Datasets:
audio_id
stringlengths 19
19
| audio
audioduration (s) 2.83
7.13
| speaker_id
stringclasses 18
values | gender
stringclasses 2
values | duration
float32 2.83
7.13
| normalized_text
stringlengths 28
155
|
---|---|---|---|---|---|
MSRDV_F_08_DV_00005 | F_08 | female | 5.9 | dawn ma kinux żminijiet sbieħ l-inkwiżizzjoni ddeċidiet li teqred lil kull min |
|
MSRDV_M_02_DV_00008 | M_02 | male | 5.8 | ovvjament kulħadd jaf li hemm l-awtur warajh però e il-pubblikatur ħafna in se inqabblu qisu l-midwife |
|
MSRDV_M_08_DV_00087 | M_08 | male | 6.76 | ta' ġenituri maltin it-tnejn maltin imma jgħixu fi brussel jew fil-lussemburgu għandhom dak iċ-ċans jagħmlu dak il-programm |
|
MSRDV_M_04_DV_00056 | M_04 | male | 6.48 | kif qed taraw ħafna drabi t-teżijiet li nmexxu huma proġetti tar-riċerka li xtaqna kieku ħdima fuqhom aħna stess |
|
MSRDV_F_02_DV_00182 | F_02 | female | 6.28 | sitwazzjoni li waqgħu fl-ingann huma il-din tkun ta' gwadann għall-eroj għall-fatt li |
|
MSRDV_M_09_DV_00013 | M_09 | male | 4.69 | e hi twieġbu bil-bil-bit-taljan tagħha li hu sqalli e |
|
MSRDV_F_08_DV_00004 | F_08 | female | 4.62 | bħal nisa bħali li qatt ma żżewġu u li ssieltu mal-ħajja biex jgħixu |
|
MSRDV_F_06_DV_00022 | F_06 | female | 6.07 | e illum e nitkellmu ħafna fuq il-midja soċjali u kif qed tgħid inti e ħafna bloggers |
|
MSRDV_F_02_DV_00075 | F_02 | female | 6.06 | għaliex tista' tgħid nistgħu naqbdu l-istat soċjali tal-eroj kif minn fqir b' dawn il-l-oġġetti |
|
MSRDV_F_03_DV_00003 | F_03 | female | 5.58 | biex fi speċi e e ir-ir-ras ċkien tal-bejbi u tinżillha iktar |
|
MSRDV_M_02_DV_00030 | M_02 | male | 4.81 | jixtri forsi kotba għat-tfal jew jaqra kotba għat-tfal ħaddieħor jiġifieri huwa kamp vast |
|
MSRDV_M_08_DV_00041 | M_08 | male | 6.39 | dawn se jkunu korsijiet li jibdew minn meta jkollok ħames snin sa meta mill-inqas tagħlaq e sittax-il sena jiġifieri mifrux |
|
MSRDV_F_03_DV_00002 | F_03 | female | 3.53 | kabbar toqba u ċekken ir-ras |
|
MSRDV_M_03_DV_00001 | M_03 | male | 5.81 | diploma fil-letteratura maltija e hija nistgħu niddiskrivuha bħala vjaġġ ta' diversi skoperti |
|
MSRDV_M_04_DV_00057 | M_04 | male | 6.1 | imma la kellna żmien u forsi l-għodda kollha sephora francalanza ħadmet fuq dan il-proġett ta' riċerka |
|
MSRDV_M_09_DV_00014 | M_09 | male | 5.62 | u mbagħad jiġri li grezzja mur imxi tir bħal għasfur id-iddur fuqha jiġifieri |
|
MSRDV_F_02_DV_00183 | F_02 | female | 6.78 | hu ħa jkun jaf x'għandu jagħmel u fl-istess ħin għall-karattri sekondarji ħa jkunu jafu li jekk l-eroj ħa jimxi b' mod |
|
MSRDV_F_01_DV_00020 | F_01 | female | 5.52 | dak li inti jkollok f' moħħok meta tikteb poeżija hu importanti li l-qarrej jifhmu |
|
MSRDV_F_06_DV_00023 | F_06 | female | 5.95 | illum m' għandniex jekk tħares hawn dejjem l-istorja x' għandek il-passat il-preżent |
|
MSRDV_F_02_DV_00076 | F_02 | female | 5.31 | il-anka id-dehra tal-ilbies fejn filli tkun qiegħda seftura |
|
MSRDV_M_02_DV_00009 | M_02 | male | 5.27 | għax huwa dak li jgħin fil-proċess biex il-ktieb jitwieled biex ngħidu hekk biex il-ktieb jasal |
|
MSRDV_F_02_DV_00100 | F_02 | female | 6.36 | dejjem il-grad dejjem jogħla l-fatt li bħalma qalet christabelle l-ajjutant mhux bilfors ikun wieħed |
|
MSRDV_M_03_DV_00002 | M_03 | male | 3.77 | għax fl-ewwel vjaġġ ta' skoperti e kien hemm il-kollegi tiegħi |
|
MSRDV_M_09_DV_00015 | M_09 | male | 4.31 | jiġifieri għandek din dat-tliet safef jiġifieri hemm hemm tradizzjoni |
|
MSRDV_F_02_DV_00184 | F_02 | female | 5.71 | tajjeb u ħa jirnexxilu f' kollox dawn ħa jiggwadanjaw milli jerġgħu jiksbu d-dehra fiżika tagħhom |
|
MSRDV_M_01_DV_00037 | M_01 | male | 3.49 | imma inti tista' tiktibha xi mkien fuq karta biex ħaddieħor snin wara |
|
MSRDV_M_04_DV_00058 | M_04 | male | 5.52 | b' ħila u passjoni kbira imma hemm ħafna xogħol ieħor ta' apprezzament akkademiku li jista' jsir |
|
MSRDV_F_01_DV_00021 | F_01 | female | 4.2 | iġifieri imma fi ftit kliem jew l-interpretazzjoni e |
|
MSRDV_F_08_DV_00006 | F_08 | female | 5.22 | jiddiletta bil-maġija smajt minn sorsi fdati li barra minn malta |
|
MSRDV_M_04_DV_00012 | M_04 | male | 3.59 | għalhekk it-taħdita tal-lum taqbad karatteristika waħda biss |
|
MSRDV_F_06_DV_00024 | F_06 | female | 5.11 | illi ta' wieħed ikompli jiskopri u jixtieqek ħafna tkompli ukoll taqsam l-esperjenzi |
|
MSRDV_F_02_DV_00077 | F_02 | female | 5.95 | u allura l-ilbies tagħha jkun imkerrah u bla valur minn biċċa minn biċċa lewża |
|
MSRDV_M_02_DV_00031 | M_02 | male | 3.64 | e sa ftit snin ilu hemm storja tal-letteratura tat-tfal kif qed ngħidu |
|
MSRDV_M_08_DV_00043 | M_08 | male | 5.89 | aħna meta qed nitkellmu fuq l-iskola primarja u sekondarja il-biċċa l-kbira inti titkellem fuq tliet livelli mill-ewwel livell |
|
MSRDV_F_02_DV_00118 | F_02 | female | 5.48 | filli l-protagonista jitlaq b' xejn fuq avventura mingħajr ħjiel ta' xejn ma jieħu xejn miegħu |
|
MSRDV_M_03_DV_00003 | M_03 | male | 5.48 | li għenuni nasal f' din l-ewwel sena tal-kors u l-letturi li għallmuna wkoll |
|
MSRDV_M_01_DV_00038 | M_01 | male | 4.48 | ikun kapaċi jdoqqha hemm u dik dejjem kellha rabta kbira mal- |
|
MSRDV_F_03_DV_00004 | F_03 | female | 3.46 | u kont nisma' lilha tgħid u mbagħad meta jiena kbirt |
|
MSRDV_F_02_DV_00185 | F_02 | female | 5.57 | għalhekk nistgħu naraw li għalkemm nistgħu naraw li hemm sofferenza tal-metamorfosi negattiva |
|
MSRDV_F_08_DV_00007 | F_08 | female | 5.9 | li nstabu ħajta jew ġew mixlija b' atti ta' maġija sfaw maħruqa ħajjin |
|
MSRDV_M_04_DV_00013 | M_04 | male | 4.93 | imma karatteristika li jidhrilna li hija fundamentali tal-lirika ta' mario vella |
|
MSRDV_F_01_DV_00022 | F_01 | female | 4.6 | u inti trid tgħidli x' fhimt jew l-ewwel ħaġa li tiġik f' rasek xħin naqrahielek |
|
MSRDV_F_02_DV_00078 | F_02 | female | 3.06 | toħroġ l-aqwa lbies u hekk tikseb tkun il-prinċipessa |
|
MSRDV_M_02_DV_00032 | M_02 | male | 4.46 | hi sejra din snin twal però dejjem kienet meqjusa qisha xi ħaġa |
|
MSRDV_M_08_DV_00044 | M_08 | male | 6.56 | għat-tieni livell għat-tielet livell hu x' inhu s-suġġett u x' inhu l-ilsien inti trid tagħmlu trid tagħmlu dak il-livell |
|
MSRDV_F_02_DV_00119 | F_02 | female | 5.93 | mingħalih li ħa jkompli l-avventura tiegħu f' nofs il-vjaġġ ħa jġarraf dak ħa jikseb dak kollu li għandu bżonn |
|
MSRDV_M_03_DV_00004 | M_03 | male | 4.04 | it-tieni skoperta kienet e ħa ngħid hekk tul il-letteratura |
|
MSRDV_M_01_DV_00039 | M_01 | male | 5.43 | mal-lingwa fil-fatt wieħed mill-ik-l-eqdem eżempji li hemm huwa mill-mil-lvant nofsani |
|
MSRDV_M_04_DV_00014 | M_04 | male | 5.43 | jiġifieri il-mod kif juża il-mod kif kif kif tuża din il-lirika mod |
|
MSRDV_F_02_DV_00186 | F_02 | female | 6.28 | u tista' tgħid li bihom sibna ħafna komuni kemm fiż-żewġ elementi tista' tgħid ukoll li |
|
MSRDV_F_08_DV_00008 | F_08 | female | 4.15 | kienet is-sena elf sitt mija ħamsa u għoxrin u jien |
|
MSRDV_F_03_DV_00005 | F_03 | female | 5.83 | issa ilni naqra f' spanja u wliedi marru jixtru dejjem kont ngħidha dejjem |
|
MSRDV_F_01_DV_00023 | F_01 | female | 6.74 | għaliex das-seħer ta' ferħ meta naf li m' għandi xejn forsi għax jiena għandi lili imma lili għada se jmut |
|
MSRDV_F_02_DV_00079 | F_02 | female | 6.67 | apparti minn hekk bħalma semmejt ukoll fil-vjaġġ temporanju eee tista' tgħid li l-oġġetti inanimati wkoll |
|
MSRDV_F_02_DV_00140 | F_02 | female | 6.44 | mhijiex il-viljakk hija pjuttost imma karattru sekondarju innoċenti mentri l-viljakk |
|
MSRDV_M_08_DV_00045 | M_08 | male | 6.63 | jiġifieri għalhekk mhuwiex veru li tbaxxija ta' standards u hekk huma tiġi mgħallem b' mod differenti pereżempju illum għandna |
|
MSRDV_F_08_DV_00009 | F_08 | female | 5.64 | u disa' u tletin mara oħra sfajna akkużati li ppreparajna u bigħhejna |
|
MSRDV_M_04_DV_00015 | M_04 | male | 3.28 | verament intelliġenti u ħafna drabi sovversiv |
|
MSRDV_F_03_DV_00006 | F_03 | female | 3.39 | minn tat-tfal jew hekk hekk kont tgħidha toqgħod tgħidha |
|
MSRDV_M_02_DV_00033 | M_02 | male | 3.62 | żbaljatament jiġifieri qisha xi ħaġa eħfef qisu li għat-tfal |
|
MSRDV_M_08_DV_00046 | M_08 | male | 6.51 | tista' titgħallem it-taljan tista' titgħallem il-franċiż tista' titgħallem l-ispanjol tista' titgħallem il-ġermaniż dak ukoll f' livell |
|
MSRDV_F_02_DV_00141 | F_02 | female | 6.55 | jieħu xeħta ta' xebba s-sewda allura hekk ħafna l-mudlam l-għeluq il-kruha |
|
MSRDV_F_02_DV_00034 | F_02 | female | 6.41 | x' għandu jagħmel allura inti l-fatt li ħa q-qed jgħidlek x' inti tagħmel qed jagħtik twissija ta' x' inhu ġej |
|
MSRDV_M_04_DV_00016 | M_04 | male | 4.06 | il-lingwa ta' kuljum il-frażijiet l-espressjonijiet l-idjomi |
|
MSRDV_F_03_DV_00007 | F_03 | female | 3.85 | fekkek waqt waqt kabbar it-toqba u ċekken ir-ras |
|
MSRDV_M_02_DV_00034 | M_02 | male | 4.7 | però jintmissu temi jiġifieri anka malta stess jiġifieri pereżempju kif qed tgħid int pereżempju |
|
MSRDV_M_08_DV_00047 | M_08 | male | 6.35 | fil-livell tlieta irid ikun e u illum jiena ngħidha din meta għandna t-tfal barranin li jitgħallmu fostna jistgħu jitgħallmu |
|
MSRDV_F_02_DV_00142 | F_02 | female | 5.51 | ħafna jiġu assoċjati mal-viljakk u dan jista' juri li l-ħrejjef xi kultant ikun għandhom huma ra- |
|
MSRDV_F_02_DV_00035 | F_02 | female | 6.03 | jibda jgħidlu għalik apparti minn hekk id-donatur jagħti wkoll l-oġġetti maġiċi l-armi li bihom |
|
MSRDV_M_08_DV_00048 | M_08 | male | 4.61 | t-taljan jew il-franċiż jew il-ġermaniż jew l-ispanjol għaliex m' għandhomx ukoll jistgħu jingħataw u jitgħallmu |
|
MSRDV_F_02_DV_00036 | F_02 | female | 6.64 | ħa ħa jkun jista' jiskansa milli jaqa' taħt xi xkiel tal-eroj apparti minn hekk ukoll |
|
MSRDV_M_04_DV_00017 | M_04 | male | 4.56 | il-qwiel il-metafori li ninqdew bihom biex nitkellmu biex nifhmu |
|
MSRDV_M_02_DV_00035 | M_02 | male | 5.64 | hemm kotba pereżempju dan jiġini f' moħħi il-qtates ta' max-xatt li jittratta b' mod delikat għat-tfal |
|
MSRDV_F_04_DV_00001 | F_04 | female | 5.66 | grazzi mill-ġdid lil ruth castillo aħna se nkomplu bid-diskussjonijiet tagħna |
|
MSRDV_F_02_DV_00143 | F_02 | female | 6.19 | zzisti wkoll għax il-il-l-iswed il-ħażin l-abjad it-tajjeb bħal hawnhekk il-li |
|
MSRDV_M_05_DV_00010 | M_05 | male | 4.43 | jiġifieri dan kollu bħala konklużjoni jiġifieri in-nies li jmorru je- |
|
MSRDV_M_04_DV_00018 | M_04 | male | 4.93 | biex nirraġunaw biex nispjegaw biex nisfidaw id-dinja li ngħixu fiha |
|
MSRDV_M_08_DV_00049 | M_08 | male | 5.65 | jiġifieri anzi jien ngħid li f' dak il-każ il-malti e għandu jkun obbligatorju imma iktar minn dan kollu irridu ngħidu |
|
MSRDV_F_02_DV_00144 | F_02 | female | 6.86 | apparti tista' tkun razzista wkoll sa ċertu punt turi wkoll is-sempliċità tagħha għax hekk turi li m' hemmx żvilupp fiha |
|
MSRDV_F_04_DV_00019 | F_04 | female | 3.71 | dan bħal meta televiżjoni aħna nagħmlu l-programmi |
|
MSRDV_F_02_DV_00037 | F_02 | female | 5.58 | għandna l-ħarsa lura fil-vjaġġ din tidher ukoll għall-fatt fl-istess laqgħa |
|
MSRDV_M_08_DV_00003 | M_08 | male | 4.21 | periklu li qed jiffaċċja ilsienna illum huwa periklu reali għax |
|
MSRDV_M_04_DV_00019 | M_04 | male | 5.78 | imma din il-materja prima rikka il-liriċista jħaddimha sew qabel ma jużaha |
|
MSRDV_M_08_DV_00070 | M_08 | male | 5.67 | li jgħidulek sewwa mela t-tfal maltin imwelldin malta minn żewġ ġenituri maltin u qed iqisu |
|
MSRDV_F_02_DV_00145 | F_02 | female | 5.81 | għax inti jekk tara affari-tibda tiffamiljarizza magħhom jekk tara l-abjad tgħid mela sens tajjeb |
|
MSRDV_F_02_DV_00038 | F_02 | female | 6.77 | mal-ajjutant jew id-donatur meta jispjega apparti li kif ħa naraw l-ajjutant għandu l-esperjenza |
|
MSRDV_M_08_DV_00004 | M_08 | male | 6.89 | li noqogħdu nivvintaw u nimmaġinaw perikli oħra li jien ngħid m' għandniex il-lussu għax tant hemm affarijiet serji li jridu jsiru |
|
MSRDV_M_04_DV_00040 | M_04 | male | 5.14 | kienet ir-rifjut tagħhom li jagħmlu dak li għamlu l-biċċa l-kbira tal-kanzunetti bil-malti |
|
MSRDV_M_08_DV_00088 | M_08 | male | 6.29 | dak il-programm ma jistgħux jagħmluh ukoll tfal ta' ġenituri f' sitwazzjoni li jista' jkun jiswa simili |
|
MSRDV_F_02_DV_00146 | F_02 | female | 5.53 | tara l-iswed tgħid sens ħażin u wkoll peress li għidna fil-bidu li hija kienet tingħad minn |
|
MSRDV_F_02_DV_00039 | F_02 | female | 6.82 | fih diġà qed turik li hu jista' jagħti l-każ li għadda minnha u ma rnexxilux allura hu jipprova |
|
MSRDV_M_08_DV_00005 | M_08 | male | 5.43 | u affarijiet veri li jaħasra li nintilfu fuq l-immaġinazzjoni e ma tagħmlilniex ġid |
|
MSRDV_M_01_DV_00020 | M_01 | male | 4.51 | biex tikklikkja fuq l-internet li ċertu affarijiet ma nindunawx bihom |
|
MSRDV_M_04_DV_00041 | M_04 | male | 4.77 | li jużaw l-istess frażijiet għajjenin l-istess xbihat sofor biż-żmien |
|
MSRDV_M_08_DV_00089 | M_08 | male | 6.04 | hekk qed insaqsi imma żgur hu li rridu insaħħu right across mingħajr ma nbaxxu l-istandard |
|
MSRDV_F_02_DV_00060 | F_02 | female | 6.01 | biex jieħu id bintu u apparti minn hekk x' jakkwista mhux biss l-imħabba imma forsi minn |
|
MSRDV_F_02_DV_00147 | F_02 | female | 5.42 | f' dinja ta' nies illeterati dawn il-konnotazzjonijiet kienu jgħinu aktar lin-nies jifhmu |
|
MSRDV_M_08_DV_00006 | M_08 | male | 4.4 | l-ikbar riskju li qed jiffaċċja pajjiżna huwa l-internet onlajn |
|
MSRDV_F_02_DV_00101 | F_02 | female | 6.62 | it-tliet aħwa u it-tlett ixjuħ rajt element paradossali hawnhekk għaliex l-iżgħar fost l-aħwa |
|
MSRDV_M_01_DV_00021 | M_01 | male | 4.41 | jiġifieri għandek kantawturi li huma forsi stabbiliti bħal corazon għall-internet għandek |
Dataset Card for masri_dev
Dataset Summary
The MASRI-DEV 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-DEV contains only the validation split:
from datasets import load_dataset
masri_dev = load_dataset("MLRS/masri_dev")
It is also valid to do:
from datasets import load_dataset
masri_dev = load_dataset("MLRS/masri_dev",split="validation")
Supported Tasks
automatic-speech-recognition: The dataset can be used as a development/validation corpus when training 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': 'MSRDV_F_08_DV_00005',
'audio': {
'path': '/home/carlos/.cache/HuggingFace/datasets/downloads/extracted/ebfc6f70292385c6e730ccadc575de507963960c28b4faf2d8bbfcd61589d2cc/dev/female/F_08/MSRDV_F_08_DV_00005.flac',
'array': array([ 0.00140381, 0.00210571, 0.00189209, ..., -0.00210571,
-0.00289917, -0.00372314], dtype=float32),
'sampling_rate': 16000
},
'speaker_id': 'F_08',
'gender': 'female',
'duration': 5.900000095367432,
'normalized_text': 'dawn ma kinux żminijiet sbieħ l-inkwiżizzjoni ddeċidiet li teqred lil kull min'
}
Data Fields
audio_id
(string) - id of audio segmentaudio
(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 speakergender
(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 validation split which has a total of 669 speech files from 9 male speakers and 9 female speakers with a total duration of 1 hour.
Dataset Creation
Curation Rationale
The MASRI-DEV CORPUS (MDVC) has the following characteristics:
- The MDVC has an exact duration of 1 hours and 0 minutes. It has 669 audio files.
- The MDVC has recordings from 18 different speakers: 9 men and 9 women.
- Data in MDVC 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 MDVC has a duration between 3 and 10 seconds approximately.
- Audio files in the MDVC are distributed in a 16khz@16bit mono format.
- Transcriptions in MDVC 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-DEV 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 validate their acoustic models in Maltese with it.
Discussion of Biases
The dataset intents to be gender balanced. It is comprised of 9 male speakers and 9 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{carlosmenamasridev2020,
title={MASRI-DEV 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_dev},
}
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
- 72