I calculate my impulse response according to this principle. Calculating the inverse filter for the (exponential) sine sweep Method. With the obtained impulse response I calculate my reverberation time according to this principle: Calculation of Reverberation Time (RT60) from the Impulse Response. Now I would like to filter the impulse response to calculate the reverberation times of certain frequencies. For this I have filtered the IR with this filter type (Biquad):
I used the filtered IR to determine the specific reverberation time. I did this with filters of different widths. The following graph shows the calculated reverberation times for the respective frequency with different bandwidths.
A bandwidth (BW) of 20 Hz provides a readable result, but I cannot detect the modes with it. However, if I go down with the bandwidth, I still don't get the peaks I expect. Here is the reference measurement with REW.
I think I'm getting lost by now, hence the questions:
- Am I using a useful filter?
- Does it make sense to filter so low frequency?
- Are there perhaps better reference measurements? I know that with REW you can look at the reverberation times for certain frequencies, but the frequency resolution here is too low for me to recognize modes.
Note: I use a sweep with 32768 samples, which goes from 10Hz to 23kHhz. I'm working on a sweep that goes from 10Hz to 500Hz, but it introduces reverberation time errors that are not yet clear.
