Title : Readme file for Kasdi-Merbah(University) Emotional Database in Arabic Speech (KEDAS):

Author(s) :---------------------------------------------------
Mourad Belhadj 
Ilham Bendellali
Elalia Lakhdari


KEDAS is an acted audio database for emotional speech geared towards speech emotion recognition systems.
It was designed in the modern Arabic language, which is symbolized in ISO 639-3 by [arb], and it was developed in two months of work.
-This work is issued by the Faculty of Modern Technologies for Information and Communication of the University of Kasdi Merbah, Ouargla, Algeria.
It is an audio representative database of emotional speech in Arabic which took two months.
Written consent was obtained prior to registration from all participants.

Individuals gave informed consent for participant data (name, surname, and age), where the participants in the database were informed that their personal information would not be
published and would be deleted, except for their audio recordings, moreover, the recordings will be
used in scientific research only. 

-------------------------------------------------------------
Linguistic material:
____________________________________________________________
• Stop now                                         توقف الآن • 
____________________________________________________________
• Please act politely                         تصرف بأدب من فضلك•
____________________________________________________________
• I am so sorry                                    أنا آسف حقا•
____________________________________________________________
• I know this is hard                           أعلم أن هذا صعب•
____________________________________________________________
• Be careful it falls down                      إحذر سيسقط منك•
____________________________________________________________
• Is there another solution?                    آلا يوجد حل آخر•
____________________________________________________________
• This is really great                            هذا رائع حقا•
____________________________________________________________
• Thank you very much                            شكرا جزيلا لك•
____________________________________________________________
• The second office is on the right         المكتب الثاني على اليمين • 
____________________________________________________________
• What can i do for you?                            تفضل •
____________________________________________________________

Overview info :

__________________________________________________________________________ 
  
Data-set name         |  kasdi-merba emotional database in Arabic speech
__________________________________________________________________________
 Date                 |  June  2022
__________________________________________________________________________
 Dataset type        |  Acted   
_________________________________________________________________________
 File Type            |  Audio only
_________________________________________________________________________
 format               |  .FLAC
__________________________________________________________________________
 Sampling rate        |  44 KHz*
__________________________________________________________________________
 ? of speakers        |  500 
__________________________________________________________________________
 ? of female speakers |  254    
_______________ __________________________________________________________
 ? of male speakers   |  246 
__________________________________________________________________________

  Age of speakers     |  [20-70 year]
___________________________________________________________________________
 ? of emotional states|  05 emotions [sadness, fear, anger, happiness, neutral] 
____________________________________________________________________________
 ? of statements      |  10 = [9 sentences + 1 word] 
____________________________________________________________________________
 ? of audio clips     |  5000 audio files
____________________________________________________________________________
 
Size                 |  282 Mb
____________________________________________________________________________

duration of each clip |  [ 0.5 s to 2.5s ]                     
_____________________________________________________________________________
hardware, Software    |  Zoom h8 device, Audacity audio editor                    
_____________________________________________________________________________
   
Duration           |  02:04   
_____________________________________________________________________________
Accessibility         |  public 


*With the exception of Actor 309 which is 48 kHz

All audio files in the database are encrypted as follows :
The name of each file is divided into seven parts in the following order

(1)- Actor id:       First actor number to last actor [001 to 500 ]
---------------------------------------------------------------
(2)- Gender:          m=male, f=female
---------------------------------------------------------------
(3)- Age group:       [01]=20-30 * [02]=30-40 *  [03]=40-50 * (years)
----------------------------------------------------------------
(4)- Emotion:         E1=Angry, E2=Sadness, E3=Fear, E4=happiness, E5=neutrality
----------------------------------------------------------------
(5)- Statement:       [01]=first statment, [02]=second statement.

ACKNOWLEDGEMENTS

The Directorate-General for Scientific Research and Technological Development (DGRSDT) and The Thematic Agency for Research in Science and Technology (ATRST) supported this work. We acknowledge their funding and support. The Linguistic Research Unit and Arabic Language Issues in Algeria (URLCLAA, Ouargla, Algeria) are acknowledged for their contribution to studying emotions in the KEDAS Dataset. The Artificial Intelligence and Information Technologies Laboratory (LENATI, University Kasdi Merbah Ouargla) is acknowledged for its support, direction, and advice.