cmu-mosi

Created by Amir Zadeh, Rown Zellers, Eli Pincus, Louis-Philippe Morency

version

1.1.1

license

CC-BY-NC-4.0

usage

research

languages

English

format

wav

channel

1

sampling rate

16000

bit depth

16

duration

0 days 04:01:39.517937500

files

93, duration distribution: 54.9 s cmu-mosi-1.1.1-file-duration-distribution 180.1 s

segments

185263, duration distribution: nan s cmu-mosi-1.1.1-segment-duration-distribution nan s

repository

audb-public

Description

Opinion-level annotated corpus of sentiment and subjectivity analysis in online videos. The dataset is annotated with labels for subjectivity, sentiment intensity, per-frame and per-opinion. Reference: https://arxiv.org/abs/1606.06259

Example

Audio/0h-zjBukYpk.wav

../_images/cmu-mosi-1.1.1-player-waveform.png

Tables

Click on a row to toggle a preview.

ID

Type

Columns

dev

filewise

file

Audio/WKA5OygbEKI.wav

Audio/c5xsKMxpXnc.wav

Audio/atnd_PF-Lbs.wav

Audio/bvLlb-M3UXU.wav

Audio/bOL9jKpeJRs.wav

10 rows x 0 columns

files

filewise

gender

file

gender

Audio/03bSnISJMiM.wav

female

Audio/0h-zjBukYpk.wav

male

Audio/1DmNV9C1hbY.wav

male

Audio/1iG0909rllw.wav

female

Audio/2WGyTLYerpo.wav

male

93 rows x 1 column

phonemes

segmented

phonemes

file

start

end

phonemes

Audio/03bSnISJMiM.wav

0 days 00:00:00.012471

0 days 00:00:01.728571

sp

Audio/03bSnISJMiM.wav

0 days 00:00:01.728571

0 days 00:00:02.117687

ow1

Audio/03bSnISJMiM.wav

0 days 00:00:02.117687

0 days 00:00:05.160770

sp

Audio/03bSnISJMiM.wav

0 days 00:00:05.160770

0 days 00:00:05.220634

dh

Audio/03bSnISJMiM.wav

0 days 00:00:05.220634

0 days 00:00:05.320408

ih1

144101 rows x 1 column

sentiment

segmented

sentiment, sentiment.binarized, sentiment.binary, sentiment.binary.old

file

start

end

sentiment

sentiment.binarized

sentiment.binary

sentiment.binary.old

Audio/03bSnISJMiM.wav

0 days 00:00:51.904533

0 days 00:00:55.945351

2.4

highly positive

positive

non-negative

Audio/03bSnISJMiM.wav

0 days 00:00:56.045124

0 days 00:01:06.780725

-0.8

weakly negative

negative

negative

Audio/03bSnISJMiM.wav

0 days 00:01:06.780720

0 days 00:01:08.736281

-1.0

weakly negative

negative

negative

Audio/03bSnISJMiM.wav

0 days 00:01:08.736280

0 days 00:01:10.542176

-1.75

negative

negative

negative

Audio/03bSnISJMiM.wav

0 days 00:01:10.542175

0 days 00:01:11.699546

0.0

neutral

non-negative

2197 rows x 4 columns

test

filewise

file

Audio/tmZoasNr4rU.wav

Audio/zhpQhgha_KU.wav

Audio/lXPQBPVc5Cw.wav

Audio/iiK8YX8oH1E.wav

Audio/tStelxIAHjw.wav

31 rows x 0 columns

train

filewise

file

Audio/2iD-tVS8NPw.wav

Audio/8d-gEyoeBzc.wav

Audio/Qr1Ca94K55A.wav

Audio/Ci-AH39fi3Y.wav

Audio/8qrpnFRGt2A.wav

52 rows x 0 columns

transcription

segmented

transcription

file

start

end

transcription

Audio/03bSnISJMiM.wav

0 days 00:00:51.904533

0 days 00:00:55.945351

anyhow it was really good

Audio/03bSnISJMiM.wav

0 days 00:00:56.045124

0 days 00:01:06.780725

they didnt really do a whole bunch of background info on why she has to fight and be prepared

Audio/03bSnISJMiM.wav

0 days 00:01:06.780720

0 days 00:01:08.736281

i mean they did a little bit of it

Audio/03bSnISJMiM.wav

0 days 00:01:08.736280

0 days 00:01:10.542176

but not a whole bunch

Audio/03bSnISJMiM.wav

0 days 00:01:10.542175

0 days 00:01:11.699546

and they i guess

2197 rows x 1 column

words

segmented

words

file

start

end

words

Audio/03bSnISJMiM.wav

0 days 00:00:00.012471

0 days 00:00:01.728571

sp

Audio/03bSnISJMiM.wav

0 days 00:00:01.728571

0 days 00:00:02.117687

oh

Audio/03bSnISJMiM.wav

0 days 00:00:02.117687

0 days 00:00:05.160770

sp

Audio/03bSnISJMiM.wav

0 days 00:00:05.160770

0 days 00:00:05.799319

this

Audio/03bSnISJMiM.wav

0 days 00:00:05.799319

0 days 00:00:06.896825

sp

49955 rows x 1 column

Schemes

ID

Dtype

Min

Max

Labels

gender

str

female, male

phoneme

str

aa0, aa1, aa2, ae0, ae1, ae2, ah0, […], uw0, uw1, uw2, v, w, y, z, zh

sentiment

float

-3

3

sentiment.binarized

str

highly negative, highly positive, negative, neutral, positive, weakly negative, weakly positive

sentiment.binary

str

negative, positive

sentiment.binary.old

str

negative, non-negative

transcription

str