Yes, FFT analysis has gotten pretty inexpensive since those early days in the mid 1980’s when we had to buy the Crown Techron 12 or 20 for something well over $10k…but regardless of when or how much it cost, once you have the FFT analysis, then what?
Read along as ASC founder & president, Art Noxon, PE responds to studio performance questions.
If the FFT waterfalls look smooth and fast, does that mean the studio sounds good?
If the frequency response measures pretty flat, does that mean your mix will hold together? What is the connection between the output of FFT analysis and being able to mix a record? Isn’t that the bottom line?
In our world of TubeTraps, we have continuously worked on recording studios since 1985. We tried to connect the dots between FFT printouts and studio performance but we ran into major problems. In particular, the uncertainty principle wiped our efforts at FFT analysis out. It’s the old dF x dT = 1 problem. In the deep bass range, say below 100 Hz, lies the first two octaves of sound, which is where most of the hard to solve, real problems in room acoustics are found.
FFT loses resolution down in the low octaves and the FFT printouts begin to become almost useless. We want to see a narrow bandwidth, say 2 Hz wide in the deep bass range and we want to see the response of the room with a resolution of at least 1/8th second. Here’s the rub. For dF = 2 Hz, we have dT = 1/dF = 1/2 = 0.5 seconds resolution and we wanted 0.125 seconds resolution. Here’s the flip side of that problem…we want 1/8 second resolution and our frequency resolution ends up being dF = 1/dT = 1/(1/8) = 8 Hz.
Wow, we are looking into the first octave, between 20 and 40 Hz and there is 12 notes in that octave which is 20 Hz wide. Each note is about 20/12 = 1.7 Hz apart. The 1/8th second resolution gives us a frequency resolution of about 8/20 = 1/3rd octave. Another way to put this is that a bandwidth of 8 Hz includes about 8/1.7 = 5 notes out of the 12 that makes up that octave. So FFT produces nothing but smeared data in the deep bass range: Time smeared or frequency smeared, take your pick, but you can’t have both in the deep bass region, which is where all our lumpy bass problems lie. All this happened to us very early on and by 1989 we gave up on FFT as being a useful tool for deep bass exploration.
So Then What?
We discovered the magical world of MTF, (Modulation Transfer Function) testing. It measures how fast your room is. It sort of measures how loud your room gets during a very short tone burst and how quiet your room gets during a very short window of silence which immediately follows the tone burst. It measures how quickly a sound level can change in your room. It turns out that engineers and other golden ears are much more interested in the high-speed dynamic nature of the room, much more than its frequency response when it comes to performance.