Skip to main content

Applications using DSP - Audio Watermarking


Patent Review: Wavelet Domain Watermarks
Publication number: US6385329 B1
Date of Patent: May 7, 2002
Inventors: Sharma et al.

Summary:
The invention relates to encoding and decoding auxiliary signals in a media signal, such as an audio, video or image, using wavelet decomposition of the signal. This method performs the wavelet decomposition of the signal and embeds a watermark into the wavelet decomposition. Another aspect of the invention is a method of detecting an auxiliary signal embedded in a media signal, where the auxiliary information is substantially imperceptible in an output form of the media signal. Another aspect of the invention is the alternative method of encoding of the auxiliary signal in a media signal. This method performs a wavelet decomposition of the media into two or more levels of resolution.
.
Link:  http://www.freepatentsonline.com/6385329.pdf


IEEE Paper Review: A robust wavelet-based digital watermarking using level-adaptive thresholding
Authors:  Kim et al.
Publisher: IEEE
Date of Conference: 24-28 Oct. 1999

Summary:
In this paper, a robust digital watermarking algorithm using wavelet transform is proposed. In order to equally embed the watermark to the whole image, the coefficients of all sub-bands including LL sub-band are utilized. The proposed watermarking method improves the performance in terms of invisibility and robustness, .by using level-adaptive thresholding to select perceptually significant coefficients and by using different scale factors according to the level of decomposition. The proposed method has been verified to be superior to other existing transform-domain watermarking schemes, especially for highly compressed images.

Link: https://ieeexplore.ieee.org/document/822889/

The lab started with Talele Sir explaining us the process of patent and paper searching and what content to focus on while reading a patent or a paper. Next we had to decide on a topic of our own choice and so we, in groups of 5, searched through numerous articles which listed and explained DSP applications. One particular lab material of Columbia university caught our eye, which had several DSP applications listed on it. We found Audio Watermarking as a very interesting application (since it involves cryptography, mathematics, hacking  yay!) and decided that we will try and implement it.

Once the topic got approved we searched for patents related to this on freepatentsonline.com and were quickly overwhelmed by the number of patents available. So we decided that we would focus on finding a working program first and then find patents and papers related to it. This strategy worked and thanks to @99sbr on GitHub, we found a well developed working program. Later we found related patents, papers; understood the program flow and combined our observations and learning into a review paper of our own. Overall it can be summarized as a lengthy and cumbersome experiment involving many visits to Talele Sir's cabin.

Comments


  1. This is a really interesting topic to be discussed. I didn't know that such complex methods were used for audio protection.

    ReplyDelete
  2. How is this compared to watermarks that we see on images?

    ReplyDelete
  3. Great concept, and its very intriguing! I would surely read more about this.

    ReplyDelete
  4. There is also something called as Audio Fingerprinting which generates unique fingerprints from the content itself

    ReplyDelete
  5. You need to clarify more on the technique used.

    ReplyDelete

Post a Comment

Popular posts from this blog

FIR Filter Design

This was a relatively easy assignment as the MATLAB function fir1() directly generates the required filter based on the given specifications. The only task was to decide the appropriate window function based on the given stop band attenuation, as taught in the lecture.

Fast Fourier Transform (DITFFT)

Using structures for performing operations on complex numbers made programming this experiment relatively simpler. I have used recursion in the FFT function which makes it functional for any radix-2 value of N, unlike the example codes provided, which had separate functions for different values of N. Because of this, the input signal of any length has to be first converted into a radix-2 form by appending the necessary amount of 0's and then calculating FFT. IFFT simply uses FFT for its calculation.

Operations on DSP Processor

This experiment refreshed the concepts learnt during the Embedded systems course that I had attended during my first year. I had to demonstrate the practical to my batchmates and we executed assembly instructions on the DSP Processor kit available at the laboratory.