AccentDB
AccentDB is a multi-pairwise parallel corpus of structured and labelled accented speech. It contains speech samples from speakers of 4 non-native accents of English (8 speakers,…
Audio and speech recordings for speech, speaker and music recognition.
AccentDB is a multi-pairwise parallel corpus of structured and labelled accented speech. It contains speech samples from speakers of 4 non-native accents of English (8 speakers,…
Speech-synthesis databases of single-speaker English recordings (18 speaker sets available in torchaudio).
Mozilla Common Voice Dataset
CREMA-D is an audio-visual data set for emotion recognition. The data set consists of facial and vocal emotional expressions in sentences spoken in a range of…
Device Recorded VCTK (small subset version): VCTK speech re-recorded through consumer devices, with train and test subsets.
A free audio dataset of spoken digits. Think MNIST for audio.
The Free Universal Sound Separation (FUSS) Dataset is a database of arbitrary sound mixtures and source-level references, for use in experiments on arbitrary sound separation. This…
The Groove MIDI Dataset (GMD) is composed of 13.6 hours of aligned MIDI and (synthesized) audio of human-performed, tempo-aligned expressive drumming captured on a Roland TD-11…
The dataset consists of 1000 audio tracks each 30 seconds long. It contains 10 genres, each represented by 100 tracks.
The dataset was collected for the purposes of music/speech discrimination. The dataset consists of 120 tracks, each 30 seconds long.
Subset of Libri-light used in HuBERT for supervised fine-tuning, with 10-minute, 1-hour and 10-hour labelled subsets.
Speech separation dataset with Libri2Mix and Libri3Mix mixtures (train-360, train-100, dev, test).
LibriSpeech is a corpus of approximately 1000 hours of read English speech with sampling rate of 16 kHz, prepared by Vassil Panayotov with the assistance of…
LibriTTS is a multi-speaker English corpus of approximately 585 hours of read English speech at 24kHz sampling rate, prepared by Heiga Zen with the assistance of…
This is a public domain speech dataset consisting of 13,100 short audio clips of a single speaker reading passages from 7 non-fiction books. A transcription is…
Uncompressed version of MUSDB18 for music source separation, with bass, drums, other, vocals and mixture stems.
The NSynth Dataset is an audio dataset containing ~300k musical notes, each with a unique pitch, timbre, and envelope. Each note is annotated with three additional…
Query-by-example spoken term detection database in several languages (albanian, basque, czech, nnenglish, romanian, slovak).
SAVEE (Surrey Audio-Visual Expressed Emotion) is an emotion recognition dataset. It consists of recordings from 4 male actors in 7 different emotions, 480 British English utterances…
An audio dataset of spoken words designed to help train and evaluate keyword spotting systems. Its primary goal is to provide a way to build and…
The TED-LIUM corpus is English-language TED talks, with transcriptions, sampled at 16kHz. It contains about 118 hours of speech.
This CSTR VCTK Corpus includes speech data uttered by 110 English speakers with various accents. Each speaker reads out about 400 sentences, which were selected from…
An large scale dataset for speaker identification. This data is collected from over 1,251 speakers, with over 150k samples in total.
FLEURS is the speech version of the FLORES machine translation benchmark, covering 2000 n-way parallel sentences in n=102 languages. XTREME-S covers four task families: speech recognition,…
Sections
No section has this name. Press “Search” to search all site content.
Favourites
My results & certificates
My academy progressMy quiz resultsKids’ progressAcademy certificatesQuiz certificatesStart any lesson or quiz and it will show up here so you can get back to it quickly.
Settings
I am the explorer's assistant. Ask me about datasets by field or task, about their licences, sources and related papers, or ask me to compare two datasets.
Answers come from the explorer's database only; it never invents numbers or licences. What you type is not stored on the server.
The assistant can make mistakes, so check the attached sources.