audio_id
stringlengths
19
19
audio
audioduration (s)
3.01
8.5
speaker_id
stringclasses
34 values
gender
stringclasses
2 values
duration
float32
3.01
8.5
normalized_text
stringlengths
26
160
MSRTS_M_17_TS_00001
M_17
male
5.92
ignazio saverio mifsud kien qed jippjana kien qed iħejji tliet volumi tal-biblijoteka maltese
MSRTS_F_08_TS_00002
F_08
female
4.52
ma tiftakarhiex lin-nanna pa tgħid xi ħaġa xi talba jew hekk waqt li qed
MSRTS_F_16_TS_00058
F_16
female
3.97
tista' tiġi ssuperata biex nibqgħu nużawha din il-lingwa
MSRTS_M_16_TS_00009
M_16
male
5.76
il-għaliex dawk huma l-attributi tal-ġustizzja u aħna mdorrijin l-allegorija narawha bħala din
MSRTS_M_04_TS_00006
M_04
male
5.81
qed nisimgħu mhux il-leħen tagħna biss imma il-leħen ta' bejn wieħed u ieħor elf sena
MSRTS_F_17_TS_00051
F_17
female
6.32
fl-elf disa' mija wieħed u tletin u fl-elf disa' mija ħamsa u sittin ninu cremona kiteb kelmtejn fuqha
MSRTS_M_14_TS_00027
M_14
male
4.7
e naħseb illi illi tapprezza illi illi lingwa ilsien
MSRTS_F_08_TS_00003
F_08
female
5.22
tiftakarha kif kienu jinterpretawha wkoll tiftakarha kif kienu jgħiduha
MSRTS_M_17_TS_00002
M_17
male
6.61
isimha kulma huwa ġabra tal-kitbiet maltin bil-malti jew bit-taljan u bil-latin
MSRTS_M_15_TS_00020
M_15
male
5.46
ifhimni inti meta speċjalment għandek hemm das-suġġetti l-kbar orrajt fuq erbat elef
MSRTS_F_16_TS_00059
F_16
female
5.48
jiena però fil-fil-ġu-fil-fl-ambitu tiegħi tal-ġurnaliżmu kont mgħallma li meta niġi biex nitkellem
MSRTS_F_15_TS_00004
F_15
female
5.33
e skont il-bżonnijiet tagħhom kif nafu mhux kulħadd jitkellem l-ingliż l-isptar fil-wards
MSRTS_M_04_TS_00007
M_04
male
3.89
ta' użu tal-lingwi u għalhekk dawn ma nistgħux ngħidulu fejn ġejjin
MSRTS_M_16_TS_00030
M_16
male
5.98
u u il-maġija kif nafu fir-rakkont jiġifieri anka fir-rakkont e e sagru sewwa fil-bibbja
MSRTS_F_17_TS_00052
F_17
female
6.8
bla ma għaraf li hi ballata isemmiha wkoll tancred borg ħarġa tal-berqa tal-elf disa' mija ħamsa u tletin
MSRTS_F_02_TS_00008
F_02
female
6.09
hemm huma komuni pereżempju jekk qed nitkellmu fuq l-imħabba l-abort e t-terroriżmu l-vjolenza
MSRTS_M_14_TS_00028
M_14
male
4.02
hija għodda importanti jekk irridu l-integrazzjoni
MSRTS_F_03_TS_00001
F_03
female
5.83
li tixtieq insaqsik kemm bħala żagħżugħa taħseb li s-segwaċi tal-brikkuni
MSRTS_M_17_TS_00003
M_17
male
5.81
jew inkella noti bijografiċi fuq l-awturi hemm u it-tieni volum kien lest
MSRTS_F_16_TS_00080
F_16
female
4.05
bħala lingwa barranija l-istess kif pereżempju konna nistudjaw il-franċiż u t-taljan
MSRTS_F_08_TS_00004
F_08
female
3.01
għax għidtli irranġajtha ftit tiftakar forsi
MSRTS_M_16_TS_00031
M_16
male
5.36
li jagħmel l-għeġubijiet biss il-maġija ġabet taqliba sħiħa fl-ordni soċjali
MSRTS_M_06_TS_00010
M_06
male
5.76
imma hemm gruppi oħrajn illi illi naħseb raw it-tip ta' liberazzjoni
MSRTS_F_02_TS_00009
F_02
female
4.17
ta' dawn huma temi li m' għandekx m' għandux x' jaqsam inti l-pajjiż fejn qiegħed u
MSRTS_M_04_TS_00008
M_04
male
5.89
il-libertà li tesprimi ruħek imma f' qafas illi waqt li jeħilsek jaqflek ukoll
MSRTS_F_17_TS_00053
F_17
female
5.34
dan induna bis-sunnetti ta' magħha fl-istess żmien il-professur vincenzo lawrenza
MSRTS_M_14_TS_00029
M_14
male
5.56
integrazzjoni e x' ħa jkun il-kontribut tagħkom f' din il-konferenza
MSRTS_M_12_TS_00001
M_12
male
5.17
il-kap tad-dipartiment tal-malti fl-università ta' malta bernard micallef qal
MSRTS_F_03_TS_00002
F_03
female
4.09
u għan-novità għall-novità naħseb li qed jisimgħu il-malti b' dak il-mod
MSRTS_F_16_TS_00081
F_16
female
4.19
kellek problemi larissa biex tgħaddi mill-eżamijiet tal-malti jekk ma jimpurtax insaqsik
MSRTS_M_17_TS_00004
M_17
male
4.23
illum huwa mitluf ħareġ l-ewwel wieħed biss hemm raġunijiet għalfejn ħareġ wieħed biss
MSRTS_F_15_TS_00005
F_15
female
4.59
il-malti m' għamiltx l-ebda kors biex nitgħallmu hemm niftak-
MSRTS_M_06_TS_00011
M_06
male
5.26
li offriet il-lirika tal-brikkuni it-tip ta' inteliġenza tip ta' saffi tip ta'
MSRTS_M_16_TS_00032
M_16
male
6.03
jiġifieri għandek tarbija mwielda f' għar li qed titqies bħala ir-re tar-rejiet allura
MSRTS_M_04_TS_00009
M_04
male
5.9
huwa din it-tensjoni inħalli lill-poplu tiegħi minn
MSRTS_F_17_TS_00054
F_17
female
5.69
dak ittraskriva il-ġurnata ta' gaetano rebull li jsemmi l-qtil ta' patri vinċenz
MSRTS_M_12_TS_00002
M_12
male
5.39
li jaqbel mal-proposta li l-malti jiġi mgħallem bħala lingwa barranija lill-barranin
MSRTS_M_10_TS_00020
M_10
male
4.36
il-kwistjoni tal-immigrazzjoni għaliex claire azzopardi tikteb dwar
MSRTS_F_16_TS_00082
F_16
female
6.57
e li kieku inti però e mhux qed tiġi kkunsidrata din l-option ta bħalissa imma ħa nsaqsihielek teoretikament
MSRTS_M_17_TS_00005
M_17
male
5.47
meta ġabarhom nistgħu inkunu nafu għax intom dawk sibtuhom qegħdin flimkien il-ħdax-il sunnett it-taljan bit-
MSRTS_F_15_TS_00006
F_15
female
3.85
karhom isaqsuni kemm ilek hawn malta u kont ilni xi
MSRTS_M_06_TS_00012
M_06
male
6.17
referenzi u qalu iwa istra illallu x' tista' tagħmel bil-bil-lin-bil-malti jekk tkanta bil-malti
MSRTS_M_16_TS_00033
M_16
male
6.3
dal-element maġiku meta tibda tarah irabbi l-għeruq tiegħu e fi struttura soċjali u jibdilha
MSRTS_F_17_TS_00055
F_17
female
6.25
oliver friġġieri ippublikata fil-ktieb tal-poeżija maltija l-ewwel volum fl-elf disa' mija seba' u tmenin
MSRTS_M_10_TS_00021
M_10
male
6.08
is-sehem il-vuċi l-ilħna tan-nisa differenti għax il-ir-realtà storika titlob
MSRTS_M_12_TS_00003
M_12
male
6.06
madankollu stqarr li tkun daqqa ta' ħarta kbira għall-malti bħala lingwa nazzjonali
MSRTS_F_16_TS_00083
F_16
female
6.63
kellek l-għażla li tistudja l-lingwa maltija e bħala l-lingwa barranija kont teħodha dik l-għażla jew
MSRTS_F_15_TS_00007
F_15
female
3.55
sitt xhur seba' xhur xi ħaġa hekk u diġà kont bdejt
MSRTS_M_10_TS_00022
M_10
male
4.18
li l-letteratura tirreaġixxi għal dawn ir-realtajiet
MSRTS_M_06_TS_00013
M_06
male
5.88
d-din tmissni fil-laħam il-ħaj din il-li lirika lili tmissni ta' vera tmissni
MSRTS_M_16_TS_00034
M_16
male
4.94
tibda tirrealizza li anki hemmhekk għandek e elementi forsi ftit iktar profondi
MSRTS_F_14_TS_00035
F_14
female
4.06
però biex nikkonkludu malli tibda tgħaqqad dawn l-attivitajiet differenti dawn
MSRTS_F_17_TS_00056
F_17
female
5.81
u f' dizzjunarju tat-termini letterarji jikklassifikaha taħt id-daħla tal-ballata
MSRTS_M_07_TS_00023
M_07
male
5.19
sena sħaħ iżur iżur il-morda u l-anzjani fl-isptar mater dei
MSRTS_F_17_TS_00010
F_17
female
5.8
għandna bħal ħaddieħor fi żmienu u bħal ta' qablu fi żmien meta malta ma kienx hawn ġurnali
MSRTS_M_12_TS_00004
M_12
male
5.17
jekk l-eżami tal-malti bħala lingwa barranija ikun jista' jsir minn studenti maltin
MSRTS_F_16_TS_00084
F_16
female
4.9
grazzi grazzi inti larissa talli ingħaqadt magħna u l-kumplament tal-lejl it-tajjeb
MSRTS_F_15_TS_00008
F_15
female
3.93
hemm nagħmel xi kliem flimkien f' sentenza
MSRTS_M_16_TS_00035
M_16
male
6.23
milli naħsbu jiena se nieqaf hawn biex biex ma ngħarraqx il-gost taż-żewġ kelliema tiegħi tgħidx kemm se jieħdu pjaċir li se jitkellmu
MSRTS_M_06_TS_00014
M_06
male
4.06
jiġifieri nista' nisma' r-radiohead jew gruppi oħrajn imma imma din
MSRTS_F_16_TS_00001
F_16
female
6.1
kif għidtilkom magħna għandna lill-ministru għall-edukazzjoni u ix-xogħol il-ministru evarist bartolo grazzi
MSRTS_F_14_TS_00036
F_14
female
4.25
l-istili differenti anke meta tibda tagħmel il-lezzjonijiet inti
MSRTS_M_07_TS_00024
M_07
male
5.97
kiseb rikonoxximent tal-ħidma tiegħu fil-volontarjat kif ukoll il-premju għall-qadi
MSRTS_F_17_TS_00011
F_17
female
6.75
kien jgħanni l-fatt fit-tielet persuna li mhux dejjem tkun omnixxjenti jew inkella mill-perspettiva tal-qatt
MSRTS_M_12_TS_00005
M_12
male
4.86
doctor micallef sostna li ma jaqbilx li s-suġġett tal-malti bħala lingwa barranija
MSRTS_F_16_TS_00085
F_16
female
3.07
hemmhekk kellna jiena nħobb indaħħal dawn l-esperjenzi l-esperjenzi
MSRTS_M_16_TS_00036
M_16
male
5.85
christabelle qalet wow huwa mod kif kif nikkalmahom ftit insomma e però però għamlu biċċa xogħol
MSRTS_M_06_TS_00015
M_06
male
4.46
u allura naħseb huwa importanti li aħna nirriflettu fuq fuq
MSRTS_F_17_TS_00095
F_17
female
6.65
taqsamha f' sittax u jiġi sittax-il faċċata e hemmhekk kellhom xi ħaġa għal xiex jirreferu
MSRTS_F_16_TS_00002
F_16
female
5.16
talli qiegħed magħna hawnhekk llum hawn propju inqalgħet ħafna polemika polemika fuq kummenti
MSRTS_F_14_TS_00037
F_14
female
4.01
hemm xi tkun ir-reazzjoni tal-istudenti jgħidlek imma
MSRTS_M_07_TS_00025
M_07
male
7.37
se ngħid kelma waħda dwar e kit azzopardi wkoll li se jagħmel dan l-intervent kritiku
MSRTS_F_17_TS_00012
F_17
female
6.53
sas-seklu tmintax iltqajna ma' biss disa' għana tal-fatt fosthom l-unika waħda bil-kitba
MSRTS_M_12_TS_00006
M_12
male
3.69
ikun ekwivalenti li qed jiġi mgħallem bħalissa
MSRTS_F_16_TS_00086
F_16
female
5.94
persuna li dak iż-żmien fl-iskola tal-knisja ma kien u din smajtha b' nies li kienu
MSRTS_M_06_TS_00016
M_06
male
7.05
kif il-il-lirika u l-mużika tal-brikkuni jiena lil mario għidtlu xi ftit ġranet ilu li fil-fehma tiegħi il-pass li jmiss huwa
MSRTS_F_17_TS_00096
F_17
female
6.14
u xorta kienu illitterati allura kienu jisimgħu 'il min jaqrawhomlhom jiftakruhom jiftakru
MSRTS_F_14_TS_00038
F_14
female
4.54
nisperaw illi ma ndumux ma nerġgħu niltaqgħu ukoll f' dan il-programm nawguralek ukoll
MSRTS_M_07_TS_00026
M_07
male
6.27
wara l-ewwel intervent ta' ġorġ agius biex imbagħad l-affarijiet jistgħu jibqgħu għaddejjin mela kit azzopardi
MSRTS_F_17_TS_00013
F_17
female
5.65
kieku tafu ħuti x' ġara u li uħud minnhom intmessew sewwa dawn il-ballati
MSRTS_M_12_TS_00007
M_12
male
5.09
id-direttur tal-matsec dario pirotta qal il-bord tal-matsec irid jirrispondi
MSRTS_F_16_TS_00087
F_16
female
4.7
li huma ta' mpari li dak iż-żmien ma kinux jieħdu gost li jitkellmu bil-malti jiġifieri għandna dawn ir-realtajiet
MSRTS_F_16_TS_00041
F_16
female
5.82
mela ħadt il-kummenti ir-reazzjoni inizjali tiegħek ir-ir-reazzjonijiet issa ta' frank adrian qed jgħid hemm
MSRTS_F_17_TS_00097
F_17
female
4.95
kull darba li jgħannuha anka l-istess għannej jibdilha kull darba pereżempju id-danny
MSRTS_F_14_TS_00039
F_14
female
3.59
għaliex għidtli fi żmien tliet ġimgħat issa ukoll e
MSRTS_F_11_TS_00001
F_11
female
6.82
is-suġġett tas-saħta u l-iskunġrar se jkun diskuss f' taħdita għada filgħaxija fl-università ta' malta
MSRTS_M_09_TS_00009
M_09
male
4.59
jevapora bix-xemx aħna nerġgħu nippumpjawh bil-mutur tal-ilma
MSRTS_M_07_TS_00027
M_07
male
8.5
huwa għalliem kittieb żagħżugħ u attivist kulturali fl-em mej tiegħu fil-masters tiegħu fil-letteratura maltija analizza il-poeżiji ta'
MSRTS_M_14_TS_00010
M_14
male
4.24
eżempju għandek limitazzjonijiet kbar fil-fid-dinja diġitali iva
MSRTS_F_17_TS_00014
F_17
female
4.46
huma parti mil-letteratura popolari maltija u mezz dinamiku
MSRTS_F_16_TS_00042
F_16
female
5.13
il-malti għandu jibqa' jiġi mgħallem kif inhu aħna soċjetà waħda u kulħadd għandu jitgħallem il-malti bl-istess mod
MSRTS_F_17_TS_00098
F_17
female
3.33
illum reġa' bidilha mill-aħħar li smajtu jgħanniha
MSRTS_M_09_TS_00030
M_09
male
5.37
u qalli int kieku tieħu ħsieb għidtlu jien nieħu għax jiena ilni naħseb hawn
MSRTS_M_07_TS_00028
M_07
male
6.51
norbert bugeja bħalissa qed jagħmel id-diploma fil-letteratura maltija li bdiet f' ottubru li għadda
MSRTS_M_14_TS_00011
M_14
male
3.99
jiġifieri jiena akkanit wkoll għall-malti inħobbu l-malti e u
MSRTS_F_17_TS_00015
F_17
female
5.36
il-ballati huma simbolu tar-reżistenza tat-tradizzjoni folkloristika orali
MSRTS_M_15_TS_00021
M_15
male
6.64
jiġifieri li semmejna jiġifieri fir-rapport tagħkom inti għandek il-firxa kbira tal-istudenti kollha jiġifieri inti
MSRTS_F_16_TS_00043
F_16
female
5.72
ikkoreġuni naqra bħalissa tfal barranin obbligati li jkollhom il-malti wkoll jew obbligati jkollhom
MSRTS_F_17_TS_00099
F_17
female
6.73
hemm meta ħadha injazju saverju mifsud ħadha bl-iżbalji b' kollox pereżempju
MSRTS_M_09_TS_00031
M_09
male
4
u naf x' irid isir hu għallimni kif irrid nagħmel

Dataset Card for masri_test

Dataset Summary

The MASRI-TEST CORPUS was created out of YouTube videos belonging to the channel of the University of Malta. It has a length of 1 hour and it is gender balanced, as it has the same number of male and female speakers.

Example Usage

The MASRI-TEST contains only the test split:

from datasets import load_dataset
masri_test = load_dataset("MLRS/masri_test")

It is also valid to do:

from datasets import load_dataset
masri_test = load_dataset("MLRS/masri_test",split="test")

Supported Tasks

automatic-speech-recognition: The dataset can be used to test 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 language of the corpus is Maltese.

Dataset Structure

Data Instances

{
  'audio_id': 'MSRTS_M_17_TS_00001', 
  'audio': {
    'path': '/home/carlos/.cache/HuggingFace/datasets/downloads/extracted/9158ecbeeb3532038f3fe3d53e0adda1f790c9363a613bac32c454a39d9c682c/test/male/M_17/MSRTS_M_17_TS_00001.flac', 
    'array': array([ 0.0020752 ,  0.00283813,  0.00167847, ..., -0.0010376 ,
       -0.00091553, -0.00100708], dtype=float32), 
    'sampling_rate': 16000
  }, 
  'speaker_id': 'M_17', 
  'gender': 'male', 
  'duration': 5.920000076293945, 
  'normalized_text': 'ignazio saverio mifsud kien qed jippjana kien qed iħejji tliet volumi tal-biblijoteka maltese'
}

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)
  • duration (float32) - duration of the audio file in seconds.
  • normalized_text (string) - normalized audio segment transcription

Data Splits

The corpus counts just with the test split which has a total of 668 speech files from 17 male speakers and 17 female speakers with a total duration of 1 hour.

Dataset Creation

Curation Rationale

The MASRI-TEST CORPUS (MTSC) has the following characteristics:

  • The MTSC has an exact duration of 1 hours and 0 minutes. It has 668 audio files.
  • The MTSC has recordings from 34 different speakers: 17 men and 17 women.
  • Data in MTSC is classified by speaker. Therefore, all the recordings of each individual speaker are stored in one single directory.
  • Data is also classified according to the gender (male/female) of the speakers.
  • Every audio file in the MTSC has a duration between 3 and 10 seconds approximately.
  • Audio files in the MTSC are distributed in a 16khz@16bit mono format.
  • Transcriptions in MTSC are in lowercase. No punctuation marks are permitted except for dashes (-) and apostrophes (') due to their importance in Maltese orthography.

Source Data

Initial Data Collection and Normalization

The MASRI-TEST CORPUS was possible due to a collaboration of two different Universities. The data selection and audio segmentation was performed by the CIEMPIESS-UNAM Project at the Universidad Nacional Autónoma de México (UNAM) in Mexico City. The audio transcription and corpus edition was performed by the MASRI Team at the University of Malta in the Msida Campus.

Annotations

Annotation process

Proper nouns and other words pronounced in languages other than Maltese (mainly from English, Italian, French and German) were transcribed in their respective orthographic system.

Who are the annotators?

The audio transcription was performed by expert native speakers at the University of Malta in the Msida Campus.

Personal and Sensitive Information

The dataset could contain names revealing the identity of some speakers; on the other side, the recordings come from a publicly repository (YouTube), so, there is not a real intent of the participants to be anonymized. Anyway, you agree to not attempt to determine the identity of speakers in this dataset. Notice: Should you consider that our data contains material that is owned by you and should therefore not be reproduced here?, please:

  • Clearly identify yourself, with detailed contact data such as an address, telephone number or email address at which you can be contacted.
  • Clearly identify the copyrighted work claimed to be infringed.
  • Clearly identify the material that is claimed to be infringing and information reasonably sufficient to allow us to locate the material.
  • Send the request to Carlos Mena Take down: We will comply to legitimate requests by removing the affected sources from the corpus.

Considerations for Using the Data

Social Impact of Dataset

This dataset is challenging because it contains spontaneous speech; so, it will be helpful for the ASR community to evaluate their acoustic models in Maltese with it.

Discussion of Biases

The dataset intents to be gender balanced. It is comprised of 17 male speakers and 17 female speakers.

Other Known Limitations

Neither the MASRI Team or the CIEMPIESS-UNAM Project guarantee the accuracy of this corpus, nor its suitability for any specific purpose. As a matter of fact, a number of errors, omissions and inconsistencies are expected to be found within the corpus.

Dataset Curators

The audio recordings were collected and segmented by students belonging to the social service program "Desarrollo de Tecnologías del Habla", it was curated by Carlos Daniel Hernández Mena and its transcriptions were manually performed by Ayrton-Didier Brincat during 2020.

Licensing Information

CC-BY-NC-SA-4.0. The copyright remains with the original owners of the video. As the data is taken from YouTube, we invoke the same argument of "fair use" as in the Voxlingua107 dataset, which is: "While YouTube users own the copyright to their own videos, using the audio in the videos for training speech recognition models has very limited and transformative purpose and qualifies thus as "fair use" of copyrighted materials. YouTube’s terms of service forbid downloading, storing and distribution of videos. However, the aim of this rule is clearly to forbid unfair monetization of the content by third-party sites and applications. Our dataset contains the videos in segmented audio-only form that makes the monetization of the actual distributed content extremely difficult."

Citation Information

@misc{carlosmenamasritest2020,
      title={MASRI-TEST CORPUS: Audio and Transcriptions in Maltese extracted from the YouTube channel of the University of Malta.}, 
      author={Hernandez Mena, Carlos Daniel and  Brincat, Ayrton-Didier and Gatt, Albert and DeMarco, Andrea and Borg, Claudia and van der Plas, Lonneke and Meza Ruiz, Iván Vladimir},
      journal={MASRI Project, Malta},
      year={2020},
      url={https://huggingface.co/datasets/MLRS/masri_test},
}

Contributions

The authors would like to thank to Alberto Templos Carbajal, Elena Vera and Angélica Gutiérrez for their support to the social service program "Desarrollo de Tecnologías del Habla" at the "Facultad de Ingeniería (FI)" of the Universidad Nacional Autónoma de México (UNAM). We also thank to the social service students for all the hard work during the audio segmentation.

Downloads last month
57
Edit dataset card