Implementation of Google Translate Application Programming Interface (API) as a Text and Audio Translator

. Through this translator program, it is craved that it can avail the general public to understand foreign language videos, can be useful in the world of education and technology, and can avail the persons with disabilities be up to communicate. The method used is a classification method that functions to detect the flow of shapes to instruct the class attribute as the task of the input attribute by generating automatic output through three stages, namely Machine Learning, Natural Language Processing, and Speech. The results showed that 90.38% of videos were successfully translated into text and audio, 9.62% of videos failed to be translated because the owner limited public interaction, and 89%-97% synchronization between text and audio. In this research, a text and audio translator program has been created using the Application Programming Interface (API). This program is a configuration of deep learning, machine translation, and text-to-speech designed using the high-level programming language python. The system used is a predictive system in which the system tries to predict the output equally the wishes of the user.


INTRODUCTION
Every person needs a medium in interacting. The medium is language. Language is a means of individuals interacting in the form of spoken sound symbols by the medium of human speech [1]. In understanding language, an object study is needed, namely linguistics. The origin of the word linguistics is adopted from the lingua which means language. Linguistics is the diagnosis of human language with science [2].
The character of a nation can be seen from the use of language. As a multilingual nation, Indonesia has hundreds of regional languages. To facilitate people interaction, one language is set as a unifier of the nation. This event has been arranged in the state constitution, the 1945 Constitution in Article 36, in which Indonesian is the official language of the state [3].
Recent research shows, that worldwide there are 7,000 languages and according to Ethnology statistics, 2.7 billion people communicate with eight languages and about 70 million humans interact with 5,727 languages [4]. With so many foreign languages as national languages, various problems will arise, either in the form of economization or efficiency. As is the case with the Philippines, although the Philippines has a national anthem (Lupang Himirang) in terms of language, the Philippines is unable to establish Tagalog as a unifying language [5].
With the globalization era, there have been changes in the times, including the linguistic aspects in it. To answer this question, English is designated as a global language with more than one billion people speaking it. The global language can broaden horizons in the field of science and technology. The dissemination of up-to-date information can help every individual around the world in getting information [6].
English has been established as a universal language with a limited understanding of a language meaning, it becomes a big enough problem for someone to communicate. A simple example is when watching YouTube videos that use foreign languages, of course, it is very difficult to understand these videos, DOI: 10.24014/coreit.v8i1.14695 especially since not all YouTube videos provide subtitles and dubbing in them. Therefore, a translator program is implemented by utilizing the Application Programming Interface (API) which is a management in the form of interfaces, tasks, levels, to forms to design software or computer software. Application Programming Interface (API) can help programmers "manifest" software that is integrated with other devices. Application Programming Interface (API) is a connector between applications that allows programmers to use disposition functions that are managed through a computer program operation [7].
This research refers to the lack of development in linguistics through Artificial Intelligence by utilizing computer programs. Then the Google Translate Application Programming Interface (API) was implemented as a text and audio translator with the main goal of making this translator program, which is expected to help the general public understand foreign language videos so that they add insight, can be useful in the world of education and technology is knowing how Artificial Intelligence works. Intelligence and Programming, and can help people with disabilities to be able to communicate audio and visually.

METHODS
The system used is a predictive system where the system tries to predict the output in accordance with the wishes of the user by utilizing the available libraries. An Application Programming Interface (API) is the design of a task that other programs can invite. The Application Programming Interface (API) functions as a link that unifies various applications from each version of the manifesto [8]. The Application Programming Interface (API) can translate foreign languages, such as English to Indonesian into script text and can project appropriate audio translations.
This program uses several libraries, namely Google API, Google Translate, and Google TTS. TTS or textto-speech is an implementation with the aim of converting from text to audio form where the resulting conversion can be adjusted for prosody speed [9]. The Application Programming Interface (API) forwards an invitation to the system in the form of translation evidence and the system displays the appropriate translation response [10].
The method used is a classification method in which each state stores various attributes, such as class attributes. This method works by detecting a stream of shapes to instruct class attributes as an assignment from the input attribute by generating the appropriate output via a language video link from YouTube. To find some class attributes, you can use a simpler programming language. Through python, the input will be more concise with the output remaining the same. Python programming language is a language that accommodates level automatic programming that can be run in various programs by managing programs quickly with simple encryption, supporting various platforms, and having a memory management disposition [11]. To understand the creation of an Application Programming Interface (API) translator program, a program creation scheme is shown. On the Jupiter notebook working page in the python application, a video link is inserted into the program, which is connected to the internet network because the created program must be connected to a stable internet connection. Make sure beforehand that you have downloaded the libraries that will be used. Libraries used are Google API, Google Translator, and Google Text-to-Speech.
The first stage is a deep learning process that is included in the Machine Learning (ML) component. Deep Learning is a subset of Artificial Intelligence (AI) that focuses on big neural network modeling, which is capable of making accurate data-driven decisions. In digital era, many high-end consumers and online technology companies use deep learning, such as Facebook which uses it to processing text in online conversations and Google, Baidu, Microsoft, etc. use it for search image and translator [12]. The system will convert the video into a script by utilizing the libraries from the Application Programming Interface (API).
The second stage is the machine translation process which is one of the constituents of Natural Language Processing (NLP). Machines will interact by utilizing natural language. Natural language is the language used in general communication that includes all mother tongues from all over the world. Natural Language Processing (NLP) is a very onerous duty in computer science where it presents an extensive variety of problems that vary from language to language. Foregoing, computer scientists broke down language into grammatical forms to become parts of speech, phrases, and others using complex algorithms [13]. Natural Language Processing (NLP) contains many algorithms, tasks, and problems in the form of input text and produces an output of useful information, such as labels, semantic representations, and so on. Other tasks in the form of translating, summarizing, and making texts are included in the form of Natural Language Processing (NLP) [14]. After the video script is obtained in the first stage, it needs to be converted into the desired translation script by making use of the Google Translate library. Google Translate is useful to answer users' needs for foreign language translations from various languages in the world.
The results of the translated script need to be converted into audio form first by using the Google Text-to-Speech library. This stage is the last stage of the Application Programming Interface (API) translator program, namely Speech. The machine will work according to the commands entered in the form of textto-speech or speech-to-text and the output from the system will not only appear on the monitor but will be stored automatically on the user's device.

RESULT AND DISCUSSION
The program is tested by translating various languages on the video with output in the form of text and audio with a duration of 3 minutes to 12 hours. This program has been tested to translate various videos on YouTube, from languages in Europe, America, to Asia with long durations ranging from 3 minutes to 12 hours. A total of 52 random video samples from various languages, durations, and content were entered into the program and the results showed that 90.38% of videos were successfully translated into both text and voice, and 9.62% of videos failed to translate because the owner limited interaction. public. Tests carried out on the Application Programming Interface (API) translator program showed that this program has no limitations, only a few conditions are needed for its use, namely the video link must be sourced from YouTube and make sure the internet connection is stable because the program is made online and automatically. In the test, there were several videos that could not be detected in a high-level programming language (python) because the videos had a personal license (license) or copy right (copyright) that could not be published, and high-level machines could not detect objects that which requires further development.  shows that in the video synchronization test, the program is able to synchronize text and video. The accuracy between text and voice has been tested and the resulting synchronization value is 89%-97%. To see the results of video samples tested through the Application Programming Interface (API) translator program, you can visit the provided link https://www.youtube.com/watch?v=zasbbNMRh8I.

CONCLUSION
The idea of making a foreign language translator program by utilizing the Application Programming Interface (API) is a new idea as well as an innovation in information techniques due to the limitations of the YouTube platform and video makers providing translations in the form of separate text and audio. In addition, there has not been much use of the Application Programming Interface (API) in linguistics, but many developments have been carried out in other ways using Google Maps, Google Assistant to Google Calendar. With the lack of depth in the field of linguistics, it becomes a challenge to develop linguisticbased programs or software that can help the general public, especially those who have limitations.
The test results show that the program is able to translate the video into text and audio with 90.38% presentation and 89%-97% synchronization with the condition that it must be connected to a stable internet connection. This research has answered the underlying problem. Lack of understanding of a language and other limitations will not be a barrier to obtaining useful information and knowledge. To help you get started, apply the Application Programming Interface (API) to any program that comes with the best software available on the Web.