s~\(f"t"'I -~

S URAJ R

COIMBATORE-641105.
EMAIL ID: shan_suraj@rediffmail.com

Abstract
Steganography is defined as the art of passing infonnation in a manner that the very existence of the message is unknown. Digital steganography is based on the fact that artifacts like bitmaps and audio files contain redundant information. That's why lossy compression techniques JPEG work. A lot of work has been done in hiding text within images. We concentrate on ISO-MPEG AudioLayer-3 (or MP3) to hide data within audio files. The widespread popularity of the MP3 file fonnat and its ubiquitous presence on storage m.edia and the Internet makes it a useful medium in which to transfer data. The. wav fonnat is compressed into the mp3 format. The Perceptual audio coder is incorporated for the purpose. In a perceptual audio coder, the codec does not attempt to retain the input signal exactly after encoding and decoding, rather its goal is to ensure that the output signal sounds the same to a human listener. It takes advantage of this process to insert hidden data during data compression. The data is hidden in the low frequency components of the audio wave. The hidden text is embedded by MP3Stego using SHA-l to generate pseudo random bits, which prevents repetitive patterns which will be apparent to the steganalyst. MP3Stego employs a passphrase and 3DES to protect the hidden data payload. The entire system has been simulated in C.

Introduction
Steganography is the art of hiding communications without the unintended party's knowledge. Steganography has again come into focus by the cryvto analytical community and the media because of the ease that infomlation can be hidden in today's many forms of digital communication. Steganography techniques can be applied in many ways including images, invisible inks, microdot printing, spread spectrum, burst communication, messages, viruses inside of a seemingly benign communication and now the Internet. The concept of hiding text within images has been under research for many years. It is possible to send a hidden data payload through an audio file. In this paper we hide data in a mp3 file and transfer it. One should not underestimate the difficulty of hiding data in audio signals. The human auditory system (HAS) operates over a wide dynamic range. The HAS perceives . over a range of power greater than one billion to one and a range of frequencies greater than one thousand to one." HAS is fairly poor with respect to its' differential range. This means that loud sounds tend to mask out softer sounds easily. Psychoacoustic models (as described later in the paper) can exploit this weakness.

Objectives and Design methodology

The primary objective of the paper is to hide the data payload within an audio file. The mp3 format has been incOl-porated for this purpose. The mp3 is a useful folmat sending covert data because ofits abundance and accessibility. An efficient technique for hiding the data within the mp3 sound file has been used. The design methodology involves the compression of a wav file into a mp3 file. The data is hidden during the compression process. This technique is a less timeconsuming and highly efficient. Random bits are introduced into the file using the SHA-l technique. This prevents the detection of data by a Steganalyst. The 3DES technique introduces a passphrase and protects the hidden data payload.

Conversion of .wav to ISO-MPEG AudioLayer-3

In a perceptual audio coder, the codec does not attempt to retain the input signal exactly after encoding and decoding, rather its goal is to ensure that the output signal sounds the same to a human listener. The primary psychoacoustic effect that the perceptual audio coder uses is called 'auditory masking', where parts of a signal are not audible due to the function of the human auditory

system. The psychoacoustic model analyses the input signals within consecutive time blocks and detennines for each block the spectral components of the input audio signal by applying a frequency transform., Then it models the masking properties of the human auditory system, and estimates the just noticeable noise-level, sometimes called the threshold of masking. In its quantization and coding stage, the encoder tries to allocate the available number of data bits in a way that meets both the bitrate and masking requirements. Two nested iteration loops are utilized in the Layer-3 encoder; the inner iteration loop (rate loop) and the outer iteration loop (noise control/distortion loop). Without going into a long, detailed explanation of the encoding process, these two loeps are involved in a "negotiation" for digital real estate. The encoding process is illustrated below:

MPEG Audio Layer-3 Encoding Process

f.ilter bank I ~ .-.I I f)ATA I perceptual ~ model M . Join stereo U .! coding ~ ~ I roNTR m . I P s L MP3 coded E Audio stream Scale factors X R Rate/Distol1ion control process .

Passphrasing using the JDES technique:

DES is a block cipher -i.e. it acts on a fixed-length block of plaintext and converts it into a block of ciphertext of the same size by using the secret key. In DES, the block size for plaintext is 64 bits. The length of the key is also 64 bits but 8 bits are used for parity. Hence the effective key length is only 56 bits. In 3DES, we apply 3 stages of DES with a seperate key for each stage. So the key length in 3DES is 168 bits. DecI")'ption is done by applying the reverse transformation to the block of ciphertext using the same key. Since the same key is used both in enclyption and decryption, DES is a symmetric key cipher. This method differs from algorithms like the RSA encryption which use different keys to enclypt and decrypt a message. The DES method of operation is shown below:

M P3Stego :
MP3Stego will hide information in MP3 files during the compression process. The data is first compressed, encrypted and then hidden in the MP3 bit stream. The hiding process takes place at the heart of the Layer III encoding process namely in the inner_loop. The inner loop quantizes the input data and increases the quantizer step size until the quantized data can be coded with the available number ofbits. Another loop checks that the distol1ions introduced by the quantization do not exceed the threshold defined by the psychoacoustic model. The part2_3_length variable contains the number of main data bits used for scalefactors and Huffman code data in the MP3 bit stream. We encode the bits as its parity by changing the end loop condition of the inner loop. Only randomly chosen part2_3_length values are modified; the selection is done using a pseudo random bit generator based on SHA-l.

We use our one-time pad (or keystream generator) to select not one pixel but a set of them, and embed the ciphertext bit as their parity. This way, the information can be hidden by changing whichever of the pixels can be changed least obtrusively. There is an interesting tradeoff: the more bits in the selection channel (i.e., the greatcr the value of k), the more bits wc can hide in the --covertext. In practice our selector channel will be a cryptographic psuedorandom number generator, and we can draw from it as many bits as we like.

Secure Hash Algorithm 1 :

Thf, Secure Hash Algorithm (SHA), developed by NIST, along with the NSA, for use with the Digital Signature Standard (DSS) is specified within the Secure Ha5h Standard (SHS) SHA is a cl-yptographic message digest algorithm similar to the ~1D4 family of hash functions developed by Rivest. It differs in that it adds an, additional expansion operation, an extra round and the whole transformation was designed to accomodate the DSS block size for efficiency. The Secure Hash Algorithm takes a message of less than 264 bits in length and produces a 1 60-bit message digest which is designed so that it should be computationaly expensi\re to find a text which matches a given hash. ie if you have a hash for document A, H(A), it is difficult to find a document B which has the same hash, and even more difficult to arrange that document B says what you want it to say. .

The Basic DES Algorithm:

The basic DES algorithm encrypts a plaintext block by a process that has 16 rounds as seen below: 8,.-1 11(, lJII IJr--l f't'/:l"/t'/ (Jph:'/: Encryption Process: .The block of plaintext is split into two halves {Lo,Ro). each of which is 32 bits long. Also DES uses the origina156 bit key to generate 16 keys of48 bits each {kj), These sub keys are used in the 16 rounds. .In each round, the function F is applied to one half using a sub key kj and the result . is XORed with the other half. The two halves are then swapped and the process is repcated. All the rounds follow the samc pattern except the last one, where there. IS no swap. .The final rcsult is the ciphel1ext {Lr,Rr). Hence the plaintext {Lo,Ro) is tranSfO\111ed to {Lr,Rr) ,

Decryption Process:

.Decryption is structurally identical to encryption. So the same machinery is used as above. However, the input here is the pair (Rr,Lr) instead of (Lo,Ro). .Also, the input sub keys are applied in the reverse order. So the ith sub key is kr-;+J instead of k;. .The final result is the original text (Lo,Ro). So the ciphertext (Rr,Lr) was decrypted to (Lo,Rt)). The 3DES Algorithm 3DES uses the same basic machinery is three times over using 3 keys kl,k2,k3 The plaintext (M) is encrypted using kl. This result is encrypted with k2 and the result is then

encrypted with k3 to get ciphertext (C). C = Ek3(Ek2(Ekl(M)). This mode of using 3DES is called the DES-EEE mode since all three keys run in the encryption mode. The other mode is called DES-EDE where the second stage is run in decryption mode. i.e. C = Ek..,(Dk2(Ekl (~1). The 3 keys mayor may not be independent. For DES-EDE mode three options are defined .The keys kl, k2, and k3 are indepcndent. .k( and k2 are independent but k( = k3 .k, = k2 = k3 -In this case, 3DES becomes backward compatible with DES.

MP3 STEGO IN ACTION :

The command line interface mode allows the user to both encode and decode files. To encode a file the following syntax is used: encode -E secret_text.txt -p passphrase hells_bells. wav hells_bells.mp3 This compresses hells.beHs. way along with secret_text.txt into hells_bells.mp3, encl")-pting the hidden text using the passphrase "passphrase." Decoding the file is accomplished using: Decode -x -p passphrase hells_bells.mp3 This uncompresses hells_bells.mp3 into hells_bells.mp3.pcm and decrypts and saves the hidden text as hells_bells.mp3.txt. ;~ c~ " SIMULA TION: Encryption & Compression: Encryption: [_WAY Fil~J 1r . Secret Stego Enclyption Ir .~ d.-nionnatlon ~ an compression MP3 File Software ~ I Stcgo key Decryption: . MP3 File Stego Declyption Secret Infonnation Software Stego Key The whole project \vas simulated in c language.

CONCLUSION :
In this paper the art of hiding a text data behind an mp3 file has been successfully pointed out. The technique has been simulated in C. The mp3 Stego using the 3DES and SHA-I technique is a simple but efficien(technique for the transfer of hidden data packets.