diffbfweights coefficients usage for audio processing

Hi, how to use coefficients from diffbfweights function to process audio data?

Answer 1

[Nithin Kumar]

Hi Marko,

I understand that you are trying to use coefficients from "diffbfweights" function to process audio data.

Once you have obtained the coefficients from the "diffbfweights" function, you can use them to process audio data using an adaptive filter.

Kindly refer to the following steps to use the coefficients from "diffbfweights" to process audio data:

• Load the audio data into MATLAB using the "audioread" function.
• Create an adaptive filter using the "dsp.LMSFilter" function. Set the filter length and the step size.
• Set the initial weights using the coefficients obtained from the "diffbfweights" function.
• Process the audio data block by block. At each block, apply the adaptive filter to the block of audio data and update the filter weights based on the difference between the filtered output and the desired output.
• Write the processed audio data to a file using the "audiowrite" function.

I hope it helps.

[Marko Jankovic]

Dear Mr. Nithin Kumar,

thank you very much for your reply.

I will try to implement the given solution and report the result...

In addition to the LMSFilter function, which is an adaptive filter that I understood (I'm not an expert in the field of filter design), is it possible to apply the given coefficients to one of the FIR filters in Matlab?

With respect

Marko Janković

Answer 2

[Marko Jankovic]

Dear Nithin Kumar,

unfortunately, I have not been able to implement an application that gives me a valid result.

With the help of the obtained coefficients, we perform processing in the spectral domain as explained in the algorithm of Mr. Jakob Benesti and Jing Dong Chen:

where represet coefficients (complex conjugate of coefficients)

or the coefficients are in the time domain?

Thanks in advance!