Audio Quality Enhancement Using Adaptive Filters

References

[1] METHODOLOGY

[2] LMS ALGORITHM

[3] The least-mean-square (LMS) algorithm is widely used in adaptive signal processing for its robustness and simplicity. It is known for its simplicity and its good steady-state performance in stationary context. Consider in general an N-tap filter, with the weight vector w(n) at time instant n denoted by,

[4] w(n) = [w1(n)w2(n) … wN(n)] T

[5] Let {x(n)} be the input sequence x(n) = [x(n) x(n-1) ……… x(n-N+1)] be its vector representation containing theimmediate past N samples of {x(n)}.

[6] The filter output 𝑦(𝑛) = 𝑊𝑇(𝑛)𝑥(𝑛)aims to follow a desired signal and the estimation error e(n) is definedby

[7] ⅇ(𝑛) = ⅆ(𝑛) − 𝑦(𝑛)

[8] An adaptive filtering algorithm adjusts the filter tap weight w(n) at each time in- stand according to the measured value of e(n). The standard LMS algorithm up- dates as

[9] 𝑤 (𝑛 + 1) = 𝑤(𝑛) + 𝜇ⅇ(𝑛)𝑥(𝑛)

[10] where mu(u) is defined as the step-size parameter which affects the convergence behaviour of the filter weights.

[11] NOISE REDUCTION ALGORITHM BASED ON LMS FILTER:

[12] We propose a noise cancelling scheme based on LMS filtering algorithm of its optimum performance. The block diagram of the noise reduction method is shown in figure. Most audio signals are time varying signals, in order to achieve effective noise reduction with LMS filtering method, the input signal must be segmented. The unprocessed noisy signal is segmented every 40ms. Let y= {y: t=1,2,…..T} be a noisy test signal with T frames and being the frame at time t

[13] In the proposed method, the problem of noise cancellation can be stated as identifying for each noisy frame y(t) a matching weight vector w(t). Since a whole unit with several frames, when treated as a segment of consecutive speech frame, can be identified more accurately from noise than the whole frames. In the LMS filter we employ the normalized least mean square (LMS) algorithm. The LMS is a version of the well-known LMS algorithm that normalizes the weight vector updates with respect to the squared norm of the regressor. This normalization makes the LMS algorithm less sensitive to variations in the input power of the adaptive filter. Therefore, the LMS algorithm is a good candidate for applications with a high degree of uncertainty about the filter input power. The statistical analysis of the LMS algorithm is complicated by the normalized weight update.

[14] The LMS algorithm can be reviewed as a special case of the LMS algorithm with a time-varying step size, in which the step size varies with the input signal strength. The tap-weight adaptation equation of the LMS algorithm is given by

[15] wn + 1=wn+μxT(n)x(n) ⅇ(n)x(n)

[16] Where 𝜇 is a parameter to be chosen and μxT(n)x(n) is the actual step size.

[17] The output signal of the proposed method is.

[18] RESULTS

[19] This normalization makes the NLMS algorithm less sensitive to changes in the input power of the adaptive filter. As a result, the NLM algorithm is suitable for applications with high uncertainty of the filter input power. Statistical analysis of the NLM algorithm is complicated by updating the fully qualified balance. Algorithm LMS can be seen in a special case in the chain frequency of the LMS algorithm. Here, the step size depends on the power of the input signal. LMS Algorithm Duret Adaptive Almine As a result, it is, leading to the obtained algorithm of the adaptive LMS filter, and gradually relaxes the noise and gradually reverses the unarmed engine noise. The noise is corrupted by the voice signal. The results show that clean speech signals can be obtained after applying the adaptive LMS filter.

[20] OUTPUT GRAPHS:

[21] OUTPUT SIGNALS

[22] NOISY SIGNAL (SIGNAL AFTER MIXING)

[23] RECOVERED SIGNAL

[24] FOURIER TRANSFORMS OF SIGNAL:

[25] MAGNITUDE RESPONSE OF ORIGINAL SIGNAL

[26] PHASE RESPONSE OF ORIGINAL SIGNAL

[27] MAGNITUDE RESPONSE OF NOISY SIGNAL

[28] PHASE RESPONSE OF NOISY SIGNAL

[29] MAGNITUDE RESPONSE OF RECOVERED SIGNAL PHASE RESPONSE OF RECOVERED SIGNAL

[30] DISCUSSIONS

[31] In this project, we have done the literature survey on audio enhancement using adaptive filter, studied different research paper on this topic, studied about the LMS algorithm the mathematical equation used in MATLAB software filters the noise signal and convert it into pure output without any noise.

[32] CONCLUSION

[33] In this report, we have presented a noise reduction method for audio and speech signals by applying adaptive linear filtering technique. The noise reduction problem has been formulated as a filtering problem which is efficiently solved by using the LMS method. In addition, the method pays attention to the nonstationary nature of some audio signal. Simulation results indicate that the proposed method can improve the performance the quality of noisy audio signal. Through computer simulations, we have demonstrated that the proposed method is quite effective in noise reduction, especially in the case of stationary white Gaussian noise

[34] REFERENCES

[35] International Journal of Modern Engineering Research (IJMER) Vol.2, Issue.3, May-June 2012 pp-792-795 ISSN: 2249-6645

[36] Adaptive Filter Theory by Simen Haykin: 3rd edition, Pearson Education Asia.LPE.

[37] B. Widow, "Adaptive noise canceling: principles and applications", Proceedings of the IEEE, vol. 63, pp. 1692- 1716, 1975.

[38] G. Goodwin, K. Sin, Adaptive Filtering, Prediction and Control, Englewood Cliffs, Prentice Hall, 1985.

[39] Jingdong, C., Jacob, B., Arden, Huang. (2007). On the optimal linear filtering techniques for noise reduction. Speech Communications, 49(2), 305-316.

[40] https://www.researchgate.net/publication/266648972_A_Noise_Reduction_Method_Based_on_LMS_Adaptiv e_Filter_of_Audio_Signals [Crossref]

[41] https://www.researchgate.net/publication/267774899_Implementation_of_the_LMS_Algorithm_for_Noise_Cancellation_on_Speech_Using_the_ ARM_LPC2378_Processor. [Crossref]

[42] https://en.wikipedia.org/wiki/Adaptive_filter

@article{1703621,
      author = {Amit Kukker, Yash Nigam, Tushar Sawle, Rajeet Kumar},
      title = {Audio Quality Enhancement Using Adaptive Filters},
      journal = {Iconic Research And Engineering Journals},
      year = {2022},
      volume = {6},
      number = {1},
      pages = {165-169},
      issn = {2456-8880},
      url = {https://www.irejournals.com/formatedpaper/1703621.pdf},
      abstract = {Audio quality enhancing plays a vital role in the field of speech recognition, communication, medical etc. The most widely used method is an optimal linear filtering that can reduce the noise level present in an audio signal and improve its signal to noise ratio (SNR). Here, we propose Adaptive LMS filtering method that can improve SNR and enhance audio quality in a very fruitful manner. The signal is filtered at once and the filter coefficients are computed adaptively in an exponential algorithm. The simulation results show a higher audio quality than the raw noise signal. Almost all practical signalling applications are difficult to implement. In this article, we propose a method to reduce noise in audio or speech signals using LMS adaptive filtering algorithms. The signal is filtered at once and the filter coefficients are computed adaptively in an exponential algorithm. The simulation results show a higher quality than the raw noise signal.},
      month = {July},
  }