audio steganography in tranform domain

Discussion in 'Programmer's Corner' started by lata, Mar 31, 2010.

  1. lata

    Thread Starter New Member

    Mar 31, 2010
    my project is based on data hiding in audio using discrete wavelet transform.the transform domain is new for me. i would like to request those who had already worked on it to guide me through the same. i have IEEE papers for reference but i do not get the starting point for the implementation.
    thanking u in advance
    looking forward for the reply
  2. Papabravo


    Feb 24, 2006
    Well the starting point is a sequence of amplitude values as a function of a discrete time index. We have essentially no information on what happens between the successive samples, but we know that any frequency components above the Nyquist are irrelevant.

    The purpose of Steganography is to provide a framework for obscuring the actual data by creating a new dataset from which the original data is either difficult or impossible to recover even by massive application of brute force methods.

    In any kind of transform method there is a mapping of a function with a particular domain and range to another function with a different domain and range. If we are lucky this transformation is 1 to 1 and onto. This is a fancy way of saying that the transform is reversible. It is the reversibility of the transform that allows recovery of the original data from its transformed representation.

    So I would start by acquiring a deep understanding of transforms and their properties. In addition to the wavelet you should be comfortable with those of Fourier, Laplace, and my personal favorite the z-transform.