Generating impulse response directly from graphic equalizer in Fourier domain?

[gavint]

Hi,

I have a set of Dolby speaker tuning parameters (bottom of this message) which I want to turn into an impulse response so I can use it in Linux (EasyEffects/Pipewire).

Currently I load a flat response (i.e. 0 dB at 5 different frequencies) as my target, fit a set of parametric biquad filters to my IEQ data (House curve) in the EQ window, generate measurement from filters, then generate an impulse response from the new measurement.

I noticed that this person generates the impulse response directly from the IEQ data using FFT:

//
FIR generation

The script generates minimum-phase FIR filters via cepstral processing:

1. Interpolate the IEQ to FFT frequency bins
2. Compute the real cepstrum (IFFT of log-magnitude)
3. Apply causal windowing to get minimum-phase cepstrum
4. Reconstruct via FFT → exp → IFFT
5. Normalize so peak frequency response = 0 dB

//
https://github.com/antoinecellerier/speaker-tuning-to-easyeffects

Is this a good way of generating an impulse function?

IEQ Data

(Hz, dB)
47, -2.97
141, 0.33
234, 4.84
328, 2.44
469, -0.91
656, 0.54
844, 2.58
1031, 0.58
1313, -0.69
1688, 0.09
2250, 0.84
3000, -0.05
3750, 3.99
4688, 3.45
5813, 1.60
7125, 0.83
9000, 1.17
11250, 1.08
13875, -0.84
19688, -2.77

 

[John Mulcahy]

Put the response points in a text file, drag the file onto REW then right click ion the graph and use "Generate minimum phase" (assuming V5.40):

 

[gavint]

ok. thanks.

can i ask if the mathematical process to generate the impulse response directly is the same as the one i described above (cepstral processing)?

also i might want to directly generate a phase-neutral (linear-phase) impulse response or somewhere inbetween ((see discussion at the bottom of the link below on Causality). is that possible with room EQ wizard?

https://github.com/erdewit/HiFiScan (bottom of page)