Wave steganography - from military to intelligence service applications (by Krivous Anatoly Anatolevich)

Download from ME

Today i want to talk about the cryptography. I've made the example of using the special cryptography - the steganography. This method hides the fact of encryption of the data. There are lot of kinds of the steganography. Today i'll talk about LSB-method when data is hided into the least significant bits of the audio file. It looks as though you are exchanged a audio files, but really you send a secret data. People who don't know about this method they will not even suspect about secret data. In some cases it can be very useful.

How does it work?

A WAVE-PCM file (without a compressions) contains sound data. Really the sound is an analog event, i.e. continuous. In order to convert it to digital form you should quantize it with lossy. This process is characterized by two parameters: bitness and sample per second. "Bitness" affects to how many levels can it contains in each sample. "Sample per second" affects to how many frequencies do you can hear:

In this case we are interested only the bitness of an audio. It can be 32, 24, 16, .... bits per each sample. Main idea of steganography (in this case) is rewrite the least significant bits to our data. The more you overwrite bits the greater the distortion.

This picture explains it graphically:

As you can see, it stores all hidden data to certain bits in the audio data (in this picture 4 bits to each sample). Also note that for storing the data you need to use the bigger file size than the source file. For instance, if you use 3 bit for the decoding the result file will have the size that is 16/3 times greater than source. I've said 16 because i use the 16 bps wave file in my example.

In the attached example i also save the original file name. In general, format of the data is described in the picture:

When the packing occurs it gets each byte from the source file. Then the subroutine extracts the necessary bits from the source file and clears corresponding bits in the audio data. Further the subroutine sets bits using bitwise-OR operator. For extracting the necessary bits it uses the masks and the shifts. The mask leaves necessary bits and the shift places them to the beginning of the byte.

Unpacking works vice versa. It extracts bits from audio data and builds file using corresponding bits.

Hope the review will be useful.

Thanks for attention.

Regards,

Кривоус Анатолий.