audio_id
stringlengths
14
14
speaker_id
int64
6.51k
19.2k
gender
stringclasses
4 values
age
stringclasses
15 values
duration
float64
0
20.1
normalized_text
stringlengths
2
109
006512-0054527
6,512
male
13-17
5.64
núna ætlar karl að hugsa um mig svaraði bella fyrir hönd þeirra beggja
006512-0054528
6,512
male
13-17
3.54
liti ekki á hann tímunum saman
006512-0054529
6,512
male
13-17
3.36
sýndi ekki mikinn áhuga
006512-0054530
6,512
male
13-17
5.46
lagði svo af með láði list sem var góð og traust
006512-0054531
6,512
male
13-17
6.18
kristín var farin að hitta elísabetu halldórsdóttur og kanna líðan hennar
006514-0054542
6,514
male
13-17
5.06
hektor starði forviða á hana
006514-0054543
6,514
male
13-17
5.94
markmiðið er að komast beint austur er það ekki
006514-0054544
6,514
male
13-17
8.68
fyrr eða síðar hverfa ónauðsynlegu hjónabandshömlurnar ein af annarri
006514-0054545
6,514
male
13-17
5.11
systir þín kallaði hana ljónatemjara
006514-0054546
6,514
male
13-17
7.2
þeir segjast báðir óviljugir að taka að sér starf ráðherrans
006517-0054582
6,517
female
13-17
7.08
á stéttinni fyrir framan hótelið fræga standa tveir menn með hatta
006517-0054583
6,517
female
13-17
2.99
fylgdu þeim fram og til baka
006517-0054584
6,517
female
13-17
4.31
það hafði ótal sinnum verið tekið upp á hana
006517-0054585
6,517
female
13-17
5.97
hann er ekki sami hrokagikkurinn og sum börn sem maður rekst á
006517-0054586
6,517
female
13-17
5.12
hann virtist allsendis óhræddur við draugana í dalnum
006518-0054638
6,518
male
13-17
4.14
þakka yður ósköp vel fyrir segir stúlkan og hneigir sig
006518-0054639
6,518
male
13-17
2.94
en á honum var ekkert að finna
006518-0054640
6,518
male
13-17
7.08
hins vegar átti ísraelsher hugsanlega harma að hefna gagnvart heyrnardaufri stjúpmóður hans
006518-0054641
6,518
male
13-17
2.88
nema að hún keypti það ekki
006519-0054642
6,519
male
13-17
4.13
þá var röð myndanna breytt á milli umferða
006519-0054643
6,519
male
13-17
3.53
ég get bara ekki munað hvað það var
006519-0054644
6,519
male
13-17
4.09
hannes er eldrauður í framan
006519-0054645
6,519
male
13-17
4.5
hvur andskotinn brotið af brísingameninu
006519-0054646
6,519
male
13-17
4.41
en ég veit hvar þeir eru sagði edda
006521-0054652
6,521
female
13-17
6.83
afsakið herra segir afgreiðslumaðurinn og tekur húfuna sína varlega ofan
006521-0054653
6,521
female
13-17
5.29
hafði javert orðið vör við þennan örlitla skjálfta
006521-0054654
6,521
female
13-17
3.39
kemur hitinn virkilega úr þessu
006521-0054655
6,521
female
13-17
7.85
rafmagnsleysið á dögunum er rökrétt afleiðing pólitískra ákvarðana undanfarinna missera
006523-0054662
6,523
male
13-17
5.25
móðir mín var þá gömluð og gleymin og líklega saknaði ég hennar
006523-0054663
6,523
male
13-17
2.83
hann er enn ólofaður
006523-0054664
6,523
male
13-17
4.32
hvert og eitt dna próf er tímafrekt
006523-0054665
6,523
male
13-17
4.55
sést nokkuð í gegnum svona glerauga er það ekki bara upp á punt
006523-0054666
6,523
male
13-17
3.62
varst þú þar stödd við aftökuna
006524-0054668
6,524
female
13-17
3.76
sævar lofaði að hringja í hann daginn eftir
006524-0054669
6,524
female
13-17
3.62
þess gerist ekki þörf sagði hún
006524-0054671
6,524
female
13-17
4.27
fyrri og seinni kona og alls fjögur börn
006520-0054672
6,520
female
13-17
3.62
þessi glæpur er tvíhöfða
006520-0054674
6,520
female
13-17
4.41
hún reyndi að bregða ekki svip
006525-0054677
6,525
female
13-17
4.97
hvert og eitt dna próf er tímafrekt
006525-0054678
6,525
female
13-17
7.24
þrátt fyrir glæsimennsku sína mistekst sigmundi ætlunarverk sitt og þrándur hefur betur
006525-0054679
6,525
female
13-17
4.55
hugsa sér allt sem úrskeiðis getur farið
006525-0054680
6,525
female
13-17
4.41
þórður er ekki að mínu skapi
006525-0054681
6,525
female
13-17
3.81
jæja við sjáumst þá sagði hún
006525-0054687
6,525
female
13-17
5.48
þá var fólk ekki heldur að springa úr offitu eins og núna
006525-0054688
6,525
female
13-17
4.04
þetta var nærmynd og þau skellihlæjandi
006525-0054690
6,525
female
13-17
3.62
auk þess sækir þú um brauð
006525-0054691
6,525
female
13-17
5.06
við verðum að stöðva loka áður en hann nær að eyðileggja göngin
006514-0054692
6,514
male
13-17
6.59
húsið hans kuba er í brekku og þar er stór kjallari
006514-0054693
6,514
male
13-17
5.67
öll pörin á gólfinu eru á leiðinni í sömu átt
006514-0054694
6,514
male
13-17
4.46
nú sting ég olnbogunum aftur í dýnuna
006514-0054695
6,514
male
13-17
11.33
vopnaeftirlitsmenn sameinuðu þjóðirnar hafi verið lítt dulbúnir njósnarar að gæta hagsmuna bandaríkjastjórnar
006514-0054696
6,514
male
13-17
2.93
nú er gaman að lifa
006526-0054702
6,526
female
13-17
4.13
best að flýta sér að setja glasið á borð
006526-0054703
6,526
female
13-17
4.04
orðið ástríðuglæpur er ekki lengur til hjá lögreglunni
006526-0054705
6,526
female
13-17
3.58
þó að ekkert bendi til þess
006526-0054706
6,526
female
13-17
4.32
þá gæti verið betra að hafa nýjan mann með honum
006527-0054707
6,527
male
13-17
3.67
ég skal laga te
006527-0054708
6,527
male
13-17
3.62
en hafðir viðkomu á salerninu
006527-0054709
6,527
male
13-17
4.46
best er að sorgin sé ósýnileg
006527-0054710
6,527
male
13-17
4.6
hún er öll varðveitt í flateyjarbók
006528-0054712
6,528
male
13-17
3.76
það væri ekki rökrétt
006528-0054713
6,528
male
13-17
3.53
bjarni starði á hana
006528-0054714
6,528
male
13-17
3.48
þeir eru áfengir og ávanabindandi
006528-0054715
6,528
male
13-17
2.97
hans dauði er ekki sorglegur
006528-0054716
6,528
male
13-17
5.48
god dag sagði hann kumpánalega þegar ég gekk hjá
006529-0054717
6,529
female
13-17
3.11
sama hvað við reynum
006529-0054719
6,529
female
13-17
4.09
þetta eru ekki bestu skilyrðin til að kynnast honum
006529-0054720
6,529
female
13-17
3.95
jafnvel mýksta veðurfar virtist gera hann veikan
006529-0054721
6,529
female
13-17
4.32
þetta voru leiðindaspurningar en það yrði ekki við öllu séð
006530-0054727
6,530
female
13-17
2.74
hver er sá maður
006530-0054728
6,530
female
13-17
3.16
brjálæði sagði stenko
006530-0054729
6,530
female
13-17
4.55
það er engin tilviljun að farsíminn minn er enn á gistiheimilinu
006530-0054730
6,530
female
13-17
3.3
hvur djöfullinn sagði guttormur
006530-0054731
6,530
female
13-17
4.6
á hinn bóginn get ég ekki verið þekkt fyrir að hafna þessu boði
006531-0054734
6,531
male
13-17
2.93
hafði fólk grátið í jarðarför herdísar
006532-0054737
6,532
female
13-17
3.63
með og á móti
006532-0054738
6,532
female
13-17
4.91
hann byrjaði að ýlfra og svo heyrðist vængjasláttur
006532-0054739
6,532
female
13-17
7.72
engin húðflúr hugsaði hún eins og fangavörður að taka á móti nýjum fanga
006532-0054740
6,532
female
13-17
4.78
klukkan var hálftíu að morgni sunnudags
006532-0054741
6,532
female
13-17
3.41
það borgaði sig varla
006533-0054744
6,533
male
13-17
4.6
það var eitthvað svo ófullkomið við að vita ekki meira um hann
006533-0054745
6,533
male
13-17
4.04
hún nefndist nasser og réð ríkjum í egyptalandi
006533-0054746
6,533
male
13-17
2.41
bjarni sagði ekki orð
006534-0054753
6,534
male
13-17
2.6
þú rannsakar óvætti
006534-0054754
6,534
male
13-17
3.95
kannt þú að garga eins og krían
006534-0054755
6,534
male
13-17
4.37
það er tekið við patrik þegar hann er fjórtán ára
006535-0054757
6,535
female
13-17
3.95
ekki mér veðja ég
006535-0054758
6,535
female
13-17
6.08
átta mig svo á því að hver verður að hafa sinn smekk
006535-0054759
6,535
female
13-17
5.53
hvar skyldi hún hafa verið fyrstu jólin eftir skilnaðinn
006535-0054760
6,535
female
13-17
4.92
ég skal fara með honum sagði guttormur
006535-0054761
6,535
female
13-17
4.41
þau ganga hreint til verks hugsar hún
006537-0054768
6,537
female
13-17
2.79
ein í heiminum
006537-0054769
6,537
female
13-17
3.72
en ég verð að finna sólina sagði sóla
006537-0054770
6,537
female
13-17
3.39
hann hefur aldrei fyrirgefið okkur
006538-0054772
6,538
male
13-17
6.36
ronnie og florian höfðu verið með honum allan tímann
006538-0054773
6,538
male
13-17
7.43
en fólkið sem þeir eiga að starfa fyrir verðskuldar aðra og betri foringja
006538-0054774
6,538
male
13-17
5.62
eigandinn gerði sem betur fer ekki vart við sig á meðan
006538-0054775
6,538
male
13-17
2.69
þeir eru keppinautar
006538-0054776
6,538
male
13-17
3.39
ég skal vaka og passa þig
006539-0054777
6,539
male
13-17
3.58
ég ræði betur við þig seinna sagði hún

Dataset Card for samromur_children

Dataset Summary

The Samrómur Children Corpus consists of audio recordings and metadata files containing prompts read by the participants. It contains more than 137000 validated speech-recordings uttered by Icelandic children.

The corpus is a result of the crowd-sourcing effort run by the Language and Voice Lab (LVL) at the Reykjavik University, in cooperation with Almannarómur, Center for Language Technology. The recording process has started in October 2019 and continues to this day (Spetember 2021).

Example Usage

The Samrómur Children Corpus is divided in 3 splits: train, validation and test. To load a specific split pass its name as a config name:

from datasets import load_dataset
samromur_children = load_dataset("language-and-voice-lab/samromur_children")

To load an specific split (for example, the validation split) do:

from datasets import load_dataset
samromur_children = load_dataset("language-and-voice-lab/samromur_children",split="validation")

Supported Tasks

automatic-speech-recognition: The dataset can be used to train a model for Automatic Speech Recognition (ASR). The model is presented with an audio file and asked to transcribe the audio file to written text. The most common evaluation metric is the word error rate (WER).

Languages

The audio is in Icelandic. The reading prompts were gathered from a variety of sources, mainly from the Icelandic Gigaword Corpus. The corpus includes text from novels, news, plays, and from a list of location names in Iceland. The prompts also came from the Icelandic Web of Science.

Dataset Structure

Data Instances

{
  'audio_id': '015652-0717240', 
  'audio': {
    'path': '/home/carlos/.cache/HuggingFace/datasets/downloads/extracted/2c6b0d82de2ef0dc0879732f726809cccbe6060664966099f43276e8c94b03f2/test/015652/015652-0717240.flac', 
    'array': array([ 0.        ,  0.        ,  0.        , ..., -0.00311279,
       -0.0007019 ,  0.00128174], dtype=float32), 
    'sampling_rate': 16000
  }, 
  'speaker_id': '015652', 
  'gender': 'female', 
  'age': '11', 
  'duration': 4.179999828338623, 
  'normalized_text': 'eiginlega var hann hin unga rússneska bylting lifandi komin'
}

Data Fields

  • audio_id (string) - id of audio segment
  • audio (datasets.Audio) - a dictionary containing the path to the audio, the decoded audio array, and the sampling rate. In non-streaming mode (default), the path points to the locally extracted audio. In streaming mode, the path is the relative path of an audio inside its archive (as files are not downloaded and extracted locally).
  • speaker_id (string) - id of speaker
  • gender (string) - gender of speaker (male or female)
  • age (string) - range of age of the speaker: Younger (15-35), Middle-aged (36-60) or Elderly (61+).
  • duration (float32) - duration of the audio file in seconds.
  • normalized_text (string) - normalized audio segment transcription.

Data Splits

The corpus is split into train, dev, and test portions. Lenghts of every portion are: train = 127h25m, test = 1h50m, dev=1h50m.

To load an specific portion please see the above section "Example Usage".

Dataset Creation

Curation Rationale

In the field of Automatic Speech Recognition (ASR) is a known fact that the children's speech is particularly hard to recognise due to its high variability produced by developmental changes in children's anatomy and speech production skills.

For this reason, the criteria of selection for the train/dev/test portions have to take into account the children's age. Nevertheless, the Samrómur Children is an unbalanced corpus in terms of gender and age of the speakers. This means that the corpus has, for example, a total of 1667 female speakers (73h38m) versus 1412 of male speakers (52h26m).

These unbalances impose conditions in the type of the experiments than can be performed with the corpus. For example, a equal number of female and male speakers through certain ranges of age is impossible. So, if one can't have a perfectly balance corpus in the training set, at least one can have it in the test portion.

The test portion of the Samrómur Children was meticulously selected to cover ages between 6 to 16 years in both female and male speakers. Every of these range of age in both genders have a total duration of 5 minutes each.

The development portion of the corpus contains only speakers with an unknown gender information. Both test and dev sets have a total duration of 1h50m each.

In order to perform fairer experiments, speakers in the train and test sets are not shared. Nevertheless, there is only one speaker shared between the train and development set. It can be identified with the speaker ID=010363. However, no audio files are shared between these two sets.

Source Data

Initial Data Collection and Normalization

The data was collected using the website https://samromur.is, code of which is available at https://github.com/cadia-lvl/samromur. The age range selected for this corpus is between 4 and 17 years.

The original audio was collected at 44.1 kHz or 48 kHz sampling rate as *.wav files, which was down-sampled to 16 kHz and converted to *.flac. Each recording contains one read sentence from a script. The script contains 85.080 unique sentences and 90.838 unique tokens.

There was no identifier other than the session ID, which is used as the speaker ID. The corpus is distributed with a metadata file with a detailed information on each utterance and speaker. The madata file is encoded as UTF-8 Unicode.

The prompts were gathered from a variety of sources, mainly from The Icelandic Gigaword Corpus, which is available at http://clarin.is/en/resources/gigaword. The corpus includes text from novels, news, plays, and from a list of location names in Iceland. The prompts also came from the Icelandic Web of Science.

Annotations

Annotation process

Prompts were pulled from these corpora if they met the criteria of having only letters which are present in the Icelandic alphabet, and if they are listed in the DIM: Database Icelandic Morphology.

There are also synthesised prompts consisting of a name followed by a question or a demand, in order to simulate a dialogue with a smart-device.

Who are the annotators?

The audio files content was manually verified against the prompts by one or more listener (summer students mainly).

Personal and Sensitive Information

The dataset consists of people who have donated their voice. You agree to not attempt to determine the identity of speakers in this dataset.

Considerations for Using the Data

Social Impact of Dataset

This is the first ASR corpus of Icelandic children.

Discussion of Biases

  • The utterances were recorded by a smartphone or the web app.

  • Participants self-reported their age group, gender, and the native language.

  • Participants are aged between 4 to 17 years.

  • The corpus contains 137597 utterances from 3175 speakers, totalling 131 hours.

  • The amount of data due to female speakers is 73h38m, the amount of data due to male speakers is 52h26m and the amount of data due to speakers with an unknown gender information is 05h02m

  • The number of female speakers is 1667, the number of male speakers is 1412. The number of speakers with an unknown gender information is 96.

  • The audios due to female speakers are 78993, the audios due to male speakers are 53927 and the audios due to speakers with an unknown gender information are 4677.

Other Known Limitations

"Samrómur Children: Icelandic Speech 21.09" by the Language and Voice Laboratory (LVL) at the Reykjavik University is licensed under a Creative Commons Attribution 4.0 International (CC BY 4.0) License with the hope that it will be useful, but WITHOUT ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

Additional Information

Dataset Curators

The corpus is a result of the crowd-sourcing effort run by the Language and Voice Lab (LVL) at the Reykjavik University, in cooperation with Almannarómur, Center for Language Technology. The recording process has started in October 2019 and continues to this day (Spetember 2021). The corpus was curated by Carlos Daniel Hernández Mena in 2021.

Licensing Information

CC-BY-4.0

Citation Information

@misc{menasamromurchildren2021,
      title={Samrómur Children Icelandic Speech 1.0}, 
      ldc_catalog_no={LDC2022S11},
      DOI={https://doi.org/10.35111/frrj-qd60},
      author={Hernández Mena, Carlos Daniel and Borsky, Michal and Mollberg, David Erik  and Guðmundsson, Smári Freyr and Hedström, Staffan and Pálsson, Ragnar and Jónsson, Ólafur Helgi and Þorsteinsdóttir, Sunneva and Guðmundsdóttir, Jóhanna Vigdís and Magnúsdóttir, Eydís Huld and Þórhallsdóttir, Ragnheiður and Guðnason, Jón},
      publisher={Reykjavík University}
      journal={Linguistic Data Consortium, Philadelphia},
      year={2019},
      url={https://catalog.ldc.upenn.edu/LDC2022S11},
}

Contributions

This project was funded by the Language Technology Programme for Icelandic 2019-2023. The programme, which is managed and coordinated by Almannarómur, is funded by the Icelandic Ministry of Education, Science and Culture.

The verification for the dataset was funded by the the Icelandic Directorate of Labour's Student Summer Job Program in 2020 and 2021.

Special thanks for the summer students for all the hard work.

Downloads last month
82