioHappy, Anger, Sadness, Frustration, and Neutral12000+ActedOn requestDataset-Specific
IESCEnglishAudioNeutral, Happy, Angry, Sad,and Fear600ActedOpen
RAVDESSEnglishAudioCalm, Happy, Sad, Angry, Fearful, Surprise, and Disgust1440ActedOpenCC-BY 4.0
Filipino Multimodal Emotion Database (FilMED)FilipinoMultimodal (Audio-Video)kasiyahan (happiness), kalungkutan (sadness), galit (anger), takot (fear), gulat (surprise), pandidiri (disgust)1,440Studio (acted) and spontaneous expressionsOn request
CaFEFrenchAudioSadness, Happiness, Anger, Fear, Disgust, Surprise, and Neutral936ActedOpenCC-BY 4.0
OréauFRFrenchAudioAnger, Disgust, Fear, Happiness, Sadness, Surprise, Neutral502ActedOpenCC-BY 4.0
EmoDBGermanAudioAnger, Boredom, Anxiety, Happy, Sadness, Disgust, and Neutral535ActedOpen
AESDDGreekAudioHappy, Sad, Angry, Fear, Disgust500ActedOpen
GreThEGreekAudioemotions are characterized using Valence(Negative/Neutral/Positive) and Arousal(Weak/Neutral/Strong).500Acted (Theatrical Plays)Open
SEWAGreekMultimodal (Audio, Video)emotions are characterized using valence and arousal.1990Spontaneous (natural interactions)On requestDataset-Specific - SEWA EULA
Tank et al. (2020)GujaratiAudioSadness, Surprise, Anger, Disgust, Fear, Happiness1296ActedOn request
Bansal And Dev et al.HindiAudio6048ActedOn request
IITKGP-SEHSCHindiAudioAnger, Disgust, Fear, Happy, Neutral, Sad, Sarcastic and Surprise12000ActedOn request
VishalBHindiAudioAnger, Disgust, Fear, Happy, Neutral, Sad, Sarcastic and Surprise3200ActedOpen
HESDHungarianAudioFear, Sadness, Anger, Surprise, Disgust, Joy, Excitement, Neutral888Acted(Non Professional Speakers - Quiet Office Room)On request
SEWAHungarianMultimodal (Audio, Video)emotions are characterized using valence and arousal.1990Spontaneous (natural interactions)On requestDataset-Specific - SEWA EULA
Fahmi et al (2020)IndonesianAudioHappy, Sad, Angry, Neutral965ActedOn request
IndoWaveSentimentIndonesianAudioNeutral, Happy, Surprised, Disgusted, and Disappointed300ActedOpenCC-BY 4.0
DEMoSItalianAudioAnger, Sadness, Happiness, Fear, Surprise, Disgust, Guilt6697ActedOn request
EMOVOItalianAudioDisgust, Fear, Anger, Joy, Surprise, Sadness, Neutral564ActedOpen
EmozionalmenteItalianAudioAnger, Disgust, Fear, Joy, Sadness, Surprise, Neutral6902ActedOpenCC-BY 4.0
JVNVJapaneseAudioAnger, Disgust, Fear, Happiness, Sadness, and Surprise1615ActedOpenCC-BY-SA 4.0
Keio-ESDJapaneseAudioAnger, joy, disgust, contempt, amusement, worry, kindness, relief, indignation, shame, ... (47 types in total)ActedOpen
Java-SEDJavaneseAudioNeutral, Sad, Happy, Surprised, Fear and Anger1680ActedOn request
Agrawal et al (2022)KannadaAudioAnger, Sadness, Surprise, Happiness, Fear, Neutral468ActedOpenCC-BY 4.0
KESKannadaAudioActedOn request
Kazakh-Russian Emotion CorpusKazakhAudioHappiness, Sadness, Anger, Neutral, Boredom, Disgust, Fear16160ActedOn request
KazEmoTTSKazakhAudio + TextHappy, Sad, Angry, Neutral, Scared, and Surprised54,760StudioOn requestCC-BY 4.0
KESDy18KoreanAudioAngry, happy, neutral, sad, disgust, fear, surprise2880StudioOpenDataset-Specific
Kurdish Emotional Speech DatabaseKurdishAudioAnger, Happiness, Sad, Fear, Boredom, Surprise, and Neutral3360ActedOn request
Nematullah et al. (2024)KurdishAudioNeutral, Sadness, Happiness, and Anger287StudioOn request
EmolahMalayAudioHappy, Sad, Angry, NeutralSpontaneousOn requestCreative Commons Attribution License
IIITSaint-EmoMDBMalayalamAudioHappy, Sad, Angry and Neutral or emotion ratings at valence and arousalOn request
Kannur Malayalam Speech Emotion Recognition (KMSER) DatasetMalayalamAudioHappy, Angry, Sad, NeutralActedOn request
Malayalam Speech Emotion DatasetMalayalamAudioHappy, Sad, Angry, NeutralElicitedOn request
MAVES DBMalayalamMultimodal (Audio-Visual)OpenCC BY 4.0
SITB-OSEDOdiaAudioAnger, Surprise, Happy, Sadness, Disgust, Fear12,110StudioOpen
ShEMOPersianAudioAnger, Fear, Happiness, Sadness, Surprise, Neutral3000ActedOpen
Cichosz and Slot (2005)PolishAudioNeutral, Joy, Anger, Sadness, Fear, Boredom240 utterancesActedOn request
nEMOPolishAudioAnger, Fear, Happiness, Sadness, Surprise, Neutral4481StudioOpenCC BY-NC-SA 4.0
Sapinski et al. (2019)PolishAudio-VisualHappy, Sad, Angry, Disgust, Fear, Surprise560 per modalityStudioOn request
emoUERJPortugueseAudioHappy, Sad, Angry, Neutral377ActedOpenCC-BY 4.0
VERBOPortugueseAudioHappy, Disgust, Fear, Neutral, Anger, Surprise, and Sadness1176ActedOn request
PEMOPunjabiAudioAnger, Sad, Happy, Neutral22000ActedOn request
Quechua Collao CorpusQuechuaAudioHappy, Sad, Bored, Fear, Sleepy, Calm, Excited, Angry, and Neutral12420ActedOpenCC BY 4.0
SRoLRomanianAudioHappiness, Sadness, Fury, Nutral396ActedOpen
DushaRussianAudioHappiness (Positive), Sadness, Anger and Neutral300000SpontaneousOpen
RESDRussianAudioAnger, Disgust, Fear, Enthusiasm, Happiness, Neutral, and Sadness1120ActedOpen
Urdu-Sindhi Speech Emotion CorpusSindhiAudioAnger, Disgust, Happiness, Neutral, Sarcasm, Sadness, and SurpriseOn request
EMOVOMESpanishAudioCovering valence and arousal dimensions999SpontaneousOpenCC-BY 4.0
ESCorpus-PESpanishAudiothree emotional dimensions Valence, Arousal, Dominance3749SpontaneousOpenCC BY 4.0
MESDSpanishAudioHappy, Sad, Angry, Neutral, Fear, Disgust432ActedOpenCC-BY 4.0
Kenyan Swahili SERSwahiliAudioSad, Happy, Surprised, Angry, CalmStudio / Acted / SpontaneousOn request
EU-Emotion Voice DatabaseSwedishAudioAdmiration, Amusement, Anger, Anxiety, Awe, Boredom, Disgust, Fear, Interest, Joy, Pleasure, Relief, Sadness, Satisfaction, Surprise, Triumph, Contempt, Embarrassment, Neutral, Pride2,159Acted (controlled settings)On request
Swedish Voice Provider Emotional Speech CorpusSwedishAudioNot explicitly labeled; inferred categories include Neutral, Emphatic, and Negative~10,000 utterancesSpontaneous (recordings from voice-controlled telephone services)On request
EmoTaTamilAudioHappy, Sad, Angry, Fear, Neutral936ActedOpenEmoTa Academic-Commercial
Ram, C. Sunitha, et al.TamilAudioJoy, Sad, Angry, Neutral, Fear95ActedOn request
TaMaR-EmoDBTamilAudioAnger, Happy, Anxiety, Sad, Neutral1202ActedOn request
IITKGP-SESCTeluguAudioHappiness, Surprise, Anger, Disgust, Sadness, Fear, Sarcasm, Neutral12000ActedOn request
Sowmith et al (2024)TeluguAudioHappy, Sad, Angry, Neutral, Surprised458ActedOpenApache 2.0
Thai Speech Emotion DatasetThaiAudioNeutral, Anger, Happiness, Sadness, and Frustration27854ActedOpenCC-BY-SA 4.0
Oflazoglu et al (2011)TurkishAudioHappy, Surprised, Sad, Angry, Fear, Neutral5304ActedOn request
TurEV-DBTurkishAudioAngry, Calm, Happy, Sad1735ActedOpenCC-BY 4.0
URDU-DatasetUrduAudioHappy, Sad, Angry, Neutral400ActedOpen
UrduSERUrduAudioAngry, Fear, Boredom, Disgust, Happy, Neutral, and Sad3500ActedOpenCC-BY 4.0
VNEMOSVietnameseAudioHappy, Sad, Angry, Neutral, Anxiety250Acted and SpontaneousOn request

Licence and access details are copied from each dataset's own documentation. Always check the original licence before using a dataset. If you use one, cite its authors. If KuralHub helped you find it, please also cite the KuralHub paper.