|
|
Home | Audio | DIY | Guitar | iPods | Music | Brain/Problem Solving | Links| Site Map
This work is licensed under a Creative Commons License.
Room To Move
June Tech’s Files
2003 by Eddie Ciletti
Two previous columns addressed low frequency perception, the region where our ears are least sensitive. When "bass judgement" errors are made, one need only look at the Equal Loudness Curve — it details the ear’s spectral sensitivity, which is uneven at all Sound Pressure Levels (SPL) from the Threshold of Pain down to the Threshold Of Hearing. This column addresses basic Control Room Acoustics issues for people who work in non-designed spaces. While the emphasis will focus on the bass region, analysis of the sample room revealed other issues that were also addressed.
COOL TOOLS
You may already have the basic tools, if not, here’s the list — a sweep oscillator, omni-directional microphone, any device with a high-resolution meter and digital converters. Some less-obvious but easily acquired tools include balloons and a rubber mallet. In addition, software for this month’s sonic shakedown included Cool Edit Pro, SMAART and Wavelab. A three-part SMAART "techniques" series is available online.
BOUNCING OFF THE WALLS
There’s no such thing as a perfect room, even those that are better by design, but the goals are the same. The audio spectrum is divided into five bands, minimizing the anomalies evenly across the spectrum. Treatment for each frequency band is not the same because bass behaves so much differently than treble. One of the most basic tests is to listen in mono for the phantom center, then switch back to stereo. Is the image stable? Can you "touch" it? Is it full bodied or scattered? Optimized room focus more sound directly at the listener. What can degrade the image? Let’s start with clutter…
To start, tidy up the crib. Remove all gear that may be obstructing or reflecting the primary path between the monitors and the listener. If things tend to slide around, redefine the center of the room, get out the tape measure and symmetrically align the work area —the monitors, the console and equipment racks. Detail the room’s particulars using familiar artistic tools (paper or virtual) taking care to make accurate measurements.
From an acoustics measurement perspective, the three tests are spectrum analysis, decay time and impulse response. Spectrum analysis displays amplitude versus frequency but does not provide any insight into why the response is so lumpy. Adding the dimension of Decay Time reveals the presence of reflections while Impulse Response can reveal specific reflections all the way down to the timing between woofer and tweeter.
That each band should have a similar decay time cannot be over emphasized. It’s pretty obvious that high frequencies — responsible for positional (localization) cues — are easily reflected by hard, smooth surfaces. To get the best "cues" you want only direct sound from the monitors. High frequency reflections can be tamed three ways — by repositioning equipment, by applying porous absorptive material in the immediate sound field and by using diffusion (uneven surfaces) in areas that will not distract the listener or corrupt the stereo image. Too much absorption will exaggerate the problems in other frequency bands.
In order for bass to bounce off a wall in similar fashion to treble that structure would have to be infinitely more rigid and dense — like sand-filled concrete block or pored concrete. Sheet rock walls can absorb, reflect or resonate when stimulated by bass and midrange. Whether by accident or on purpose, this example shows how wall construction and problem-specific traps can be built to physically tune the room, like an equalizer and including "Q" (bandwidth). This can be good, bad or ugly depending on where in the room this is happening, room size, shape and the ratio of length, width and height. Get out your fist or rubber mallet and pound the walls listening for resonance — it will be all over the map.
What other accidental contributors might there be? Windows were found to be a problem in the test room (not the OS) as well as the ductwork inside the walls. It can be a nightmare, especially when using a generic room for sonic purposes. If you are at the planning stage, eval the room before putting all the gear in because ripping out sheet rock is a messy business.
NOTES:
Room Dimensions — the relationship of height (h), length (l) and width(w) can be magical or disastrous; a cube-shaped room is least desirable. The ratio of these dimensions will play a large part in determining the "modes," places where build-ups occur. Two examples of preferred ratios are 1:1.14:1.39 (a small room that’s 10h x 13.9l x11.4w feet) and 1:1.6:2.33 (a larger room of 10 x 23 x 16 feet).
The Reflections — not a cover band but three "modes" that detail the number of surfaces sound will hit, and bounce from, in an untreated room — axial (two surfaces), tangential (four surfaces) and oblique (six surfaces). The dimensions and the number of reflections will create sonic bumps and dips across the frequency spectrum. Larger rooms have less destructive modes and support lower frequencies.
SHAKE DOWN
A sweep oscillator is all that’s needed for the first test — it’s not for absolute measurement purposes but feel free to put up an omni-directional mic at the sweet spot and document the proceedings. Start at 1kHz and slowly sweep down into the bass region while listening for (and fixing) rattles. This might keep you busy for a while. If taking notes, do the test three times for Left, Right and both monitors. You’ll be amazed how different the response for each will be.
While an oscillator is not the right tool for the job, it does provide some insight. It should be perfectly flat, but I’ll bet you noticed lots of peaks and dips during "the sweeps." A well-played and dynamically processed bass part can be similarly flat, although a funky room can make it seem inconsistent. Knowledge of music and of the notes that seem boomy or lost can be integrated into the interrogation process. A standard electric bass covers over three octaves, the translation of notes into frequency are shown in Table-1.
TABLE-1: Electric Bass Guitar Frequencies
E StringString A String
D String
G String
E1 41.20
MIDI#28
A1 55.00
MIDI#33
D2 73.42
MIDI#38
G2 98.00
MIDI#43
F1 43.65
MIDI#29
A#1/ Bb1 58.27
MIDI#34
D#2/ Eb2 77.78
MIDI#39
Gb2/ Ab2 103.83
MIDI#44
F#1/ Gb1 46.25
MIDI#30
B1 61.74
MIDI#35
E2 82.41
MIDI#40
A2 110.00
MIDI#45
G1 49.00
MIDI#31
C2 65.41
MIDI#36
F2 87.31
MIDI#41
A#2/ Bb2 116.54
MIDI#46
G#1/ Ab1 51.91
MIDI#32
C#2/ Db2 69.30
MIDI#37
F#2/ Gb2 92.50
MIDI#42
B2 123.47
MIDI#47
A1 55.00
MIDI#33
D2 73.42
MIDI#38
G2 98.00
MIDI#43
C3 130.81
MIDI#48
A#1/ Bb1 58.27
MIDI#34
D#2/ Eb2 77.78
MIDI#39
Gb2/ Ab2 103.83
MIDI#44
C#3/ Db3 138.59
MIDI#49
B1 61.74
MIDI#35
E2 82.41
MIDI#40
A2 110.00
MIDI#45
D3 146.83
MIDI#50
C2 65.41
MIDI#36
F2 87.31
MIDI#41
A#2/ Bb2 116.54
MIDI#46
D#3/ Eb3 155.56
MIDI#51
C#2/ Db2 69.30
MIDI#37
F#2/ Gb2 92.50
MIDI#42
B2 123.47
MIDI#47
E3 164.81
MIDI#52
D2 73.42
MIDI#38
G2 98.00
MIDI#43
C3 130.81
MIDI#48
F3 174.61
MIDI#53
D#2/ Eb2 77.78
MIDI#39
Gb2/ Ab2 103.83
MIDI#44
C#3/ Db3 138.59
MIDI#49
F#3/ Gb3 185.00
MIDI#54
E2 82.41
MIDI#40
A2 110.00
MIDI#45
D3 146.83
MIDI#50
G3 196.00
MIDI#55
Original Chart Copyright © 2000-2002 by Grant Green
Modified for Electric Bass by Eddie Ciletti and permission granted for use ã 2003
Once the problematic frequencies are known it is possible to start calculating for the offending distances using the formula…
Wavelength = 1,126.8/frequencyThe "constant" is 1,126.8 feet per second, the speed of sound at room temp. Having documented all of the bumps and dips, plug a few of those frequencies into the formula to see if any wavelengths correspond to the room’s obvious dimensions. My room, depicted in Figure-1, had a bump at 160Hz — just a bit shy of E3 — its 7-foot wavelength corresponded to the ceiling height. An online wavelength calculator can be found at www.eatel.net/~amptech/elecdisc/frequncy.htm — just plug in the numbers and play.
Figure-1: The very modest control room under scrutiny.
The orange rectangles represent type 703 fiberglass panels used to absorb reflections.
HEAVY ARTILLERY
After the basics were out of the way, I consulted with Dave Meyers at www.overkillaudioinc.com. He brought over balloons and we popped quite a few in several places — in front of the monitors and in each cavity (entrance and storage). Even before analyzing the recording this test helped us find several sympathetic resonators — like the window behind the left speaker and ductwork along the right speaker wall. These remarkably obvious problems were otherwise hidden when listing to music.
Dave suggested SMAART Live ($695 list), a powerful and affordable analysis tool that runs on a PC. A demo version is available from www.siasoft.com — it runs for 30-days SMAART’s spectrum analyzer has resolution up to 1/24th octave (1/12th was used). Pink noise was pumped into the average undersized and under-treated control room (mine) and considering the sine sweep, it wasn’t too frightening. The "before" image was saved and combined with the "after" as detailed in Figure-2.
Figure-2: The BLUE spikes represent the original curve pre-tweaks.
GREEN represents the curve apres tweaks. Post tweak "gains" are shown in RED.IMPULSIVE
In a large space, like a studio or concert hall, Reverberation time is stated as RT-60 (the time required for the signal to be attenuated by 60dB). It’s a little different in a Control Room where the "R" is more like "resonance and reflection" than reverberation.
Steinberg’s Wavelab has great time-domain capabilities. Using its signal generator to create an impulse, its time domain tool helped determine resonance (for low and mid freqs) and reflection (mid and high freqs). You can do a similar test on any editor by recording a balloon "pop" and then measuring the reflection in the waveform. Cool Edit Pro. Figure-3 shows the waveform as an insert (top right) along with the full-color analysis below. In this instance, time is from rear to front with the audio spectrum from left to right. This relationship can be reoriented to your preference.
Figure-3: The insert at top right is an impulse reproduced by the monitors and captured
to measure reflections (the double arrow) and analyzed for resonance below.As we learned via the balloon tests, the window contributed several very low frequency aberrations (plus rattling springs) while the ductwork quacked in the mid-band around 300Hz. Above 160Hz the decay is obvious but below that frequency there is almost no decay within the half-second range of capture. In the insert, time moves from left to right, the large double arrow is the distance between the impulse and it’s first reflection. When the area is highlighted in Wavelab’s waveform window, the time in milliseconds is displayed. From there it was easy to calculate the actual distance of the reflection using a ratio, again based on the speed of sound…
1,126.8 feet / 1sec = x / .01sec (10mS) = .01 x 1,126.8 = 11.268 feet a.k.a. room width
("x" is the distance we’re looking for. Just cross multiply. Since both denominators are in seconds and the left denom is "1," there’s no need to divide.)
THE FIXES
I spent the next day applying damping materials to the windows, window springs and ductwork — none of which was considered the final treatment, but just to see if anything measurable would happen. It did! I also reoriented an absorber panel, opposite the left monitor, by 90-degrees. The results are displayed in Figure-2 where GREEN represents the curve after tweaks. Any "gains" are shown in RED while the BLUE spikes represent resonance reduction (attenuation). The gains at far right may have been microphone orientation. The losses at left were most likely the result of window dampening. In between, many of the losses were in the 3dB to 4dB range while 1dB to 3dB gains were realized.
TADA!
For each attempt at tweaking control room acoustics I have been rewarded. These have all been little gains and while this project’s goal was to improve the low frequency response — the sonic upgrade was full-spectrum. Most interesting was that the midrange listen-ability improved to the point where I wanted to pump up the volume. (This directly correlates with the previous column on directional mics — many have a midrange bump to compensate for proximity effect.) The phantom center became more like a sonic hologram, the improved stereo image revealed that the D/A outputs required L/R calibration. Previous tweaks — an overhead absorber — improved the localization range between 5kHz to 7kHz. Further tweaks extended this into the mid-band. Now there is more depth, more impact and more intimacy.
I did not go into detail about the materials used — mostly panels of type 703 fiberglass — because the experiment is ongoing. (The 700 series of glass fiber from www.owenscorning.com is the choice of acousticians because of its effectiveness, density and fire rating.) The two windows in the room should really be replaced and the sheet rock hiding two lengths of ductwork removed for full damping. There are still some peaks to tame, materials to try (3M Thinsulate, it’s not just for gloves and coats) and traps to build — the makings of a future article for sure.
Ultimately a control room should minimize the guesswork. Of course you could hire a professional and I highly recommend that you find one that’s compatible with your needs. But as always, I also recommend the DIY approach for the learning experience — worst case you’ll know what questions to ask.
If you want to learn more, install a copy of "How to Build a Small Budget Recording Studio From Scratch," recently updated by Mike Shea (TAB books), into your porcelain office of choicce. It’s an excellent resource of info, materials and sources.
Eddie would like to thank Dave Meyers, Terry Hazelriga nd Wes LaChot for sharing their expertise.
Home | Audio | DIY | Guitar | iPods | Music | Links | Site Map | Contact
