38

Journal  of  ICSAR
ISSN (print): 2548-8619; ISSN (online): 2548-8600
Volume 3 Number 2 July 2019: 38-40

Implementation of Speech to-Text Application for Deaf Students on 
Inclusive Education Course

Irdamurni, Johandri Taufan

Universitas Negeri Padang, Indonesia
Email:  irdamurni@fipunp.ac.id

Abstract: Inclusive education is an educational services system where people with disabilities can attend 
to the nearest regular school of their place, included college. Persons with hearing impairment is one of 
the minorities who is still marginalized from the right, including the needs of access to communications 
and information technology. Because of that, people with hearing impairment are not able to participate 
in instructional activities. Based on the observation and the interview, to graduate and undergraduate 
students with hearing impairment of Special Education department, they said that they still have problems 
in communicating with lecturers or other students, because of their inability to understand the explanation 
of the lecturer. This paper aims at providing an analysis of using speech to text app in an inclusive 
education course, which can record the voice of a lecturer and converted to text that is displayed through 
the power point, so that, the students with hearing impairment can read the text when the learning occurred. 
A classroom action research with tw0 cycle is used to recognize a summary of the instructional which 
attended by two students with hearing impairment and twenty-eight regular students. Observation, test, 
and interview conducted by team teaching show that speech to text based on power point has an impact in 
increasing comprehension of students with hearing impairment on inclusive education course.
Keywords: speech to text, hearing impairment, inclusive education

INTRODUCTION

The inauguration of inclusive education system 
that stated in Ministrial Regulation number 70 of 2009, 
people with disabilities have been accepted in the 
regular school and university. Irdamurni & Rahmiati 
(2015) states that inclusive education, as a education 
service that required all children with disabilities 
(visual impairement, hearing impairement, physical 
impairment, autism etc.) are served in schools nearby, 
in regular classes, together with regular students. 
Therefore, it should be emphasized to conduct the 
restructuring of schools and modification of science 
and technology in the learning process, in order to 
serve all students in appropriate learning process. 

Hearing impairment or deaf is one type of 
disability. Persons with hearing impairment are 
individuals who lost all or half of the hearing so they 
difficult to communicate verbally although it has been 
given hearing aids and still required special education 
services (Irdamurni, 2018). Deaf students can receive 
information when they read the lips of speaker. 
Conversely if deaf students do not pay attention to 
the speaker, such as the teacher/lecturer, in explaining 
the  material, they will not understand the material 
presented.

The development of information technology 
improves educational inovation in educating deaf 
students. Gales & Steve (2008) conducted research to 
apply algorithm Hidden Markov Models for Speech 
Recognition. This study reveals the development of 
Speech Recognition using algorithm Hidden Markov 
Model was greatly benefited from performance aspect 
and complexity because the Hidden Markov Model can 
display the rows of the vectors that needed by Speech 
Recognition (Fauzan, Arwani, & Fanani, 2018). The 
use of this Algorithm is also have high accuration of 
the Speech Recognition.

The author developed speech to-text applications. 
Speech to-text applications can be used to change 
the voice into text that can be displayed to the power 
point. The sound form of the input can be converted 
to text that can be read by the system. It is useful for 
deaf students because they can read the materials that 
delivered by teachers/ lecturers. Thus with the speech 
to-text app, deaf students with diverse characteristics in 
the classroom canunderstand the material that presented 
by teacher/lecturer.

The excess of speech to-text app is this app can 
affected human weakness, by the sound as the control. 
Through saying a word we can immediately get directly 
response  from the computer (Hepril, 2017). 



39Irdamurni, Johandri Taufan, Implementation of Speech to-Text Application for Deaf Students on . . . .

Tabel 1. The mean score of pretest

No  Subject Mean 
score

Descrip-
tion

1 Students with hearing 
impairment

  58 Poor

2 Students with visual im-
pairment

  70 Fair

3 Regular students   78 Good

Table 2. Mean score of post test

No Subject Pre-
test

Post-
test

Descrip-
tion

1 Students with 
hearing impairment

65 80 Good 

2 Students with 
visual impairment

75 80 Good

3 Regular students 81 85 Very 
Good 

Figure 1. Mean score of pretest

Speech to text/ Speech Recognition, can be found 
on windows Xp, windows vista and windows 7. On 
windows 7, speech recognition software is one of 
the items that are already on the menu control panel. 
Speech recognition is effective because it is easier for 
us in doing activities related to computer (Khan, 2019). 
We don’t need to type on the keyboard, and don’t need 
to move the mouse to enter the input. Because only 
through the sound and spelling of clearly Bahasa, the 
system can easily run to translate into text.

Through this way, we can determine when we 
need to type or edit the text immediately, and continue 
dictating to make the more appropriate sentences.  
Speech to text can be used to change the voice into 
text, so that it can be easier for users when they want 
to translate a language. The purpose of this article is to 
determine whether the speech to-text application can 
improve students’ understanding of deaf students in 
learning material.

METHOD 

This study used classroom action research. This 
research is conducted in the second semester of the 
2018 academic year. The sample of this research is 30 

students consists of a student with visual impairement, 
one deaf student and 28 regular students who take a 
inclusive education class in the department of special 
education, Education Faculty of UNP

Data collection techniques in this study using 
multi techniques. The learning activities conducted in 
the classroom on the subjects of inclusive education, 
and collaborated with team in inclusive education 
course. Data of the deaf students performances in 
inclusive education course achieved through pre-
test and posttest. Pre test conducted at the beginning 
of the 1st cycle and the posttest was conducted after 
cycle 1 and cycle 2. This is conducted to determine the 
improvement of deaf students’ performance in inclusive 
education course. 

The data collected in classroom action research are 
quantitative data and qualitative data. The quantitative 
data is data numbers obtained from the pre-test 
and post-test. Quantitative data was analyzed uses 
descriptive analysis and visual presentation in tables 
and graphs. Presentation of data in tables and charts 
describes that the actions can causes an improvement. 
While qualitative data was qualitatively.

FINDINGS AND DISCUSSION

The implementation of speech to-text applications 
on inclusive education course conducted after mid 
semester, and before the mid semester, the learning 
process was not using the speech to-text app. The pre-
test conducted to measure the performances of students 
on inclusive education course. The mean score of pre-
test before using the speech to-text app are presented in 
table 1. For more details can be seen in Figure 1.

Based on Figure 1, it can be conccluded that the 
pre-test score of students with hearing impairment is 
poor, meaning that students with hearing impairment 
do not understand the material presented by the 
lecturer. After pre test score obtained, then the learning 
process continued with a cycle two, with implemented 
speech to-text app in the learning process. The material 
provided in according to the planning program of 
inclusive education courses. Cycle one is conducted in 
three meetings.

Besides the implementation of the test to the 
participants, addition data was also collected through 
interviews to students about the implementation of 
speech to text app. Opinion from interviews state that 
it need to increase the font size so that students is easy 
to read. In the first cycle, speech to-text app used 12 
point Time New Roman font, but in cycle two, it used 
16 point Times New Roman font. In the second cycle 
was conducted three meetings with the mean score are 
listed in table 2. For more details can be seen in Figure 
2.



40 Journal of ICSAR; Volume 3, Number 2, July 2019: 38-40

Figure 2. Mean score of post test

Speech to-text application is used to make deaf 
students focus to the material that delivered by the 
lecturer. They can understand  the material through the 
text in power point. What material that delivered by the 
lecturer will be recorded through the screen of the power 
point. This is in according to the principles of learning 
for hearing impairment. Irdamurni (2018) stated that 
principle of directional orientation, the teacher should 
stand in front so that the face of the teacher, especially 
the teacher’s mouth can be seen by the deaf students. It 
will help deaf students understand what is explained by 
the teacher. It needed to have face to face instructional 
with deaf students in the learning process so that the 
material can be understand by the students. This is in 
line with the opinion of the Tarmansyah (2009) that 
deaf students can receive information from what they 
saw.Through the speech to text app, the principles of 
instructional for deaf can be resolved. It means that 
the deaf students should not be always looking at the 
lecturer when explaining the learning material. 

Tantowi (2016) explains about how to use  Speech 
to text, here are the steps to activate it: 1. Click start , 
in the start button bottom left-hand corner, 2. Control 
Panel, 3. Ease of Access, 4. Speech Recognition . next 
connect the computer with the infocus along with the 
power point that was prepared for the regular students 
according to the material,

If teacher ever try this out, and about to set 
up the microphone, then just click the Start Speech 
Recognition. However, if this is the first time teacher 
try this software, then teacher should read to the steps 
how to upgrade by click on Start Speech Tutorial. Then 
if teacher have not set up a microphone  yet, teacher 
can click Set Up Microphone. Select next, and continue 
to finish. Speech to text is already to used. Open at the 
windows display then there will be Speech Recognition 
box display.

On the box Speech Recognition, it written 
Sleeping which means software is not active. To 
activate it, teacher just need to Click on the button 
under a microphone picture and say “START”. Then, 
the box will be inscribed the words “Listening” which 
means the software is ready to used. To continue, say: 

“I am now using speech recognition to dictate to the 
computer”.After that, it is free to open any application. 
Speech Recognition will only work if we say it clearly. 
Unclearly spelling will only burden its work. Usually, 
when the intonation of pronunciation is not clear, the 
software remains silent. Therefore, to avoid this, the 
spelling should slowly and clearly.

CONCLUSION

Based on the result and discussion, it can be 
concluded that : The implementation of speech to-
text application in learning process can improve deaf 
students’ performances. Inclusive  education course 
by using the speech to-text app can improve learning 
motivation of the deaf students . The speech to-text app 
can be used by connecting with the internet network. 
Because of the process uses google voice as a library 
for the matching process of the pronunciation of sounds 
and also as a validation. The text displayed of speech 
to-text application must be using  16 point font size so 
that the deaf students can read clearly. 

REFERENCES

Fauzan, A., Arwani, I., & Fanani, L. (2018). 
Pembangunan Aplikasi Iqro ’ Berbasis Android 
Menggunakan Google Speech.

Gales, M., & Steve, Y. (2008). The Application of 
Hidden Markov Models in Speech Recognition. 
Foundations and Trends® in Signal Processing, 
1(3), 195–304.

Irdamurni. (2018). Memahami Anak Berkebutuhan 
Khusus. Jawa Barat: Goresan Pena.

Irdamurni, & Rahmiati. (2015). Pendidikan Inklusif. 
Sebagai Solusi dalam mendidik anak istimewa. 
Jakarta: Paedea.

Khan, B. (2019). Android speech to text tutorial. 
February 5, 2019. Retrieved from https://www.

simplifiedcoding.net/android-speech-to-text-
tutorial.

Hepril, A. S. (2017). Implementasi Speech Recognition 
Dalam Aplikasi Game Kuis Edukasi Untuk Anak 
Usia Dini Berbasis Android (Doctoral dissertation, 
UIN Sunan Gunung Djati Bandung).

Tantowi, F. (2016). Aplikasi Pencarian Surah Pada 
Audio Player Untuk Difabel Netra Dalam 
Menghafal Al-Qur’An Dengan Metode Speech 
Recognition Berbasis Android.

Tarmansyah. (2009). Pendidikan Inklusif. Padang: 
UNP Press.