DESIGN OF A REAL-TIME SPEECH SCRAMBLER BASED ON PERMUTATIONS OF FREQUENCY COMPONENTS AND ITS IMPLEMENTATION FOR AUDIO COMMUNICATION SECURITY ON AN FPGA

• Main Author: Pramudita Rumanti, Emmanuella
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/87731
• Institution: Institut Teknologi Bandung

Some people worry about the potential eavesdropping of the communication devices they use, as well as the risk of misuse of personal information. These worries may be aimed towards the communication device or network whose methods of handling data cannot be known by users for certain, such as instant messaging (IM) applications provided by third parties. Additionally, these worries may also be related to the potential unknown existence of spyware or malware on the private devices of the user or the conversational partner, whose nonexistence is very hard to prove. One of the methods that can be used to implement end-to-end encryption (E2EE) in this situation is by doing the encryption on the message yourself before the message enters the communication device, and inversely doing the decryption yourself after the encrypted message has left the communication device. In the case of secure audio communication, this may take the form of encrypting speech after it enters the microphone but prior to it being received by the cell phone or computer, and decrypting speech that has been received from the cell phone or computer but before it is sent to the headphones or speaker. To implement this, we designed and developed an audio communication security device that scrambles speech by changing the permutations of its frequency components. The audio communication security device has mass and dimensions that enables it to be placed on a small table with ease, can be connected with a private communication device such as a cell phone, laptop, headphones, or speakers using a 3.5mm audio jack, and can process sound and scramble it very fast in real-time. This capstone project discusses the development of the design and implementation of an FPGA-based secure audio communication device. The designed speech scrambling system implements Fast Fourier Transform, the permutation of frequency components X[k] in the frequency domain, and Inverse Fast Fourier Transform to produce scrambled speech in the time domain. The results of this development has been successful in implementing real-time encryption and decryption of speech in real-time on an Altera Cyclone V SoC 5CSXFC6D6F31C6 FPGA.