Robust Sound Encryption

Introduction

            By the development in technology, people have started to be interested in multimedia applications.  Sound is one of the main components of these applications.  With the increase in importance of sound, security problems are attached more importance. Researches have guaranteed sound security in digital environment, but there are still problems about security of sound entering to analog environment. In this research, a sound encryption algorithm, that is robust to filtering and quantization processes, has been developed.

Robust Sound Encryption

            The process of sound transmission is done by taking samples in definite periods. Thus obtaining an equal copy of the sound signal is impossible.  Depending on digital environment is needed for coding sound samples one by one or as blocks but it is not possible to get sound back as intelligible after passing to analog environment and modification is needed.   

To solve this problem a robust sound encryption algorithm whose main mixing scheme is based on noble prime numbers has been developed. For every noble prime number a time depended block has formed. Every block contains bunches of samples returned in milliseconds.

            Noble prime numbers are special states of prime numbers. For every noble prime number there is a condition that sum of remainders divided by one is equal to noble prime number minus one. There is a sample application for a noble prime number, “19”, below: 

                 Sample format: 44100 Hz 16-bit

                 Noble prime number =19    

                 if one block is ms = 50 milliseconds then it contains samples.

                   Let’s form a mixing array starting with 1:

                           1 mod 19 = {1,10,5,12,6,3,11,15,17,18,9,14,7,13,16,8,4,2,1}

                   Used mixing array: {1,10,5,12,6,3,11,15,17,18,9,14,7,13,16,8,4,2,19}

                   Mixed array: {A, J, E, L, F, C, K, P, R, S, I, N, G, M, Q, H, D, B, T}

            The start of mixing process may be done by using whichever numbers between 1 and 19. Thus, a more robust coding can be formed.

Experimental Results

            A sample CD quality wave file has been encrypted and decrypted. The file had the same quality with original one.

            High-Pass, Mid-Pass and Low-Pass FFT filters had been performed on the sample file. It’s found that 80% of filtered sample is perceivable and can be audible. It was found that there is some repeated noise on particular frequency after decryption. This noise occurs because of the transition between the blocks during filtering process.

            Quantization test was performed by reducing a sample file from 44100 Hz 16-bit stereo to 22050 Hz 8-bit mono and bringing back to original format. Doing so sample file loosed a significant amount of quality, especially high frequencies has gone In spite of this loss, after decryption, sample sound was still audible. There was still a constant noise.

Future Studies

            Noise scatters on certain time intervals between two blocks. This noise is in particular frequency bands. A special FFT filter can be applied to erase the noise.

            In addition to this by examining the occurrence of the noise more accurate solutions can be found.

References

[1]    M. Yahya Karsligil, Noble Prime Numbers, Lecture on Data Structures and Algorithms, 2001.
 

[2]    Chasan Chouse, Audio Encryption Using Noble Prime Numbers, Lecture on Data Structures and Algorithms, Research no 12, 27-05-2001

Visual Studio 7 Source Code

     Source code is available. Grab it from here.
     Also Compiled version is here.
 

Sample Sound Files

     Files are in mp3 format. You must convert original mp3s to PCM wave format(44 KHz 16-bit stereo) in order to use.