Rational Acoustics



Dave Gunnell
January 17th, 2013, 05:36 PM
Hi All:

I am refining my sub tuning methodology (EAW BH760 subs), and I'm looking closely at the 25-50 Hz range. I believe Smaart is a Fixed-Point Per Octave (FPPO), or Multi Time Window (MTW) tool, where multiple FFTs of different sizes are computed and pieced together to form the end display. Within any given FFT, the lower frequencies will have lower resolution than the higher frequencies. Does anyone know if there's a way to increase the resolution in this VLF area? Also, how do I know what the transfer function resolution is at, say, 30 Hz? In certain rooms I see some bumps below 40 Hz, but I can't tell if their BW is being 'spread' by lower Smaart resolution at these low frequencies.

Any thoughts would be appreciated. Thanks.

Dave

Arthur Skudra
January 17th, 2013, 07:04 PM
If you're using MTW, it's 1 Hz or better in resolution from 20 Hz to 140 Hz. Do a capture of a trace, export it as ASCII, then have a look at the datafile and the resulting resolution!

However, be really careful in doing anything at these low frequencies, particularly indoors! Anything below the critical frequency of the room, and you will be dealing with room modes and *very* position dependent anomalies.

How to calculate the critical frequency?

Fc = 3C / RSD

Where Fc is the critical frequency, C is the speed of sound, RSD is the room's smallest dimension.

Typically the RSD is the ceiling height of the room. So for a RSD of 20 feet,

Fc = 3 x 1130 / 20 = 170Hz

So anything below 170 Hz, I'd advise a lot of caution before you start tweaking away!

Moral of the story, measure and optimize subwoofers outside!!

gluis
January 28th, 2013, 05:00 PM
Hello Arthur,

I have one question: What's the reasoning behind that formula? I'm not questioning it, I just want to understand what's behind it.

Regards,

GS

Harry Brill Jr.
January 28th, 2013, 11:22 PM
I'm curious what makes you think Smaart has a lower resolution at low frequencies? It's actually the other way around sort of. The resolution is musically distributed the way our ears hear. Arthur is correct though,...1Hz down low.

Arthur Skudra
January 30th, 2013, 01:14 PM
Hello Arthur,

I have one question: What's the reasoning behind that formula? I'm not questioning it, I just want to understand what's behind it.

Regards,

GS
Hi gluis,
Sorry I haven't responded sooner, been really busy here! Anyway to answer your question, below that critical frequency, in an enclosed space, the room behaves like a giant bowl of jello, room modes are plentiful. The problem is that the low end response below this frequency will be *very* different from one position to another, depending on whether you're in a peak or null at certain frequencies. The smallest dimension of the room is the determining factor, usually the ceiling height of the room. I should note that this is not a hard transition, but a gradual one, kinda like a transition zone.

Edit: You can read up on it more in Davis' and Patronis' book "Sound System Engineering" 3rd edition, page 178. There they reference previous work done by Bolt, Beranek, and Newman on room acoustics. To quote Davis & Patronis: "Fc coincides with the dimension of the room equal to the lowest wavelength that can fully develop across that dimension." Bolt, Beranek, and Newman developed a chart called "controllers of steady-state room acoustic response" (see the graphic below) as it pertains to small room acoustics. A lot of this is applied in the design of LEDE control rooms. Here they describe four zones: the pressure zone, modal zone, diffusion zone, and the specular reflection zone, that determines how acoustic treatment is best used at different frequencies. From this you may deduce some cool implications on how we optimize sound systems.
663
My personal observations here are that in the modal zone (below Fc), response varies widely dependent on where you put your mic. Even with multiple mic averaging, you really need to be cautious! I'll typically look at the upper "envelope" of all the individual traces combined in this frequency region. In the diffusion zone (between Fc and 4Fc), here you benefit significantly from multiple mic averaging, and incidentally this is the region of frequencies that you do most of your work in sound system optimization (for rooms with a smaller dimension). In the specular reflection zone (above 4Fc), you really need to be very cautious that you're not trying to correct for a specular reflection causing addition or cancellation in your response, so here a coherence weighted spatial average is helpful using multiple mics. This is also typically where loudspeakers are most directional, so mic position is also very important. Again, Fc should not be considered a hard and fast change from one mode to another, but rather a gradual transition. Also, my sound system optimization "observations" in relation to where 4Fc is positioned falls apart in much larger rooms (ie large auditoriums or stadiums), but in those situations, a specular reflection problem is pretty obvious with your ears and you take whatever measures need to be taken (eg re-aim the system or treat the offending surface).

All this to say that the primary "take home" idea here is knowing at what frequency the modal behaviour of the room starts to take over (Fc) and thus you need to exercise judicious caution in your system optimization below that frequency. Hope this helps!

gluis
January 31st, 2013, 05:53 PM
Thanks Arthur,

Actually I'm clear about the concept, just curious about the formula. Specifically why the 3 times C in it.

Regards,

GS

Arthur Skudra
January 31st, 2013, 07:05 PM
Thanks Arthur,

Actually I'm clear about the concept, just curious about the formula. Specifically why the 3 times C in it.

Regards,

GS
Because of this:
To quote Davis & Patronis: "Fc coincides with the dimension of the room equal to the lowest wavelength that can fully develop across that dimension." (p 178, Sound System Engineering).

They are saying that you need 3 full wavelengths across a room's smallest dimension to fully develop at that frequency and higher. Which makes sense...the higher in frequency, the shorter the wavelength, the more periods that fit in the room's smallest dimension, thus the density of the peaks and nulls increases and are more "diffuse" over a given area. Go lower than Fc, and the peaks and nulls become more and more obvious as you walk through the room and observe the fluctuations in response.

Hope this helps!

gluis
February 1st, 2013, 08:23 AM
It certainly does :-)

I should had checked on my books before asking, but I recently moved, and all my books are in my former residence in Margarita island. As I'm not sure how long I'm going to be where I am (maybe a couple of years), I don't want to move my stuff internationally.

THanks for your help

Regards,

GS

Harry Brill Jr.
May 9th, 2013, 01:16 PM
Very nice Arthur!


Hi gluis,
Sorry I haven't responded sooner, been really busy here! Anyway to answer your question, below that critical frequency, in an enclosed space, the room behaves like a giant bowl of jello, room modes are plentiful. The problem is that the low end response below this frequency will be *very* different from one position to another, depending on whether you're in a peak or null at certain frequencies. The smallest dimension of the room is the determining factor, usually the ceiling height of the room. I should note that this is not a hard transition, but a gradual one, kinda like a transition zone.

Edit: You can read up on it more in Davis' and Patronis' book "Sound System Engineering" 3rd edition, page 178. There they reference previous work done by Bolt, Beranek, and Newman on room acoustics. To quote Davis & Patronis: "Fc coincides with the dimension of the room equal to the lowest wavelength that can fully develop across that dimension." Bolt, Beranek, and Newman developed a chart called "controllers of steady-state room acoustic response" (see the graphic below) as it pertains to small room acoustics. A lot of this is applied in the design of LEDE control rooms. Here they describe four zones: the pressure zone, modal zone, diffusion zone, and the specular reflection zone, that determines how acoustic treatment is best used at different frequencies. From this you may deduce some cool implications on how we optimize sound systems.
663
My personal observations here are that in the modal zone (below Fc), response varies widely dependent on where you put your mic. Even with multiple mic averaging, you really need to be cautious! I'll typically look at the upper "envelope" of all the individual traces combined in this frequency region. In the diffusion zone (between Fc and 4Fc), here you benefit significantly from multiple mic averaging, and incidentally this is the region of frequencies that you do most of your work in sound system optimization (for rooms with a smaller dimension). In the specular reflection zone (above 4Fc), you really need to be very cautious that you're not trying to correct for a specular reflection causing addition or cancellation in your response, so here a coherence weighted spatial average is helpful using multiple mics. This is also typically where loudspeakers are most directional, so mic position is also very important. Again, Fc should not be considered a hard and fast change from one mode to another, but rather a gradual transition. Also, my sound system optimization "observations" in relation to where 4Fc is positioned falls apart in much larger rooms (ie large auditoriums or stadiums), but in those situations, a specular reflection problem is pretty obvious with your ears and you take whatever measures need to be taken (eg re-aim the system or treat the offending surface).

All this to say that the primary "take home" idea here is knowing at what frequency the modal behaviour of the room starts to take over (Fc) and thus you need to exercise judicious caution in your system optimization below that frequency. Hope this helps!

FILO4PRES
March 21st, 2014, 05:50 PM
Good points below,

Id also suggest using a log swept sine or (Pink Sweep). Three averages should be plenty couple seconds say three to sweep the subs. You'll be able to drive the subs harder without damaging them improving the signal to noise ratio in the measurement and giving you a stable phase trace even in big rooms or noisy rooms.

Ten years ago you might have to tweak the sample rate etc but I dont believe thats necessary anymore.