This matter is interesting as long as I have not fully understood room correction for the last 12 years . However I understood that it was enough to use in Acourate pro a psychoacoustic treatment with a frequency dependent window of 15/5 ie 15 for the bass and 5 for the treble . I thought that would correct mainly the loudspeaker inacuracy for the treble and make a real room correction for the bass only. So the correction is on the complete spectrum but room correction is limited to the bass .Am I right ?
This explanation is a bit involved. First i'll explain what it is we want to correct, and then i'll tell you the difference between Acourate's FDW (frequency dependent windowing) settings of 15/15 and 15/5.
What to correct
Take a look at this graph which shows the in-room frequency response taken with MMM of the same loudspeaker in 9 different rooms:
From
loudspeakers.audio (go down to "Acoustics and EQ" and click on "Correction at mid and high frequencies" to view this graph).
Notice that above 800Hz the measurements hug each other tightly, the only variation is the downwards slope of the treble response. Some have more tilt, some have less. The variation is about 5dB.
Below 800Hz, the FR is increasingly chaotic as wavelength gets longer.
From this graph, we can conclude that the room and listening position has a substantial influence on bass frequencies, so this is what we want to correct. For higher freqs, it is listening distance that influences the treble tilt.
Now, these are MMM's which means that the treble response has been spatially averaged. If you use Acourate, most people typically correct for a single microphone position. The problem with a single mic position is that it is a single point in space. For low frequencies, the wavelengths are very long - a 100Hz sound has a wavelength of 3430mm which is very large when compared to a typical human head (150mm). So a single mic position is likely to be representative of what you will hear.
As wavelengths get shorter, that mic position becomes more and more specific to that particular location in space. A 20kHz sound has a wavelength of 17mm, which is much smaller than the width of your head. This means that what the mic captures is highly unrepresentative of what you will hear. The solution to this is:
1. Do
not use a high resolution single point measurement as the basis of correction for high frequencies. Use a spatially averaged measurement like an MMM.
2. Do
not perform a "room correction" on the high frequencies at all. Instead, do an anechoic or quasi-anechoic correction of the high frequencies (Acourate lets you do this).
3. If needed, apply a treble tilt to taste.
Acourate FDW
Acourate lets you use any measurement as the basis of your correction. Most users use a single point measurement for the sake of simplicity. You
can use an MMM or a spatially averaged measurement as basis for correction, and I recall describing how to do it in my
Acourate guide. If you use a single point measurement, you can adjust the smoothness of the correction using different FDW settings.
For bass correction, we want to correct the speakers together with the room. So we use FDW 15. This lets 15 cycles of sound into the correction and will produce a fairly high resolution picture of what is going on.
For treble correction, we do not want a high resolution correction. So I personally use FDW 1 or FDW 5 depending on my mood
The lower the FDW, the shorter the window, the smoother the curve. This is not quite the same as a spatially averaged correction, but it is very close. If you are a purist, you should be using an MMM or spatially averaged measurement as the basis of your correction. But you are likely to find that the shape of the correction is pretty similar to an extremely smoothed high freq single point measurement.
I've done it both ways, and have some gear where you can't limit it too.
