The Chain is as Strong as Its Weakest Link


An article written for the first Hong Hong HiFi Show, 1993
Translated and published in Chinese
by Arthur M. Noxon, PE, MSME, Physicist

The Chain is as Strong as Its Weakest Link
The electronic components and interconnects available in today's hi end audio are superb. Yet there is always room for improvement and changes in both the technology and style of audio. Playback audio will continue to be improved in every detail but if the minute improvements are not perceptible, they will not be heard and appreciated. New equipment that presents no audible difference from older equipment will not sell so easily and without good sales, further product development cannot be supported. The evolution of the product line slows to a halt and sales become driven by marketing hype, not performance.

The Listening Room
If a technical improvement is real it should be audible, but if it cannot be heard it becomes useless, or at best a curiosity. As long as an improvement is within the realm of human perception, it can be heard except for only one factor - the noise floor. Whenever improvements are buried in the noise floor of a system they cannot be heard. Only when the noise floor is lowered can the hidden, inner detail become revealed.

The electronic "hiss" noise floor had been a problem in audio for many years. More recent improvements in signal processing have driven the electronic noise floor down to almost the threshold of hearing. The noise floor now is more mechanical - transport decks, air conditioning and outside noise intrusion. But there is one less obvious noise floor and it is currently the single most significant sound barrier to better listening. It is the self noise of the listening room acoustic. Lingering sounds from one musical moment become the masking noise floor for the next musical moment. The chain is as strong as its weakest link and for today's audio, the weakest link is the listening room acoustic.

There are two phases of listening room acoustics and both must be understood and properly handled in order to reduce the room acoustic noise floor. We begin by recalling the piano keyboard. It is divided in half with notes belonging to the bass cleft to the left and the treble cleft at the right. For all practical purposes, it is sufficient to divide the room acoustic noise floor into these same two regions, bass arid treble.

When most people envision sound, they think of how the sound in the treble cleft behaves. It bounces around the room as if it were a ball and that is something with which we all some experience. It is because we are familiar with bouncing balls that we have an intuitive feeling for sound in the treble cleft. The bass cleft is quite different. With bass, sound doesn't beam forward and bounce around, bass goes in all directions equally - as easily backwards as forwards. Place a speaker outside, in the yard and play it while you stand behind it. What do you hear? Bass without treble. Despite this most obvious fact, the behavior of bass continues to elude people. Most of us do not have experiences that lead us to an intuitive understanding for sound in the bass range.

Old Wives' Tales
The first idea the hifi person usually has about bass these days is the idea of room modes or room resonances. There have been articles about this in the press. The basic expectation is if the room has the correct ratio it will sound great. Another popular idea is to build a room that has no parallel walls, even the ceiling could be at an angle. This is supposed to eliminate standing waves. Another often heard notion is that of killing resonance by adding bass traps or an equalizer. Amazingly, all of these too common ideas about bass and rooms are simply incorrect. Not one thing about these notions is true, but the people are not at fault. They have been taught these ideas by the sci fi writers of the hifi press whose imagination exceeds their training or experience in acoustics.

The search for the perfect room dimension is a very old one and it is based on the good idea of evenly distributing the resonances. This means each room resonance is as evenly spaced away from adjacent resonant frequencies as possible. Originally this concept belonged to the art of building reverberation chambers. Reverb chambers are used to test acoustic properties of materials and they need to have evenly spaced resonances to give good results. Reverb chambers have their speaker located in one of the tri-corners (floor/wall/wall) and the microphone in the opposite tri-corner (ceiling/wall/wall). Every room mode can be stimulated by the speaker in the tri-corner and the mic will hear every mode loudly. This very unique property only belongs to the corners of the rooms.

There are certain room ratios that give evenly distributed resonant frequencies. However, the hifi listening room is not an acoustic test chamber. Hifi speakers are almost never sitting in a corner except for Klipsch and the listeners are surely never sitting in a tri-corners of the room. In a hifi room the speaker is located away from the corner and from there it cannot stimulate all the room resonances. Which room resonances are stimulated depends on where the speaker is located in the room. Only modes stimulated by the speaker need to be well spaced apart. There is absolutely zero guarantee that a room with reverb chamber quality well spaced modes will also have well spaced modes when the speaker is moved out of the comer.

The non-parallel wall is another magic carpet concept in audio. It is true that non-parallel walls keep the flutter echo down but that is a treble effect. Sound is energy. Put this energy into a room and it stays there until one of two things happen. Sound can be absorbed by bass traps and wall friction or it can be leaked out of the room. Sound energy is stored in the room in either of two forms, organized or disorganized. Disorganized sound dies out faster in rooms than does organized sound. Organized storage of acoustic energy is a room resonance. Because the room resonance is so efficiently stored, it is the sound in the room we mostly hear. It causes "room boom" and we wish it would go away. No matter what shape the room has, the room still has room resonances. If a room has walls it has resonances. The only effect the shape of a room has on resonance is which frequency it is and how the resonance is laid out in the room. Resonance is not eliminated by changing the angle of walls.

An equalizer is irrelevant to room resonances. What is more is that they are almost never used in high end audio. A parametric equalizer changes the loudness of sound at a specific frequency. If a room resonance is "equalized", then what used to sound loud now sounds normal but the room is physically the same. The long decay of the resonance still fills the room - it is just not as loud.

Bass Traps actually improve bass in rooms. Designer built, pro sound recording studios have the bass traps usually built into the walls and they are not usually visible. Only in hifi and the smaller project or midi studios are bass traps being set up in front of the existing walls and corners of the room. Here, bass traps have become quite visible. Regardless of their location, bass traps can never "kill" room resonances. Properly designed bass traps do not weaken bass in the room. They actually increase the bass power delivered to the listener. Bass traps do not effect the direct wave from the speaker. They do reduce the strength of the room resonances that interfere with the direct wave. Bass traps reduce the sound of the room and let you hear more of the sound from the speaker.

Rooms have resonances because rooms have walls, floor and a ceiling. This is a fact of nature. Bass traps do change the sharpness or "Q" of the room resonance, no different than adding resistance to a resonant electrical circuit. No matter what the resonance, there will always be an intense sound pressure in the corners of the room. That is why pressure zone bass traps loaded into the corners of the room are so effective. By converfing 1% of your listening room volume into highly efficient corner loaded bass traps, an amazing conversion, a stability in the room acoustic sets in. Bass trap all four vertical corners and the ceiling perimeter comer with a soffit bass trap. This is the best way to control the behavior of bass in the room and causes as little visual impact as possible.

Dimensions and Floorplans
Rooms are being built exclusively for listening and home theatre more often in recent years. People always ask what are the best dimensions. It is better if the room is designed to fit the sound system, letting the dimensions fall naturally into place. Speakers are usually 7 to 8 feet ( 2.15 to 2.4 m ) apart and the listeners in biff like to be 8 to 12 feet ( 2.4 to 3.6 m )away from the speakers. Speakers should be about 3 to 4 feet ( 0.90 to 1.20 m ) off the sidewalls and 4 to 6 feet ( 1.20 to 1.80 m ) off the front wall. The listener is best located 3 to 5 feet ( 0.90 to 1.50 m ) off the back wall. This puts room dimensions to be 13 to 16 feet ( 4 to 4.80 m ) wide and 15 to 23 feet ( 4.50 to 7 m ) long. If ratios are desired, any combination of 7, 9 and 13 works well, such as 7 to 9 or 7 to 13 or 9 to 13. Standard cone drivers should never be located 25% of the width away from the side walls. The better dimension is 29 to 32% of the width from the side walls. The distance from the cone to the rear wall is best set within 10% at 1.4 or 2.4 times the distance from the cone to the side wall.

With dipoles and towers there is a strong need for a good length/width ratio and less of an issue exists about the height ratio of the room. Tall speakers stimulate strongly the front to back and side to side modes of the room, but not too much in the vertical direction because the speaker drivers line up most of the room's vertical height. If ratios are preferred only the width/height ratio is important. The speakers create ringing in the front of the room, "head end ringing". The 13/7 ratio is usually best for height but the other ratios can be used. Room height can vary between 7 feet and 9 feet ( 2.15 to 2.80 m ) but heights over 10 feet ( 3 m ) should be avoided except with tall towers.

The higher the ceiling, the lower the floor to ceiling resonant frequency. An 11 or 12 foot ( 3.30 to 3.65 m ) ceiling has a fundamental resonance of about 50 Hz and is easily stimulated because most bass drivers and especially subwoofers are located near the floor. Probably the worst ceiling to use is a peak or vaulted ceiling. The peaks reach substantial heights and their megaphone effect is not fun to listen to. If angled ceilings are used, they should be low over the speakers and open up behind the listener.

Another listening room floor plan is coming into vogue. It is the short, wide room. These rooms can be 15 or 16 feet ( 4.60 to 4.90 m ) deep and 24 or more feet ( 7.30 m ) wide and the listener is against the back wall. The room is so wide the bass build up at the back wall is not as bad as one would expect. The two wings become ambience storage chambers and the sound has a strong image presence with a strong lateral ambience. This, like all listening rooms, needs 1% volume dedicated to bass traps. The width/depth ratio of 3/1 must be avoided by 15%. The speaker separation must be 15% more or less than 1/3 the width and definitely not 50% of the width.

Another listening room arrangement that some people try to work with is to set the axis of the system down a diagonal of a square room. Huge bass traps are needed in the corner between the speakers and the corner behind the listener. The walls behind the listener should be pretty absorptive to keep imaging clear. This room can be frustrating to set up but it seems to have some ambience and stage depth properties that keep people trying to make it work.

When dedicated rooms for hifi are not available then a portion of the house is used for playback. The first rule for set up, above all others is acoustic symmetry. Without acoustic symmetry the imaging and fidelity will be highly distorted. Probably the most common asymmetric room is an "L" shaped listening room. One speaker has its back to the corner and the other speaker is against the wall in the center of the room. The corner loaded speaker requires heavy use of bass traps to balance out the bottom end so both speakers begin to sound the same. Sometimes furniture can be rearranged to create reflections that mimic the presence of a wall for the speaker that is out in the open. A cabinet can be moved in against the back wall, just to the outside of the speaker to simulate a wall corner. A cleverly positioned lamp shade can reflect the treble. Work to get symmetry in the acoustic space.

Absolutely the worst bass performance occurs when there is a room next to the speakers and the doorway is open. This coupled resonant cavity puts such a "double kick" on the bass transients it is nearly unlistenable. Close the room up or at least put heavy drapery to lower the strength of the report from the coupled cavity. Opening doors to adjacent rooms behind the listener can add to low frequency ambience which can be desirable.

When building or remodeling a listening room, the walls, floor and ceiling materials become particularly important. Today's audio gear is powerful and can deliver real punches to the surfaces of the room. Although acoustic pressure maybe quite small in a mechanical sense, when spread out over the surface of a wall, real force can be developed. At the very least the head end (speaker end) of the room needs to be mechanically inert. That means when you thump the walls or ceiling with the heel of your fist, not much happens. Concrete is inert but many homes are built out of lighter weight materials such as wood studs and sheet rock. The wood stud acts like a spring and the sheet rock acts like a weight connected to the spring. The result is "wall/stud resonance", a sympathetic resonance of the walls at about 70 Hz. This drumhead effect can drive anybody in audio completely mad. Ceiling hop occurs at lower frequencies for the same reasons. Good audio rooms will have all surfaces mechanically inert.

The wall cavity can be filled with insulation. It is always a good idea but doesn't help the wall/stud resonance. Some people wish to pour sand into their walls to add mass and establish a damping action inside the wall. The big problem with this is that sand settles, gradually increasing pressure on the lower wall until the sheetrock wall bulges, wedged apart by compacted sand. The best method is to use constrained layer damping in the wall construction while suspending the sheetrock off the studs. There is an item used to increase the isolation between walls of apartments. It is a resilient metal fir strip often called "Z-metal" "RC-1". This is installed onto the studs then the sheetrock is screwed to the it. The result is that the spring of the studs become disconnected from the mass of the wall. Now there is no spring/mass to vibrate and the 70 Hz wall boom problem is gone.

However, a new problem will have developed. We now have a large freely suspended sheet of gypsum board, a wall that can make quite a sound of its own. If visco damping material is applied between the Z-metal and the sheetrock, damping of the wall panel will occur. The next improvement is add a second layer of sheetrock with visco elastic material sandwiched between the layers. The wall really becomes inert. The ceiling can be treated exactly the same way. An alternative to the thin 1/16" thick double sided adhesive visco-elastic sheet, would be to 100% glue both sides a layer of sound board (firtex or celotex) between the two layers of sheetrock. Constrained layer damping means extra labor and cost but the unmuddled sound in the room is well worth the extra effort.

Once the surface of the room has been rendered inert and the strip type corner loaded bass traps have been installed in the four vertical corners and the ceiling perimeter corners, the listening room is ready for occupancy and detailing. It is also at this point that a number of regrets become realized. The size and placement of windows and doors can cause untold grief if they were planned by the architectural standards that apply to residential rooms and not hi end audio playback rooms.

Doors should be located on the side walls and definitely nowhere near the speakers because they rattle. Locate doors behind the listener but not flush to the corner. At least 2 feet ( 0.6 m ) of solid corner wall should always be maintained. A door on the back wall is a lessor choice due to the intense bass pressures on the back wall. The back wall door should not be located in the center or in either corner. It should begin about 2 feet ( 0.6 m ) from a corner. One of the best set ups possible is to have open french doors or archway directly behind the listener. This is a classic set up. It eliminates problems and adds benefits. Adjust heavy drapery for best effects.

Windows are another item that commonly occupy a position in a room. Large windows are to be avoided, they are too tympanic. Tall, narrow windows are the choice. Use laminated glass, the type used for glass shelving in stores. Windows are best located behind the speakers on the side walls but not close to the corners. They can also be located to either side of the listener, preferably slightly behind the listener, but slightly in front of the listener is okay too. One of the biggest mistakes is thinking a thermal window is good for acoustics - it isn't. It's terrible. The glass is very thin and the air space is negligibly small for acoustic purposes. Standard sliding glass doors are never to be used in the hi end room as they resound with tympanic thunder.

Windows allow distracting outside noise into the room as well as let sound leak out. To minimize sound transmission problems, use a double window. The two sheets of glass should be separated at least by 4 inches of air space. The glass sheets ought to be of different thickness. The space between the glass is to be vented into the wall cavity. Frame work and trim molding need to be sealed with expandable foam and acoustical caulking. Set all glass into a bed of visco-elastic damping material.

Speakers don't play by themselves, they require equipment to drive them. Power amps are best left near the speakers with short connecting cables. The preamp and all other equipment is best positioned slightly behind the listener, against the wall. Some people are building equipment closets, much like the equipment rooms of recording studios. A glass door and remote control work well together.

There is a tradition in mid fi that tends to creep into the hi end. The electronics are piled into a custom built wood and glass equipment cabinets centrally located on the wall, directly behind the speakers. All cabinets, or shelving in hi end audio must be acoustically porous, not reflecting. Any kind of cabinetry or equipment stacked up between the speakers locates it directly in the middle of the sound stage. The last thing any of us really want is their sound stage filled up with sound reflecting equipment, blinking lights, and hollow resonant boxes. The sound stage needs to be absolutely free from visual or acoustic distraction. It needs to be an acoustically quiet backdrop upon which sub fie images play out their action.

Lighting in the playback room needs to be versatile. Bright lights are needed during set up, adjustments or upgrade installs. But for playback and imaging, it is best if there are no strong visual distractions in the front of the room. We want our eyes to be relaxed, wide open but without anything to catch our visual attention. Visual activity competes with acoustic imaging. It is best if the front of the room appears to be nothing more than vague visual shadows in a dim grey fog.

The lights for playback are best located behind the listener and they should be diffuse. Hollow can lighting is attractive but they make a lot of noise. You will hear them in a good room. You can also hear singing metal waste paperbaskets in a rood room, (use wicker). Best light has a ceiling bezel and lens of thick slightly rounded glass. A dimmer is appropriate but many of the electronic dimmers create a buzz of the filaments when dimmed down and we prefer to avoid introducing noise into a playback room. A variable transformer voltage control for lights does not produce noise. Low voltage lighting should be considered. There may be electronic dimmers that are quiet but be sure to test them before installing them.

Wall reflections and ceiling reflections obscure imaging detail, stage depth, smooth lateral positioning and musical accuracy. This is because the early, first few reflections are still strong and arrive within the sound fusion time period. Our listening process correlates all the early reflections that arrive within 20 milliseconds of the direct signal. What we think we hear is really the sum of the direct and early reflected signals. This is primarily a treble range effect. The single worst reflection is the crosstalk reflection. This occurs for example when the left speaker plays off the right wall and we hear the reflection in our right ear. Just after our right ear had received a signal from the right speaker it is fed another signal from the left speaker. The result is essentially the loss of stereo and we recognize it because of the "hole in the middle" soundstage. If there is but only one reflection to control, it is the crosstalk reflection.

Placing absorption at the reflection point will reduce the problem of reflections. The question becomes, where is the reflection point? The crosstalk reflection is on either side wall, at ear level and located about 1 1/2 feet ( 0.46 m ) behind, towards the listener from the halfway point between the speaker and the listener.

A very accurate and easy method to locate the reflection point does exist. It is an optical alignment technique known as "ray tracing" and it is learned by all students of acoustics. Tape up a strip of mylar reflecting plastic sheet on the wall or have someone hold a mirror fiat against the wall. Use a strong, narrow beam flashlight and hold it near your head and shine the light on the mirror. Observe where the light lands. Adjust the mirror and light beam until the light lands on the midrange speakers and tweeters. Mark the spot on the wall where the reflection occurs because that is where your wall acoustic treatment will be placed.

The wall absorption should be effective at least through 200 Hz in order to cover the treble cleft. There is one problem with simple, commodity type absorption. They usually do not absorb in the lower treble octave and even worse these office panels, which are fabric covered noise control panels or blocks of corrugated foam, leave a "black hole" in your acoustic periphery. This is very distracting for those of us who worked hard to develop a sensitivity for sound and imaging. Wall absorbers do need to absorb the early reflection but they can also be fit with reflectors. Sound that bounces off the rear wall, behind the listener will graze back up the side walls, impact the reflectors and backscatter towards the listener.

The best of both worlds is now attained. Early reflection absorption is followed by time delayed, low level ambient reflections and all from the same spot on the wall. This technique of using time delayed backscatter sound absorption produces the very desirable effect of lateral ambience recovery. It is used in properly designed recording studios and it can also be set up at home. The important feature in hi end sound absorbers is the presence of some sort of reflective quality. It may seem like a contradiction of terms but it is really a matter of refined taste.

There is another reflection in the room that is particularly hard on the listener. It is the rear wall bounce. Sound passes by the listener only to hit the rear wall and bounce back past the listener again. If the positive phase of a sound wave is bounding back just when the negative phase of the wave passes the listener - they cancel. This occurs when the round trip distance between the listener and the back wall equals 1/2 wavelength. For example, sitting 4 feet ( 1.2 m ) from the wall means cancellation occurs for a 16 foot ( 4.8 m ) wavelength, a 70 Hz frequency. This is not a room resonant problem. The low frequency "cold spot" depends only on the distance the listener is from the rear wall. Bass traps behind the listener against the wall must be used to control this problem. Remember, treble range reflectors across the face of any bass trap keeps the ambience from feeling dead. The goal in hi end acoustics is of course acoustic control, but control done in such a tasteful way that you can never hear where the control comes from.

The Last Link
By following the above details, the perfect room will still not be realized but the room you have will be so good that you will be very happy with it for a long time. You will recognize it is a good room because the speakers will seem to completely disappear. They become silent projectors of an acoustic holograph that appears in the front of your room.

Just remember, acoustic modifications, detailing and upgrades are every bit a part of today's hi end audio system as the speaker. Modern hi end audio electronics are of such quality that the room acoustic simply has to be improved in order to even begin to hear the hi end part of hi end audio. The audio chain can be only as strong as its weakest link, which for now is the playback listening room.