Forum Discussion

Altera_Forum's avatar
Altera_Forum
Icon for Honored Contributor rankHonored Contributor
10 years ago

I have an issue in audio filtering and i need help please

Hello there

I'm a beginner in using DE2 board and I need some help.

I have this code which input and output audio through line-in/line-out on altera DE2 board (cyclone II, Wolfson 8731). I have been asked to implement a highpass FIR filter with 1 kHz cut-off frequency, and 48 kHz sampling frequency. also I have to calculate the distance between samples of the filtered signal. when I tested the filter using sine sweep (20 Hz- 20kHz), I can't hear anything before 1 kHz which means my filter works!. However when I connect the output to the oscilloscoop, the filtered signal goes to zero ( i can't see the filtered signal). my questions are:

1. Is this the right way to implement a FIR filter on the board ? if not, what should I do for the filter to work properly?

2. im still confused about where to implement the filter, should it be in the wolfson chip or in the FPGA itself?

the project files are attached

any extra information or suggestions is much appreciated

thank you,

4 Replies

  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    10 years ago

    Do you have a testbench and a set of test vectors? how did you design the filter initially, can you compare the results to the origional algorithm in simulation (it will be far easier to debug than on hardware).

  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    10 years ago

    hi Tricky, thanks for responding

    I designed it using " fdatool" MATLAB, It also generate a testbench for the filter, but im not sure about test vectors.

    so there is nothing wrong in the connection of the filter, but the problem might be in the filter itself. do you mean that ?
  • Altera_Forum's avatar
    Altera_Forum
    Icon for Honored Contributor rankHonored Contributor
    10 years ago

    I have no idea about the filter - I dont know the algorithm and Im not really a filter design guy.

    But the basic pricipal of having a testbench with good testvectors you can use in both your model and your design is essential. Without them you're basically doomed to a life of painful hardware debugging.

    So put the same data through your model and your firmware design - do the results match?