Automatic Room Correction has won the Subwoofer Wars

The premise/purpose of the Fourier transform is simple: Any recorded sound can be broken up into discrete sinusoidal components. For instance, a square wave, no matter how perfectly square, can be decomposed into odd-order harmonic sine waves, despite the original square wave looking anything but sinusoidal.

It is not magical however, and the results will vary based on the portion of the recording analyzed.

FFT is also not a substitute for all digital signal analysis. You don’t need FFT to tell you what a visual inspection of an impulse response will, such as looking for reflections and time aligning speakers. Nor do you need FFT to create an algorithm to automatically set speaker delays.


https://en.wikipedia.org/wiki/Fourier_transform


Where the Fast Fourier Transform really transformed acoustics was in the nearly infinite resolution. We went from band limited (octave, 1/3 octave, etc) measurements to resolution bounded only by the sample length on the low end and the Nyquist frequency on the top. Outstanding.


https://en.wikipedia.org/wiki/Nyquist_frequency


And of course, let’s not forget the waterfall plots.

Best,

E

"Windowing" the impulse response to look at only the initial part shows us the frequency response of the direct sound with little or no contribution from the room.

Yes, but this windowing feature also limits the lower boundary of the FFT. The smaller the time window, the higher the minimum frequency of the FFT, and as pointed out, get low enough and the room is unavoidable.

But the entire premise of ARC is to actually correct for room behavior, so not sure how we got here. We want / need the room in measurements for ARC to work.

Quasi-anechoic measurements ( i.e. simulating a speaker without a room ) is the opposite of what we are looking for here.

Best,

E

We got here by Duke pointing out that consumer room correction, at least currently, (to our knowledge) does not take into account reflected/directly energy when doing level balancing, and hence while it can correct for phase/amplitude, it lacks a degree of freedom in setup.

Ahhh.  Well, Toole dislikes them, but I think that over the last decade it's gotten better. 

Personally, I do all my own EQ.  However!!!! Teaching a noob is a lot of work, and a lot of debate and a lot of different answers. 

ARC I think, over the last few years, has gotten better, and yields better than average audiophile results.

Could Duke or myself do better by hand?  Absolutely.