Se også denne video:

http://www.princeton.edu/3D3A/

Introduction to 3D Audio with Professor Choueiri

Karma-Audio.dk

Lytterummet/stuen

Absorptionskoefficienterne for æggebakker på en væg.

Man kan se, at den er ganske effektiv ved høje frekvenser, men at den ingen virkning har ved lave frekvenser.

-sæt en rockwoold bat bag æggebakken så virker det.!!

Ligegyldigt hvor mange penge du bruger på dit hi-fiudstyr, så afhænger det hele i sidste ende af lytterummets indretning og akustik.!

Karma-Audio.dk

Lyden i rummet

November 2007

Akustik er ikke underligt

Hvad er det der gør at et rum lyder som det gør og hvordan får man det til at lyde som man gerne vil?

Eddy Bøgh Brixen, akustiker

Det med akustikken kan godt forekomme at være en tricky affære, der ikke sjældent ender et helt andet sted, end det man havde tænkt. Men sådan behøver det ikke at være - man skal bare tage højde for akustikken så tidligt som muligt når man planlægger kontrolrum, studier, demorum, AV rum, auditorier mv. Det er nemlig som regel dyrere at rette på et rum med dårlig akustik end at bygge et, der er i orden fra starten.

Her vil vi se på det, der hedder rumakustikken, dvs. lyden i selve rummet. Dette har i princippet ikke noget med lydisolering at gøre - det er en helt anden disciplin.

Videnskab i 100 år

Akustikken som videnskab er godt 100 år gammel. Dengang var det den amerikanske fysikprofessor Wallace Clement Sabine, der gik rundt på Havard universitetet og udsendte lydimpulser med en orgelpibe, hvorefter han med sit stopur kunne konstatere, at lyden havde forskellig udklingningstid i forskellige rum. Han fandt en sammenhæng og skabte den formel, der i dag bærer hans navn.

Når vi skal se akustisk på et rum er der to grundlæggende faktorer: Rummets størrelse (volumen) og mængden af lydabsorption.

At finde rummets volumen er jo relativt simpelt, til gengæld kan det være noget mere besværligt at beskrive lydabsorptionen i rummet.

Efterklangstid

Sabine definerede begrebet efterklangstid, dvs. den tid lyden er om at dø ud efter lydkilden er stoppet.

Efterklangstiden er den vigtigste parameter inden for akustikken. Afhængigt af rummets anvendelsesformål vil der være en optimal efterklangstid. Er det et kontrolrum vil efterklangstiden som regel skulle ligge mellem 0,2 og 0,4 sekund. Er det et studie, vil efterklangstiden være afhængig af hvad det skal bruges til. Et TV studie må som regel gerne være godt dæmpet og afhængig af størrelsen ligge i området 0,3 til 0,6 sekund. Et auditorium kan have brug for lidt mere klang til at bære stemmen igennem, men ikke så lang at forståeligheden mudrer til. Den længste efterklangstid man i praksis har brug for er omkring 2,2 sekunder i koncertsale til symfonisk musik.

Hjemme i stuen, hvor surround anlægget står, vil det være hensigtsmæssigt at sigte efter en efterklangstid omkring 0,5 sekund. Dog er tidens møbleringsskik måske lidt for sparsom til, at man når helt derned.

Når man har styr på hvilken efterklangstid man har behov for, er det vigtigt at vide, at efterklangstiden kan variere med frekvensen. Dvs. efterklangstiden kan være én i bassen, men en anden i diskanten. Det gælder om at opnå samme efterklangstid i hele frekvensområdet.

Absorption

Som udgangspunkt har store rum længere efterklangstid. Men det der i den sidste ende styrer efterklangstiden er lydabsorptionen. Hvis man forestiller sig et rum med et åbent vindue ud til det fri, så vil lyd der rammer denne åbning forlade rummet og aldrig vende tilbage. Dette areal kan da betragtes som fuldt absorberende. Det kan også udtrykkes, at det har en absorption på 100% eller en absorptionskoefficient på 1.

Nu forholder det sig sådan, at der i praksis stort set ikke findes materialer, der absorberer lige så godt som det åbne vindue. De fleste absorbenter, det vil sige materialer og bygningsdele med en lydabsorberende virkning, kan være gode til at absorbere i en del af frekvensområdet og mindre absorberende i en anden.

Det medfører, at man har et puslespil, hvor det gælder om at sikre, at materialerne i rummet tilsammen har lige meget absorberende virkning ved alle frekvenser. Ellers bliver det "mærkeligt" at opholde sig i rummet.

Absorbenter

De fleste opfatter materialer som mineraluld, tæpper og æggebakker(!) som lydabsorberende. Det er også rigtigt. Men afhængigt af opsætningen vil disse såkaldte porøse absorbenter kun absorbere i diskanten. Et godt akustikloft, nedhængt 30 50 cm vil dog virke pænt ned i frekvensområdet, dog uden at være en egentlig basabsorbent. Generelt er det sådan, at jo tykkere absorbent (eller jo større afstand fra den fast bagvedliggende flade), des lavere frekvenser absorberes.

I bassen har man som regel behov for membranabsorbenter. Det er materialer som gipsplader (gipsvægge og lofter), trægulve på strøer, evt. store vinduer. Dvs., at hvis man starter i et råt betonrum (beton har ingen absorption), kan man være sikker på, at der skal anvendes så mange membranabsorbenter man kan komme afsted med. Derefter kan man tænke på resten af frekvensområdet.

En tredje form for absorbenter er resonansabsorbenterne. Det er sådan noget som perforerede plader, spaltepaneler mv. Deres typiske absorption ligger i mellemtonen.

De egentlige akustikmaterialer er specificerede mht. absorptionskoefficienter mv. Til gengæld skal man søge akustiklitteraturen, hvis man ønsker at kende andre materialers akustiske egenskaber.

Hvad skal der til?

Har man allerede et rum, hvor man kunne tænke sig at gøre noget ved akustikken, kan man iføre sig de akustiske briller. En akustiker er normalt kendetegnet ved at klappe i hænderne (ét klap), når vedkommende kommer ind i et rum. På den måde konstaterer man hvordan efterklangen er og om der skulle forekomme såkaldt flutter ekko. Dvs. et ekko, der nærmest optræder som en snerrende tone. Det er mellem hårde parallelle flader, at den slags forekommer.

Hvis man så går rundt og banker på vægge, gulv og loft, så kan man få et indtryk af om der er noget, der kan absorbere i bassen. For en ordens skyld skal det lige nævnes, at akustikeren naturligvis også har måleudstyr til konstatering af de akustiske forhold.

Bryd parallelliteten

Det kan være en god ide at bryde parallelliteten af modstående flader, f.eks. 5 grader. Herved reduceres muligheden for det der kaldes stående bølger, dvs. lave frekvenser der særligt fremhæves fordi rummet er afstemt nærmest som en guitarstreng. Dette kan ved samme lejlighed fjerne flutterekkoer.

Derefter er det vigtigt at vide, at den absorption man tilfører gerne skal være fordelt på fladerne. Hvis man f.eks. har et akustikloft og et gulvtæppe, men ikke noget på væggene, kan der i princippet være lyd, der kører frem og tilbage mellem vægfladerne uden at møde absorptionen på gulvet eller i loftet.

Hvis man laver lytte , demo eller kontrolrum, er der i øvrigt behov for, at rummet er rimeligt højre/venstre symmetrisk. Ellers vil højttalerne i den ene side ikke lyde på samme måde som højttalerne i den anden side.

Styr dine refleksioner

For at hjælpe lyden hen til de absorberende flader, er det godt at have diffuserende elementer i rummet. Gerne så store som muligt. Mikserpulte, stole, sofaer, reoler mv. er gavnlige, hvis de bare ikke reflekterer lyden direkte hen til lytteren, men mere tilfældigt ud i rummet.

Refleksionerne i rummet skal så vidt muligt dæmpes eller "styres væk". Det gælder især det der kaldes 1. ordens refleksioner. Det er reflekteret lyd, der fra lydkilde til lytter har ramt en enkelt flade. Den slags refleksioner vil farve lyden på uheldig vis.

Højttalere i rummet

Når højttalerne så sættes ind i rummet bør man undgå, at disse befinder sig 50 100 cm fra begrænsningsfladerne. På den måde begrænser man udfasninger, dvs. mangel af lyd i især frekvensområdet 80 120 Hz. Det er nemlig et frekvensområde, som det er ekstra vigtigt at man har styr på. Hvis man har en subwoofer, må man flytte rundt med den for at finde den optimale position.

Husk

Det akustisk gode rum er en forudsætning for al akustisk lydproduktion og for al kontrol af lydmateriale uanset hvor digital lyden end måtte være.

Eddy Bøgh Brixen

Karma-Audio.dk

 

EBUs anbefaling til efterklangstidens variation med frekvensen i lytte /kontrolrum (Ref: EBU Tech 3276 og EBU Tech 3276)

Akustikkens indflydelse på frekvensgangen (20 Hz 150 Hz) for en subwoofer, der flyttes rundt i forskellige positioner.

Absorptionskoefficienterne for en "berømt" absorbent: Æggebakken.

 Man kan se, at den er ganske effektiv ved høje frekvenser, men at den ingen virkning har ved lave frekvenser. Målingen er foretaget af Lydteknisk Institut.

Karma-Audio.dk

Bas-absorbent og diffusor.

http://www.dagogo.com/View-Article.asp?hArticle=919

In 1950 the chief engineer Harry Olsen in his classic technical book, Acoustical Engineering.  (FIG 1)

Found in this book are untold numbers of RCA audio/acoustic lab secrets.

Still, it took another 30 years and a patent search before I discovered the functional bass trap and the great book where it was disclosed.

Frankly, I was relieved to discover my work fell right in line with and was a natural extension of earlier work in this same area. 

Even the same peak efficiency of 140% reported by Harry Olsen for his functional bass trap is a standard measurement of the TubeTrap product line.

Karma-Audio.dk

Acoustics - Basic room treatment

For a HiFi Listening Room for a Home Theater click here

Add Half-Rounds to the back wall.

Again, stacked Half-Rounds are better than single Half-Rounds.

The TubeTrap's chrome dots should point to the listening chair. Rotate Tubes to increase treble absorption.

Add Full-Round TubeTraps to all four corners, start at the speaker end of room.

Stacking two Full-Rounds on top of one another in a corner is better than a single Full-Round in a corner.

The TubeTrap's chrome dots should point to the listening chair. Rotate Tubes to increase treble absorption.

Step 2

Add Wall Panels,Diffuser Panels, or a combination of both to the side wall.

Diffusers turn hard reflections into ambiance. Panels absorb hard reflections.

This removes ghost images in the horizontal plane, making images more focused.

Step 3

Add Half-Rounds to the front wall..

A pair of stacked Half-Rounds are better than a single Half-Round.

Step 4

Larger speakers will benefit from using Half- Rounds instead. Their additional bass absorbing capability will enhance bass dynamics.

This improves dynamics, unmasks midrange/treble detail, and reduces “room boom”. Launching a clean wavefront at the front of the room, and cleaning up the excess energy at the rear of the room make the most dramatic improvements.

This deepens the sound stage, by reducing ghost images from the front wall. Treblereflectors maintain room ambiance.

Step 1

Capturing bass energy at the rear of the room allows the direct bass sounds to have more impact, free from masking, cancellations, and room modes.

Karma-Audio.dk

Home Theater Acoustics

Volume Three

The proper placement of subwoofers in your home theater system is crucial to the quality of the desired sound. Placing them in the correct location creates a bass sound level smooth with frequency.

BY ARTHUR NOXON

The subwoofer generates very low frequency sounds. The size of these sound waves compares to the size of the listening room. If the subwoofer is placed in the wrong position in the room, we hear "room booms" instead of the musical bass scale. On the other hand, if we get the subs into the proper location, the bass sound level becomes smooth with frequency. Subwoofer extension into deep bass is achieved along with significant punch capacity. In this section of work, we will study both the good and bad placement positions for subwoofers located in smaller sized listening rooms, the kind most of us have. Bad speaker positions are those that allow the speaker to stimulate room resonance (modes). Good speaker positions are those from which the speaker cannot stimulate such "room boom" effects. These golden spots are called the anti-mode speaker positions.

If the woofer is positioned at one end of the big pipe and a frequency sweep is delivered to it while a sound meter is positioned at the opposite end of the pipe, we will see evidence of the modes. At first, in the very low frequency (LF) range, there are no special changes in the sound level meter. Sooner or later, there will be some frequency where the meter needle gets pegged. The sound got exceedingly loud at this opposite end of the tube, marking the first or "fundamental" resonant frequency and mode.

As the frequency sweep continues upwards, the meter level drops back to normal for a while, but finally peaks again. This next frequency marks the second resonance mode and is called the first partial or first harmonic. Curiously, the frequency of this second resonance is exactly twice that of the first resonance. We go up some more, only to find another resonance, the third resonance or second partial which is exactly three times the fundamental resonance frequency. This harmonic. Series goes on and on with this same pattern.

Needless to say, if we moved the speaker to the opposite end of the pipe, exactly the same harmonic series would be developed. However, if the speaker were moved to the exact middle of the pipe, the first resonance would not sound out. Nor would the third resonance, the fifth, and so on. Odd numbered resonances cannot be stimulated in a closed pipe when the speaker is located in the middle of the pipe. From the middle of the pipe the speaker can only stimulate half of the total number of resonances available to the pipe, the even numbered resonances.

RESONANT MODES

To gain some understanding of mode vs. anti-mode speaker positions, it will be very helpful to consider a one-dimensional acoustic space. In a regular room, sound can travel in any direction. If, however, the speaker was located at the end of a long, narrow pipe, the sound could only travel in one direction, along the axis of the pipe. A pipe is a one-dimensional acoustic space. If we plug up both ends of the long pipe, then the "boundary conditions" of a one-dimensional room are met. This is a similar idea to a room having walls.

This position dependent selectivity does not stop with the ends or middle of the pipe. Move the speaker to a position one third from either end or, presto, only the third, sixth, ninth, and so on harmonics can be stimulated. Then we move to a position one quarter of the pipe length from either end and are not surprised to find only the fourth, eighth, twelfth, and so on harmonics. And next the fifth ... and so on.

The reason for harmonic selectivity is not in magic numbers, or any other form of audio voodoo. It's more like simple physics, otherwise known as the nature of things. A play set swing can provide a good example for this effect. As children, most of us learned to "pump the swing" by coordinating our leg/body action with the position, more accurately, the phase of the swing's position. It's all in the timing and it is pretty hard to explain, so we teach by showing. Monkey see, monkey do. If we can get the timing right, up we go, almost like magic.

The swing system is a resonant system and a pipe filled with air is also a resonant system. Applying the right kind of force at the right place and time can pump either up. In a closed pipe, which has been stimulated into its first resonance condition, we will find that the sound is very loud at either end of the pipe and very quiet at the halfway point, the middle. These loud areas are called sound "pressure zones"; and, if the speaker is located in either of these pressure zones, it efficiently couples to and can pump up the resonant condition. Conversely, if it is not so located, it can't pump.

The second harmonic of a closed pipe has three pressure zones, one at either end and one in the middle. If we located the speaker in any three of these pressure zones, we can stimulate the second harmonic. However, if we locate the speaker in the middle pressure zone, we cannot stimulate the first resonance but we can still stimulate the second one. Once the understanding of these variables has been made clear, it becomes easy to expect what will happen if a speaker is located in any particular location.

It seems that no matter where a speaker might be located in a closed pipe, one resonant harmonic series or another will become stimulated. However, subwoofers are always rolled off just below the beginning of the vocal range, about 85 Hz. This means that the subwoofer cannot stimulate resonances above the roll off frequency. Now, if the first resonance is 25 Hz, the second will be 50 Hz, and the third 75 Hz. The fourth resonance will be at 100 Hz. The fourth resonance and all of those higher than it are above the 85 Hz roll off frequency of the subwoofer. This means that the speaker need only be positioned so that it doesn't stimulate the first, second, or third resonances. The speaker has to be located somewhere, but not at either end, not at the middle, and definitely not at the third waypoints.

For our example room, the distance off the end wall had to be five feet. The distance off the side wall could also have been set up at five feet. We could have had two, ten-foot round trip waves impacting the speaker with a time delay of 10/1130 = 1/113 second. This would create the self-cancel effect to occur for a frequency whose period is twice that time or 2 x 1/113 = 1/56 second. This would be the frequency of 56 Hz which is well below the 85 Hz roll off frequency of the subwoofer. A better choice for the subwoofer position might be 2-1/5 feet up, 3-3/4 feet out from the side wall, and five feet off the end wall.

A graph can be used to help with this latest decision whenever there is a range of speaker positions available. For any axis in which the third harmonic is engaged, the speaker position is fixed at 25 percent. There is flexibility in speaker position for any axis that only engages the first or second harmonic. Outside of keeping the three dimensions as different, as far apart as possible, there is one other detail. We need to keep the bicorner bounces from overlapping the wall bounces. The only opportunity for trouble here is if the distance to the corner formed by the two shorter dimensions equals the third longer dimension.

To use the timing graph provided here, you darken the arcs whose radius equals the fixed, third harmonic, 25 percent dimensions. Then you darken the straight lines that correspond to the ranges available in speaker placement for the other, lower harmonic axis.

For our example, the room length engaged the third harmonic and the distance off the back wall became fixed at five feet. An arc with a five-foot radius is darkened on the graph. The width and height of the room were not long enough to engage the third harmonic. The corresponding ranges for speaker placement are plotted on the graph, one axis for each graph axis. It doesn't matter which room axis goes on which graph axis. Here, the side wall was placed on the vertical axis and the height range was placed on the horizontal axis.

0 35 50 67 100%

There is another factor that limits the remaining options for speaker placement. The pressure zone is not a pinpoint-sized space; it spreads out. If the speaker is located near enough to the center of the pressure zone, the resonance can still be stimulated. A pressure zone effectively extends about one quarter of the distance between adjacent pressure zones and the speaker should not be located inside the effective pressure zone space. For all practical purposes, the speaker should be located 25 percent away from the end of the pipe to best avoid stimulating any of its first three harmonics. There is no location towards the middle of the-pipe that suits a subwoofer position, as the pressure zones there are overlapping.

A listening room can be approximated as if composed of three intersecting pipes. These pipes would lie along the three room axes -- front to back, side to side, and floor to ceiling. This means that the subwoofer location for best, non-resonant playback will be about one-quarter of the ceiling height off the floor, one-quarter the width of the room off the side walls, and one-quarter the room length off the front or back wall. When discussing speaker location, it is only the dimensions to the center of the driver cone that count. The location of the edge of the box really doesn't matter.

No computer program is needed to properly position the subwoofer in a room; a tape measure is your only investment. Note also that the currently popular "rule of thirds" placement formula is not consistent with the understanding of an aresonant speaker placement. This over publicized "rule of thirds" would only be applicable if the subwoofer roll off was set so that the speaker did not play the third harmonic.

The concepts of subwoofer placement have by now been well developed and now some practical applications can be considered. Two things need to be shown - the roll off frequency of the subwoofer and the first resonance frequency of each pipe axis of the room. Typical roll off is set at 85 Hz.

The shortest dimension of a room is the floor to ceiling distance. If this dimension is eight feet, the first vertical resonance occurs at: 1130/2x8 = 70.6 Hz. The second at 141 Hz is well above roll off and can be ignored as well as any higher partials. The vertical position range for aharmonic playback will be to locate the subwoofer anywhere in the middle half of the room, keeping it at least two feet away from either he floor or ceiling.

The next shortest distance in a room is the width, typically about 15 feet. The first resonance for this is 1130/2x15 = 37.7 Hz. The second is twice that at 75.4 Hz and the third is three times that or 113.1 Hz. The second harmonic is within the subwoofer range but not the third. The sub has to be placed more than 25 percent away from the wall because of the first harmonic, but not in the central one-eighth width of the room due to the second harmonic. The sub can be located anywhere between three-quarters and 6-3/4 feet from the side wall. Lastly, the length of a room might easily be 21 feet long. The first resonance for this would be 1130/2x21 = 26.9 Hz. The second is 53.8 Hz and the third is 80.7 Hz. The fourth at 107.6 Hz md above are all well above the roll off frequency and can be ignored. For the length of the room, the sub position should be one-quarter of the room length or five feet off either end wall.

So, a room 8 feet by 15 feet by 20 feet will have the smoothest bass if the piston of the subwoofer is located two to six feet off the floor, between 3-3/4 and 6-3/4 feet off the side walls, and five feet off the end wall. This is true as far as avoiding strong coupling of the speaker to the room modes, but there is more than modes to worry about as far as speaker smoothness is concerned.

SOUND CANCELLATION

Incidentally, these silent areas located between the pressure zones deserve a little attention as well. They are "cancel zones" because sound is cancelled at these locations. Sound cancellation is being used a little more often these days, particularly with industrial noise control applications. Sound cancellation seems to possess a form of sci-fi lure for some people. The idea of beaming "anti-sound" waves to quiet freeway noise is one of the more popular of these energy-out-of-water type schemes. To the literal reader, words create reality. But to the engineer and scientist, reality exists independently from words. Just because someone can dream up a sentence that seems to make sense doesn't mean that it physically does make sense.

Normally, sound cancellation applications remain limited to the control of sound in pipes. For example, if we take a closed pipe that contains a harmonic condition and drill a hole into the pipe, we will get varied results, which depend on where the hole is located. For the first harmonic, with a pressure zone at either end and a cancel zone at the middle, we can drill a hole into the pressure zone at either end and kill the resonance. But, if we drill through the wall of the cancel zone, there is absolutely no change in the resonant condition. A hole in either pressure zone allows pressure energy to leak out. But there is no pressure energy in a cancel zone, so a hole that leaks pressure doesn't affect anything.

This is not news -- the ancients knew about it. The flute and clarinet type instruments use this open/closed hole effect to select pipe resonances, heard by us as notes. Let's consider what can happen if the closed pipe is engaged with its second harmonic. There are three pressure zones and two cancel zones. A hole could be drilled through the pipe wall at each cancel zone and not affect the existence of the resonance. Now we have made a closed pipe into an open pipe; and, if we blow air into one hole, it will come out the other hole. We have discovered a pathway to conduct air through a pipe filled with sound without having any of the sound leak out.

With industrial sound canceling, the tonal sounds of a blower that moves air in a closed duct can be cancelled at an air outlet. One can use either this standing wave pipe process or a speaker/microphone/computer system to create this same sound canceling effect at the opening of the pipe. Although the sound at the opening can be cancelled, the sound elsewhere in the pipe is very loud. If two forces are applied equal and opposite, there is no force imbalance, hence no movement. That doesn't, however, mean there is no stress on the material. There is twice as much stress to the material than if only one force was applied.

So it is with sound. If two sounds are applied equal and opposite, there is no sound at some point, but that doesn't mean there is no stress on the material. There is, in fact, twice as much stress in the material than if only one sound had been applied. If we move away from the point where there is no sound, we'll find twice as much sound everywhere else. That's the point. Sound cancellation doesn't mean sound energy cancellation. The energy is still there. In fact, it has become twice as strong. Just because we can't hear it at one location only means we will hear it twice as loud at another.

This brings to mind freeway noise cancellation and many other sound cancellation schemes. The real rule for sound cancellation engineering is that if we arrange to not hear sound in one place, then to someone else it has become twice as loud. We always have to watch out where that loud zone has become located. If it is onto our neighbor's property, we might get sued. Sound is energy. We also know that energy plus energy equals more energy, not less. We can steer sound around somewhat by adding more sound, but we can't simply erase sound with "anti-sound" waves. Except, of course, in the imaginations of those who read and write sci-fi stories.

Under certain conditions, a speaker can cancel its own sound. Consider what happens when a positive part of a sound wave meets a negative part of the same sound wave. We have sound cancellation. When a speaker is near a wall, sound from the speaker expands out from the speaker, impacts the wall, and rebounds back toward the speaker. At some certain frequency, the timing of the rebound wave will be exactly one-half a period of the tone.

The period of a wave is exactly the time it takes for one cycle to occur. Middle C of the musical scale has the frequency of 256 Hz. That means the period takes 1/256 second to occur. A half period for 256 Hz would be 1/512 second. If sound could go from the speaker to a reflecting surface and back to the speaker in 1/512 second, the positive of the reflected wave would mix with the half-period-later negative of the wave at the speaker face and there would occur sound cancellation. The round trip distance covered would have to be 1130xl/512 =2.2 feet. A wall located 1.1 feet away from the speaker could reflect sound back to the speaker and create this self-cancellation effect.

A single bounce is bad enough, sometimes creating a three to four dB reduction speaker output at and around the self- cancel frequency. But to have two walls reflecting waves back to the speaker at the same time is nearly intolerable. Whenever we have a speaker near a corner, there results three wall reflections, three corner reflections, and one tricorner reflection. In order to keep the self-cance11ing effect to a minimum, every one of these round trip distances should be as different from one another as possible. The most obvious setup is to keep the distances the three walls as different as possible.

ANTI-MODE, ANTI-CANCEL SUB SETUP

The result is a rectangle with an arc passing through the lower corner. The distance off the floor and side wall can take any pair of values inside the rectangle, except those on or close to the arc. They also shouldn't be equal to each other, so the pair of values needs to stay away from the "equal" line on the graph. There is another consideration. Subwoofers sound weaker when played out in the open and stronger when played near sound reflecting surfaces. This wall or floor loading effect is a form of horn loading which always makes low-frequency speakers more efficient. In addition, we elected to keep the sub as close to the side wall as possible, out of t, he middle of the room. The coordinates of 2 to 2-1/2 foot height and 3-3/4 off the wall meets all of our requirements.

Subwoofer setup is usually accomplished by listening to music, inching the box around the room, and trying to find the smoothest location. This sport is more like fishing than anything else, to be specific, bass fishing. What we have tried to do here is debunk some of the practices of audio voodoo, reduce your dependency on the audio personality or guru, limit your searching for magic numbers, and the purchase of guru computer programs. We have tried to replace them with simple graphs, the otherwise desperate and often misdirected groping for that elusive, but real, subwoofer sweet spot.


Home Theater Acoustics

Volume Four

As we survey audio systems for home theater, a trend appears. We consistently find one or two subwoofers, two or three stage speakers, and two ambience speakers. In the last two sections, we studied the subwoofer as it fits into and plays the listening room. Here we will study ambience speakers, the kind that are becoming a standard for home theater.

The Dolby surround signal is a mono signal usually fed to two speakers located towards the back of the room. This signal is unique in audio because it is rolled off at 100 Hz. This doesn't mean that there should be no deep bass in the ambience effect. It does mean that the deep bass is generally understood to have no directionality. Our head is too small, our ears too close together, and our hearing too insensitive to be able to tell which direction low frequency sounds are coming from. Remember how no one worries where the subwoofer is placed, except for visual or room mode control? That's because we can't tell where the bass is really coming from. The way we "know" where bass comes from is by focusing on where the upper partials of the bass sound are coming from. The Dolby surround signal contains the upper partials of the ambience bass, so we think the ambience bass is coming from the ambience speakers. But really, it's the subs and main speakers that get the signal and do the generating of the ambience deep bass.

A good demonstration on the directionality of a speaker can be achieved by setting a small loudspeaker outside of the house on a table that is placed in the middle of the open yard. Then, while keeping some fixed distance away, walk all the way around the speaker while it is playing some tune with which you are familiar. You will hear the full range of sounds of your speaker when you are in front of the speaker, but as you move to the sides, and especially when behind the speaker, the highs drop off substantially, but not the lows. Male vocals, for example, sound pretty much the same no matter where you are, but sibilance, the "tsss" sounds, dramatically drop off behind the box.

If you get an identical second speaker, wire them up in phase and place them back to back. You'll hear bass range everywhere and the sibilance will be heard in two beams, One forward and the other opposite. Listening directly off to the side of the speaker pair, you'll hear the midrange and bass. Now reverse the phase of the two speakers and listen. All of the bass drops out, yet the two mid/high back to back beams remain. To the side, there is a strange drop in all sound. So it is with the dipole speaker. The dipole effect is limited to the upper ranges of the speaker because the bass shorts out, acoustically speaking. At some low frequency, the dipole speaker simply sloshes air back and forth around the edges of the speaker and makes no more sound. This is nc different than listening to a bare speaker and then mounting it onto a piece of plywood. We increase the distance between the front of the driver and the back and, in doing so, give the speaker more range in the bottom end.

Because the surround dipole speakers are fairly small, they short out at fairly high frequency, around 400 Hz. And, so, there must be another system in place to generate sounds below this natural dipole cutoff. There are a number of ways to accomplish this. The most straightforward way is to use a single lower frequency driver reversed, large-sized, and directional midrange drivers. Offset or coaxial tweeters will accompany these large midrange drivers to get full high frequency range. The main thing to keep in mind during the evolution of this style speaker is that the orientation of the low frequency drivers is irrelevant as to the directionality of the lower registers. Omni is omni and it doesn't matter which direction the midbass speaker(s) points.

There seems to be only a couple of rules to follow when placing the surround dipole speakers. Mainly, they have to be placed high on the side walls, directly to either side of the listener position. They can be positioned in front or behind the listener somewhat, but must be angled so that the side of the speaker points to the listener. Above all, never place them in bookshelves no matter how convenient it may seem. The honky, tonal resonances this setup produces will be almost unbearable, not to mention that the walls of the bookshelf will catch the ambience signal before it gets to the room. These surround speakers are to fire along the side wall towards the front and back walls. Next, there are three factors to be considered in the placement of ambience speakers -- resonance, self-canceling, and flutter.

Whenever a speaker is placed in a room, it needs to be positioned so as to minimally stimulate room induced coloration effects. This is especially true for ambience speakers because their effects are in direct competition with the room's natural ambience for the listener's attention. If the ambience speakers are located improperly, they will strongly stimulate the local room effects and their capability of generating the desired audio track ambience will be reduced by the sound masking effects of the room's acoustics.

We know the ambience speaker is to be located high on the side wall by the listener. Beyond that, we seem to be left to our own resources. The lower frequency play of the speaker can be used to determine the most neutral vertical location on the side wall. The high frequency characteristics of the speaker can be used to determine the most neutral front-to-back position for the speaker. In the following sections, we go over the details that determine the most neutral position for the ambience speaker.

ANTI-RESONANCE AND SELF-CANCELING

In the previous chapter, we studied how to determine the most neutral position for the placement of subwoofers in the listening room. Two factors came up to impact the coloration of the sound quality. The first and most familiar was room resonances. We determined that placing the speaker so as to least stimulate the room resonances would be most appropriate. In addition, there is the complication due to placing a speaker near a wall, floor, or corner - a self-canceling effect. The nearby reflection actually weakens the strength of the speaker at a certain frequency.

These lessons also apply to the ambience speaker positioning. The ambience speaker is essentially a single, mid-bass driver with two reversed phase, mid/hi drivers, back-to-back. The vertical position of the speaker on the side wall is determined by the speaker's low frequency coupling to the floor/ceiling parallel surface system. We saw that when the frequency range of the speaker spans many resonances, the best location for the speaker is at the 25 percent mark from one end. However, for the ambience speaker, it is rolled off at least 100 Hz or higher. This means the first floor to ceiling resonance, typically at about 70 Hz for an eight-foot room height, cannot be stimulated. By studying the pressure distribution for the first three resonances and ignoring the first one, we see that the minimum position for stimulation of the second and third resonances lies 20 percent from one end of the dimension. This means the best, anti-resonant location will be a distance down from the ceiling that measures about 20 percent of eight feet or 1.6 feet (19 1/4 inches) down from the ceiling or up from the floor.

For the ambience or surround speakers that are mounted high up the side walls, the 20 percent down position is easy. However, for those ambience speakers that are on speaker stands, putting the speaker 19 inches off the floor is not a normal thing to do. Most speaker stands are set up to position the speaker about ear height, 42 inches off the floor. There is another position, not nearly as good as the 20 percent position, but at least it is a relatively minimal position. This is at the 40 percent point, where the first and second harmonic curves cross just below the 50 percent point. The traditional speaker stand positioning of 42 inches places the speaker at the 44 percent height point for an eight-foot high room. It is not easy to change the height of a metal or even a wooden speaker stand, nonetheless ... we are at this time concerning ourselves with good acoustics, not convenience.

Every time the speaker is located near a reflecting surface, the problem of self-canceling comes up. For a speaker mounted 20 percent down from the ceiling, the self-canceling frequency occurs at a wavelength that equals four times that distance or 80 percent of the room height. The wavelength that goes with an eight-foot high room will be about 6.4 feet, which corresponds to 1130/6.4 or 177 Hz. By the way, there will be reinforcement at twice that self-cancel frequency at 354 Hz and then a cancel at 530 Hz, and so on. Every 177 Hz there is a self-induced effect that alternates between cancel and boost. This is on the order of a four to six dB magnitude and stops only when the speaker becomes so directional that it doesn't illuminate the reflecting surface, typically about 600 to 700 Hz.

It is very easy to remedy this self-canceling problem. Simply, bass trap the bounce back point. But not just any bass trap will do. The low frequency cut off for the bass trap should be set about a half octave below the lowest frequency that needs to be trapped. For 177 Hz, this is figured as follows: A full octave below 177 is 88 Hz, so a half octave below is half of 88 or 44 Hz. The half octave below 177 Hz is 177-44 or 133 Hz. Now that you know all about it, the simple formula is that the lower half octave point is 75 percent of the given frequency.

The floor standing ambience speakers seem to luck out as far as self-canceling effects go. Their drivers will be 39 to 42 inches off the floor and self-cancel at four times those distances, for the 15- and 14-foot wavelengths. The frequencies for these are 87 and 80 Hz and both are well under the 100 Hz cutoff for the Dolby ambience signal. So these high mounted, floor standing speakers do not self-cancel off the floor. But floor standing speakers tend to be set up away from the wall. While the floor bounce may be too far to self-cancel, the nearby wall bounce can be a problem. We know the omni speaker is rolled off at about 400 Hz. The 1/4 wavelength dimension for this is 8 1/2 inches, which becomes the maximum distance this driver should be away from the wall and not self- cancel from the wall bounce.

Why, one might ask, should we be careful of the range of the bass trap we use? Also, who needs a "bass trap" anyway? Don't acoustical foam or wall panel type products absorb sound and at a lot less cost? The questions are proper to ask and deserve an explanation. They all involve the balancing of frequency characteristics, those of the speaker to those of the absorber.

A speaker loves to be near a corner when reaching for its lowest registers. The "horn loading" effect due to placing a speaker near a wall, floor, or corner increases the efficiency of the speaker in the bottom end, more bass power at no extra cost. If a bass trap is placed in the corner, we usually do not want it absorbing the deep bass. We want the opposite, horn loading to reinforce the deep bass. For this reason, we need the bass trap to roll off its absorption in the range where the speaker output is also rolling off and the benefits of horn loading are being called into action. For small, full range boxes, this 3 dB down point (50 percent power) can typically be about 60 Hz. But as mentioned above for the home theater ambience speakers, the roll off is set at about 100 Hz or more.

Now we'll move onto acoustical foam and wall panels. These fairly common acoustical products are good only for the midrange and high frequency ranges. This range includes only the top three octaves of the piano keyboard and does not include anything in the lower 4 1/2 octaves of the keyboard. Only bass traps can cover this lower range of sounds. The middle of the keyboard is C4 at 256 Hz. In our example, we needed the absorption half power point to be at 133 Hz and that's almost one full octave below middle C. It also is two full octaves below the roll off point of commodity foam and wall panels. Bass traps are the only absorptive devices that can correct acoustical problems.in the lower 60 percent of the piano keyboard.

CONCLUSION

Ambience speakers, like all others, engage the room acoustics. Because of their limited bandwidth, they do not couple to the lower resonances of the room. That gives us the most neutral, anti-resonant position yet for the speaker position, 20 percent off the floor or down from the ceiling. Something new has been added to help smooth out the acoustic space for the speaker - the bass trap - the self-canceling bounce back point. The best ambience sound is colorless, except for the ever changing signatures in the ambience track.

Karma-Audio.dk

Home Theater Acoustics

Volume Five

There are two basic kinds of ambience speakers these days, although more may pop up as time goes on. The first, most basic type are simply small, book shelf type speakers on speaker stands or mounted ot the wall. The ambience signal can be beamed either: directly at you or away from.you/bouncing around the room a bit before it hits you. If the speaker is aimed directly at you, you will hear it and know where it is. Our hearing is very sensitive to sounds beaming directly into one ear. After all, what do we do whet we can barely hear some sound? We turn our head to the side, so one of our ears can hear the sound more directly. For the ambience speaker setup, the orientation of the speaker is a matter of personal choice and the experiments should be made. Many people prefer not to hear the ambience signal directly and their ambience speakers are turned somewhat towards the wall and face either forwards or backwards.

The second type of ambience speaker is called a surround speaker and is recommended by the THX people. In this system, the choice about how we hear the ambience signal has been made for us. This speaker is mounted high on the side wall and set up to not beam any sound directly at the listener. These speakers are specified to be primarily dipole type speakers. This means that they play backwards and forwards equally strong, but not at all to their side, which is, of course, where the listener is located.

The dipole speaker familiar to us in hi-fi is usually a thin sheet of material that is forced back and forth by either magnetic or electric fields. The forward wave is exactly out of phase from the backwards wave. When the sheet moves forward, a positive pressure wave is sent forward while a negative pressure wave is sent backwards. Not so for most surround sound speakers. This type is often comprised of two dynamic speakers wired out of phase and playing back to back. There still is a positive wave sent out in one direction, while a negative wave is sent in the other, being equal in strength but opposite in phase. There are numerous dipole speakers and the goal here is not to propose or evaluate which might be better than the other, if such would even be possible. The goal here is to explore the effect on the sound of these speakers that is imposed by the room in which they are located.

The dipole type surround speaker is a strange kind of speaker to the world of audio and it will, without a doubt, undergo a number of transformations as it evolves into its mature form. To begin with, it is not a full range speaker because the surround channel is rolled off at 100 Hz. For the most part, these speakers nave been a small speaker cabinet with two speaker baffle boards, one set to face forward and the other to face backwards. Usually, we see each panel forward and the other facing backwards. Usually, we see each panel mounted with a single driver. Two-way speakers are also used, sometimes with the tweeter offset from the main driver, other times with coaxial drivers.

The intent of this style of speaker is to "play forwards and backwards" so as to illuminate first the room and not first the listener. This directional effect only works for a limited frequency range of the speaker. Small-sized drivers are directional for upper, mid, and high frequency ranges, but become omni-directional for the lower ranges. This directionality effect occurs at a predictable frequency based on the size of the drivers, as well as the cabinet in which they are mounted.

In hi-fi, home theater, and even most recording studios, the parallel wall surfaces are within the range of 15 to 30 feet apart. That means we don't hear flutter echoes but do hear the flutter tones. Flutter tones are sounds that have a low-frequency character, but they are not to be confused with room modes which also are low frequency in nature. The control of the low frequency flutter tones, as we will soon see, is accomplished with high-frequency type diffusion or absorption. Of course, control of the low frequency of room modes is accomplished only by means of larger-sized bass traps, usually best located in the corners.

The low-frequency flutter tone is a pseudotone - a trick on our hearing system played by the rapid staccato of high-frequency noise pulses. Sometimes a careful listener can become confused as to how a seemingly low-frequency sound can be eliminated by the introduction of a paper thin reflector or fabric, especially when common sense leads us to expect that only those large-sized bass traps should have been needed. In order to eliminate the detection of a flutter echo pseudotone, we need only to break up the flutter echo process. It takes very little scattering or absorption of high-frequency sounds to break up the flutter echo sequence, and thereby el.iminate the accompanying impression of the low-frequency sounds of the flutter tone.

Audio parlor tricks, such as making bass reverberation disappear with nothing more than a carefully placed scrap of paper, are accomplished with the magician's classic technique, a distraction of words and slight of hand. Only this time, we say that to create the illusion, the hand must be moving faster than the ear. Actually, the clue to the trick will be found in the presentation. The guru claps the hands and says to listen to the low-toned overhang. If you spectral analyze the energy content of a hand clap, you will find no energy below 400 Hz, yet the hand clap generates the perception of typically a 50 Hz sound. It's a great trick. Practice it and amaze your friends with your superpowers. You could even start up your own business, selling little tinfoil "bass traps" and you'll probably even get away with it, for awhile.

Another important position to stand at is the end of the hall. We already know the flutter echo occurs at half the rate as when we stood in the middle of the hall. But let's look at the pulse timing detail. Again, two pulses expand from the clapper's position, one heads toward the far end wall and the other toward the near end wall. The first reflection, off the near end wall, hits us after an overall travel of only three or four feet. It races by and follows the other pulse down the hall, lagging by six to eight feet. They both hit the far end wall and return towards the clapper's position. The leading pulse flashes by and on to hit the nearby end wall. By the time it again hits the clapper, the lagging pulse also hits the clapper. This creates the effect of a single-hitting, double-strength pulse. Then the lagging pulse moves past and towards the nearby end wall. It reflects and, after a bit, again passes by while heading for the far end wall. In the meantime, the leading pulse had already long left the scene, heading again for the far end wall and a repeat of the cycle.

What we have here is a triple pulse event whose timing is that of a full round trip in the hall. The three pulses are so close together that they sound as if they were one pulse. This combining effect is well-known in pro and high-end audio. It is called the Haas effect, after the scientist who did a lot of work in this area of hearing. What he found is that when high-frequency reflections, such as those in the hand clap arrive within ten to 15 ms (thousandths of a second), they fuse together and sound as one.

Next, we take a few steps down the hall and repeat the hand clap test, listening for any changes in the sound of the flutter tone. If we moved five feet off the end wall, the two pulses would be 20 feet apart and heard as separate pulses because they arrived outside the sound fusion time period. However, the same sequence of events still occurs. The only difference is the separation of the two distinct and small pulses. In the middle position, double-strength pulse effect still occurs. As we change positions along the length of the hall, we change the timing of the discrete echoes that make up the flutter tone. We also find that as we approach the middle of the hall, the two single echoes get far away from the double pulse and closer to each other. When they are within about six feet of each other, the fusion effects begin and the two pulses start sounding as if they were one and the upper octave flutter tone is heard. Get just a few feet off dead center of the hall and the upper octave disappears and the lower flutter tone begins to reappear.

The timing of the two separated pulses is what accounts for the changing of the character of the flutter tone. As we move closer to either of the end walls, the timing between the two separate pulses gets closer together, sandwiching the double-strength pulse until the end wall is reached and they are essentially all on top of each other. As we move closer to the center of the hall, the timing between the two separate pulses again gets smaller. This time, they do not sandwich and are as far as possible from the double-strength pulse. Finally, at the center, the time between them goes to zero, creating a second, double-strength pulse.

All the pulses contain energy, the same amount of energy. Whenever they return to the clapping position, together they combine into a stronger, double-strength pulse. Even more, when they arrive at the clapper's position within six feet of each other, they still combine into a single, double-strength pulse. When a clap originates within three feet of an end wall, all of the pulses arrive at effectively the same time and the result is heard as a four-times stronger, low-frequency flutter tone. Then again, if the clapper is within three feet of the middle of the hall, the separated pulses arrive close enough together to combine and double up in strength. Either of these extreme conditions is about as easy to detect.

When the two separated pulses are not close to the doubled-up pulse, the lower flutter tone is quieter, less noticeable to detect and that is good. Also, when the separated pulses are not combined due to a midpoint clap position, the upper octave flutter tone is not heard. That is also good. Clearly, we now know that the most non-stimulating position for flutter tone generation will be more than four feet away from either end wall and a few feet off the center of the room. By experimenting, additional information is developed. Anywhere in the end third of the room seems to strongly stimulate the lower flutter tone. The thirdway point seems to stimulate the third octave, along with the fundamental flutter tone. The middle of the room really generates the second octave flutter tone within a foot or two of the center point.

Using our 20-foot room as an example, the ambience speaker ought to be located ahead of the 1/3 point, but two to three feet off the center. That puts it at about seven to eight feet off either end of the room, probably the rear wall for home theater. As a general rule, the ambience speaker can be placed 38 percent of the room length off the back of the room. This position will ensure that minimal flutter tone coloration is introduced into the room.

This section has been intended to be a baseline guide for the anti-flutter tone positioning of the surround speakers. To this, we next add some enhancement devices to both increase the presence of the ambience signal and to continue to reduce the telltale presence of flutter tones in the home theater setup.

DIFFUSION OF FLUTTER

In addition to positioning the speaker to weakly stimulate the distracting flutter tones, another element of acoustics can be brought into the battle and put to good use. Diffusors are devices or surfaces that scatter sound. The home theater ambience speakers are located high on the sidewalls and directed to illuminate the upper outside areas of the front and back walls. The first idea about scattering sound tends to be directed to these areas. Why not add a curved or otherwise irregular surface to these areas of direct illumination?

As it is, we can hear the flutter tone that comes from the ambience speaker because its multiple reflecting wavefront not only shuttles back and forth between the front and back walls, but the wavefront expands while doing so. What we hear is the expanding edge of the flutter echo circuit. Now if we add diffusion to the end walls, we will certainly reduce the time that the flutter tone is sustained because the diffusors are redirecting some of the flutter energy away from the flutter circuit at each reflection. This redirected energy is not absorbed but scattered more fully into the room. That means that the listener is getting an even stronger flutter tone signal than before. Not only does the listener hear the expanding edge of the flutter echo, but now additionally hears the scattered sound off the diffusor. Ironic as it seems, adding diffusors to the end walls is a trade-off treatment with mixed results. The flutter tone becomes louder but shorter-lived. It is a change, but is it an improvement? Better, worse, or merely different, this now is something for you to decide for yourself.

CONCLUSION

Over the last two sections, the dipole ambience speaker has been shown to best be placed about 38 percent of the room length off the back wall, and 20 percent of the room height down from the ceiling. Located directly above it there needs to be a bass trap good through 100 Hz. Along the upper sidewalls there should be distributed a set of ambience kickers. Attend to these details and the ambience speakers can safely play into your. room without inducing coloration or distracting distortions. Only then can the true shading and hue of the signal on the ambience sound track be heard.

Let's look at another technique. The flutter echo runs back and forth along the length of the room, hugging the upper sidewall/ceiling corner. Sound-scattering devices can be placed along the upper sidewalls of the room. Again, sound is depleted from the flutter echo circuit. As energy from the flutter echo is redirected into the room, the flutter echo lifetime is reduced. However, this time the scattering takes place between the end wall reflections and not in lumped reflections off the end walls.

These deflectors can be slightly angled down so as to not only kick the reflection to the side, but also downwards. After all, the listener is nearer the floor than the ceiling. Such deflectors are sometimes called ambience kickers in the professional world of recording studios. Another aspect in the setup of these kickers is their spacing. Just as the regular timing of end wall reflections manifests itself to us as a flutter tone, regular timing of reflections off the deflectors can also create a flutter tone. Additionally, we don't want to place the deflectors so that their signal arrives at the same time as any of the regular flutter echo signals. In such a case, the work accomplished would be minimally different from that by diffusors on the end walls.

Clearly, we won't want the deflector to be located the same distance towards the front of the room as the distance the ambience speaker is to the rear wall. This would give the same timing to both reflections being received at the listener's position. The side scattering deflector has to either be in front of or behind this position. Since the ambience speaker is located about 38 percent off the back wall, the ambience kicker should avoid the location of 76 percent off the rear wall. As a first guess, we could locate it almost halfway between, about 52 percent off the rear wall. This produces two new reflections spaced out between the timing of the end wall reflections. The strength of these reflections will be similar to the end wall reflections because of the longer distances involved.

Another deflector could be placed about halfway between the ambience speaker and the rear wall. This one will produce a reflection that arrives somewhat before the rear wall reflection and helps to fill in that big time gap. How many other such ambience kickers can be installed is not so easily predicted. The side fill they produce and its value to the listener belong, in a large degree, to the listener's taste and judgement.

The sonic impact produced by upper sidewall diffusors is quite different on two levels. First, the scattering reflections are distributed all around the listener rather than coming from just in front of and behind the listener. This more diffuse "source" of the ambience signal seems to promise to be more supportive and involving for the surround sound effect. Second, is the relief provided due to multiple reflections that crop up in between the end wall reflections. These intermediate reflections spoil the perception of the otherwise clear and distinct end wall reflections. The result is that distributed, upper sidewall deflectors produce a signal that masks out the flutter tone. The result is a lively, diffuse, and colorless ambience signal.

Karma-Audio.dk

A hand clap contains only high frequencies. For a loudspeaker, the high frequencies are directional, forward of the speaker box. To properly administer a hand clap that mimics the high-frequency beaming pattern of a loudspeaker, the hands must meet at waist height while the clapper is facing the same direction that the speaker does. The body of the clapper blocks the expansion of the clap sound backwards. The listener is no longer in the clapper position, the listener is now seated in the listening position. This time, the hand clap is cast forward from the speaker position and is heard by the real listener. It is how the listener hears the speaker that counts and not so much how the speaker sounds to itself, at least in hi-fi playback settings.

In order to properly evaluate the consequence on the listener of the strange sound we heard when standing on the chair and clapping our hands overhead and near the mounting position of the ambience speaker, we must repeat the test while a listener is seated in the listener's chair. True enough, in this case, the zing we hear when we clap is also heard by the listener. And so, is the sound we hear, good, bad, or inconsequential? Certainly this sound effect is distracting and that alone is enough to warrant its eradication. On the other hand, we want to retain an overhead liveliness so as to promote the ambience signal. We can't sacrifice the lively quality of the overhead space in the room, yet we must try to get rid of its distracting effect known as flutter echo.

FLUTTER ECHO/FLUTTER TONES

Before we try to solve our problems, let's spend some time learning about it. When we administer a hand clap test while located between a pair of uncluttered and parallel walls, we hear a flutter echo. It has a "zing" sound. The flutter echo actually does sound like a tone. The frequency of the tone depends upon the timing of the flutter. A flutter echo is how we hear what really is a rapid sequence of noise pulses. When we clap our hands in the outdoors, we simply hear the single, sharp pulse of noise we call the clap sound. If we clap our hands while standing some distance away, yet facing a wall or building, we will hear a single rapport of the clap, its echo. Then, if we relocate and stand between a pair of more nearby and parallel walls, that single pulse reflects back and forth rapidly between the parallel walls and we hear what we call a flutter echo.

If the walls are far apart, some 60 feet or more, we actually hear the flutter sequence of the echo reflections. But if the walls are closer together, the distinct detail of the staccato seems to disappear, but only to be replaced by a new sound, one of tonal quality. If the walls are far apart, say 60 feet, we hear the slap back at a rate of 1130/60 or 17 times a second and it sounds like the tap-tap-tap of a true flutter echo. However, if the walls are closer, say 20 feet apart, we will hear that slap back pulse of sound at a rate of 1130/20 or 57 times per second. When we, the human listeners, hear a click or noise pulsed at 57 times a second, our ears/brains are tricked into perceiving a buzz-like tone of 57 Hz. And so, the flutter echo we hear when the walls are farther apart becomes a zing-sounding flutter tone when the walls are closer together.

FLUTTER TONE SCIENCE

If we stand at the end of a long, narrow room such as a hallway and clap, we will hear the flutter echo as it returns to us each round trip. If the hall is 20 feet in length, the flutter echo returns after every 40 feet of travel. The time for the round trip is controlled by the speed of sound. In this example, the sound of the clap makes a round trip some 1130/40 or 28 times a second, which sounds like the note of 28 Hz, a half octave below the lowest note of the piano keyboard. However, if we stand in the middle of the room and clap, we hear a different flutter tone. In this situation, part of the clap sound travels towards each end wall. Being in the middle means that each end wall is only ten feet away. Both sounds return to us after only 20 feet of travel. They pass by and head off towards the opposite wall, only to return to us after another 20 feet of travel. This situation produces a flutter tone of 1130/20 or 57 Hz, a full octave above the basic flutter tone of the hall.

If we were really doing this experiment, we would quickly find that we must stand to the side of the hall so as to let the two end walls have a clear view of each other. If we stand in the center of the hall, the flutter is quickly damped out because of the absorption of our body. In this position, with our back to the side wall, sound travels away from the clap equally in both directions, up the hall and down the hall. When we stand at the midpoint of the hall and clap, the two wave fronts race towards the two end walls, reach them and reflect back to soon pass by the clapper at the same time. These two pulses, having arrived at the same time, are heard as one loud pulse. Positions non reversed, the two pulses race for the opposite far walk, and again repeat the course. For this position, the double-strength pulses are heard every time they make half of a full round trip of the hall.

Eugene, Oregon USA

This paper first presented at the

81st Convention 1986 November 13 - 16

DOWNLOAD PDF

Room Acoustics and Low Frequency Damping

by

Arthur M. Noxon

Acoustic Sciences Corporation

Room Acoustics and Low Frequency Damping

The quality, “Q,” of a resonant system identifies its response characteristic. High-Q systems are sharply resonant. They are easy to drive and have a strong response at the resonant frequency (Fo). Low-Q systems respond less strongly and over an extended frequency range. A flat response system has zero Q.

Introduction (not presented but provided for clarity)

The Sabine type formulas of decay rates are derived for diffuse sound fields. This restricts their use typically to 300 Hz and above. Standing wave modes dominate the lower frequency range form of acoustic energy storage. Dissipation of this energy from the room occurs in two forms: transmission out of the room and absorption within the room.

Rooms used for acoustic work frequently have heavier than usual walls to increase isolation from exterior noise. This results in less opportunity for transmission type of energy loss from the room which increases its dependence on internal acoustic absorption to provide sufficient decay rates.

Absorption of acoustic energy is by means of friction effects applied to kinetic energy components of the sound waves. This friction is usually “wall friction,” where the reflecting wave is locally transformed by the stiff and heavy wall impedance. The surface normal component of the waves’ kinetic energy density converts to extra pressure and the tangential component is exposed to opportunities for surface frictional dissipation.

There are three types of low frequency wave containment in a room: Longitudinal, tangential and oblique. The decay rates of these are not the same. The longitudinal modes are one dimensional, axial standing waves and present the lowest amount of kinetic energy density to the wall surfaces, hence they have the longest decay rates. The tangential modes impact two pairs of wall surfaces and the oblique impacts all three pairs of walls. The tangential and oblique modes produce about twice the decay rate as the longitudinal mode because their grazing impact on wall surfaces provides for more wall friction. Sabine type equations also account for this type of activity.

Bass traps are discrete devices as contrasted with a wall surface. Their performance depends on their placement relative to the energy distribution of the various modes of vibration. At a particular location, the trap may provide significant absorption at one frequency, and minimal absorption at another. Traps located in the tri corners of a room contact pressure fluctuations associated with each room resonance.

Corner loaded bass traps pull energy out of the standing wave with each pressure change that occurs. Low frequency presents pressure changes at a slower rate than would be by a higher frequency. Calculations of decay rates that are based on this understanding are derived by distributing the energy in the room into the number of pressure zones that exist for the particular mode, then dissipating a fraction of that energy each half cycle, depending on the number of traps located in these pressure zones.

This new method of calculation predicts the number and frequency response of the bass traps required to attain specified decay rate frequency response of a room. Calculation and measurements in test chambers are found to agree. For example, a 2000 ft3 chamber with each of its 8 tri corners loaded with an efficient bass trap produces an RT-60 of 0.3 seconds at 113 Hz.

The formula developed to handle this viewpoint decay rates includes a term which counts the number of fluctuating pressure zones in a room. Its appearance is very similar to the equation that predicts modal density. Another curious effect noticed with very efficient bass traps is the saturation effect of absorption. Decay rates are proportional to the amount of absorption in a corner, but they become less sensitive with higher absorption and reach a limit, indicating that a finite rate of energy can be withdrawn from a resonant field, i.e., no more than all the energy contained in the half wave length held by the corner can be extracted per half cycle, in spite of the “amount” of absorption available.

One of the first things the novice acoustician does upon entering a room is to deliver a sharp clap of the hands. This is followed by a grave shake of the head and comments about how bad the room sounds. Next comes a proposition to fix the room and the fee. The unsuspecting client then administers a sharp hand clap, nods the head in agreement, and gives the guru a retainer. The only problem here is that these people are busy buying and selling modifications to the sound of their own hand clap. We don't listen to a speaker while holding it in our hands, yet we can be tempted to consider acoustics based on the sound of our own hand clap.

Home theater audio systems have an ambience channel. It usually delivers a bandwidth-limited (no bass), mono signal to a pair of speakers that have been mounted high on the wall and to the side of the listener. If you stand on a chair and clap your hands in the location of the ambience speaker, you will hear a very funny and undesirable sound effect. Is this really something we can hear? If so, do we want to listen to this sound effect or provide it to our clients? If not, there might be something we can do about flutter echo colorations.

THE ACOUSTIC CLAP TEST

On a practical basis, the only time that a self-administered and self-audited hand clap is directly relevant to anything in audio is when the recording engineer is setting up mikes in a studio. Only in this special circumstance does the desired audio signal leave from and return to the same place. Listening to one's own hand clap duplicates this round trip, acoustic process and thereby is a relevant test. If someone ever wants to know how a loudspeaker sounds to the listener, a different technique must be followed, one that mimics the actual speaker/listener acoustical path.

The frequency response curve of a speaker may be flat from 20-20,000 Hz in the test chamber, a room without reflections. Place the speaker in a real room with a microphone at

the listening position. Measure again the response. A series of peaks and valleys are recorded. Move the speaker or mic and a different curve is developed. A room has many resonant frequencies. Which of them are stimulated is dependent on speaker placement. Each peak and null in the spectrum identifies a resonant condition.

Any physical resonance will have a pressure distribution in space. The microphone at a pressure peak will register a strong signal. Move the mic ¼ wavelength to a node and no signal is received. In either case resonance is evident.

Definitions of “Q”

The “Q” of a system can be measured from its frequency response curve. The ratio of the resonance center frequency to the bandwidth that accompanies the ½ power or 3 dB down point comprises one definition of the “Q” of a system.

Usually room response curves are presented dB vs. log frequency format. Resonances occur at different center frequencies. If the “Q” is the same, the response curve shape is the same no matter which center frequency is chosen. The “Q” of an average room lies between 10 and 40. The “QP of a free piano string is 1000.

Resonant systems with slight resistance have High-Q responses. Add energy dissipations (resistance) to lower the "Q". Another definition of "Q" is 2pi times the ratio of the energy of the system to the energy lost per cycle.

Decay Relations

Ordinary resonances decay out following an exponential curve in time. The time constant (T) of the decay is the time required for the system to drop to 1/e of the original energy level.

The exponential decay equation can be used to develop the definition of “Q” for the system. If the exponent is a small fraction, less than 1/10, then a simple approximation arises. “Q” equals 2p times the resonant frequency times the decay constant.

The traditional presentation of decay measurements is the RT60; the time required for the energy to drop 60 dB. The exponential curve appears as a straight line in its dB vs. time plot.

By combining the dB level version of energy with the exponential version, the RT60 is resolved to be 13.8 times the decay constant.

“Q” and Decay Constants

The resonance response Q can be expressed in the traditional measure of decay, RT60. It is developed by combining the lightly damped Q relations with the RT60 decay constant relationship.

The result of the previous analysis is the linear relationship between the resonant frequency of a listening room and its “Q” for a fixed RT60. For example, a room may well have an RT60 of 1 second at a resonant frequency of 90 Hz. This means that the room has a “Q” of 50 for that resonance. A current spec for listening rooms is an RT60 of .5 seconds. If this applies to room resonance modes, their “Q” varies from 5 to 100 in the 20 to 400 Hz range.

Resonant Bandwidth Relations

The “Q” of the resonant mode is linear with frequency for a constant RT60. By referring to the half power bandwidth relationship, the bandwidth is definable in terms of RT60. For a constant RT60 the bandwidth is constant.

The frequency response of a listening room can be taken with a linear frequency sweep. This will show the fixed bandwidth resonances to have the same shape regardless of center frequency.

The initial RT60 of the room is .73 seconds. The additional absorption added is sufficient to establish alone in the room an RT60 of 1.1 seconds. The result of the total absorption produces an RT60 of .44 seconds.

In order to provide the correction (dQ), a fraction of total energy (F) must be removed from the resonant mode each cycle. The Sabine type equations do not apply here. They are based on absorptive surfaces exposed to diffuse sound fields and are valid above 300 Hz. Here is low frequency absorption and it is related to the volume and position of the absorption relative to that of the standing wave.

If it is determined that the ”Q” of some mode needs to be reduced, the proper resistance needs to be added. The energy relations for “Q” yield the required (dQ) addition based on initial Qi and final Qf values.

Example

The bandwidth of the 100 Hz room resonance mode may be found to be 3 Hz giving an initial Qi of 33. The desirable bandwidth might be 5 Hz for a “Q” of 20. The correction required has a strength of 50. It is developed by adding the proper amount of absorption to the resonant mode.

Resonant Decay by Discrete Absorption

A basic view of energy absorption allows a fraction (F) of the energy remaining in a system to be removed at a regular rate (1/N times a second). This leads to the exponential decay relations whose “RT60” expression is well known. If the fraction is less than 20%, the system is “lightly damped,” and the log term can be simplified in approximation.

The decay equation is very general. It remains only to define the rate and fraction of energy absorption for any particular system and the RT60 can be predicted.

One Dimension Resonance Decay

The “Impedance Tube” provides a device in which standing waves can be generated and then their decay monitored. The absorption device is located at one end of a tube while the sound source is at the other.

Work is done at the absorption each time there is excess pressure. This occurs twice each cycle, once when the pressure goes positive, and then again when it goes negative. The rate of absorption is twice the resonant frequency.

The fraction of energy lost by each absorption depends on the position and number of traps in the resonant field. A trap located at one end of the impedance tube (A) experiences pressure pulses and can absorb energy. The same trap located at a pressure node (B) experiences no pressure change and does no work.

The single trap at the end of the tube has access to one-half the total energy in the tube. There are two pressure zones, ¼ wavelength in size for the first harmonic.

The second harmonic has its energy split amongst four ¼ wavelength zones. The trap has access to only ¼ the total energy stored in the resonant condition.

The third harmonic has six discrete pressure zones. The trap only works 1/6 of the total energy in the field. The relative size of the trap to the zone increases with higher mode (j) numbers, so its efficiency increases.

Multiple traps in a resonant field increase the fraction of energy removed each pressure pulse. Two properly placed traps in the third mode or harmonic has access to 2/6 or 1/3 of the system’s energy.

The total number of ¼ wavelength pressure zones is twice the mode number. The fraction of energy lost per pressure pulse is the ratio of trapped zones (J) to the total number of zones (2L) times an efficiency term.

The RT60 equation can be written for one dimension trapping. For small absorption, the approximation is made.

The simple Sabine decay formula for one dimension is a classic derivation. A pulse is injected into the impedance tube. Absorption is located at the tube end. The fraction of energy lost upon impact is the absorption coefficient (a).

The PZT decay formula can be converted into a form like the Sabine. Any frequency of resonance belongs to one of a harmonic series. It is the multiple of the mode number (L) and fundamental frequency (fo). Since absorption is only at one end of the tube for both cases, only one pressure zone is trapped.

The efficiency term (n) in PZT analysis and the absorption coefficient (a) in Sabine calculations have the same physical definition. It is the ratio of energy lost to initial energy. For the one dimension systems, PZT rationale results in the same conclusion as does the classic Sabine analysis.

Two Dimensional Decay Rates

The two dimensional physical space is outlined by an X and Y dimension. Each resonant mode is identified by a “mode number,” a set of two whole numbers (L,M). If one of the mode numbers is zero, the one dimensional model develops.

The standard equation for the frequency of a resonant mode has two components. They can be converted into wave numbers by dividing each mode number by its associated physical length. The mode frequency equation can be rewritten in terms of wave numbers.

The primitive cell in two dimensions is the (1,1) mode. Positive pressure in opposite corners with negative pressure in the other two marks the energy distribution at one moment. A half cycle later the polarity reverses. Between these moments are complimentary patterns of kinetic energy distribution.

There are a total of 4 quarter wavelength zones in the pressure distribution of the primitive cell. They are in the corners. All the energy in the resonant cell is found within these four zones twice each cycle. 80% of a zone is found contained within the radius, 1/6 of the wavelength from the corner.

Higher mode numbers are simply more such cells packed into the same space. A (2,1) mode has two cells in the X axis and one cell in the Y. A (2,2) mode is two cells wide by two cells high. The total number of cells is the product of the two mode numbers.

The total number of pressure zones (K) will be four times the number of cells in a mode. If some number (J) of them are absorptively trapped, the fraction of pressure zones trapped is known if the efficiency term is included.

The RT60 formula derived for PZT methods is general and can be applied to this two dimensional case. For light absorption, a further simplification results.

Three Dimensional Modes

The three dimensional model of Pressure Zone Trapping also has a primitive cell, (1,1,1). It has eight corners, each containing a quarter wavelength pressure zone. If all eight zones were placed together a complete sphere would be formed.

Harmonics of the fundamental are built in terms of complete cells. The (1,1,2) will be one cell high, one cell wide, and two cells deep. It will have 8 x 2 or 16 pressure zones. The (1,1,3) mode is one by one by three cells in configuration and has 8 x 3 or 24 pressure zones. The (2,2,2) mode is accordingly two by two by two cells for a total of eight and 8 x 8 or 64 pressure zones. The total number of pressure zones for any (L,M,N) mode is 8(LMN). They momentarily hold all the energy of the resonant field two times per cycle for any standing wave mode in a three dimensional field.

The basic Pressure Zone Trapping formula still applies. The more complicated term for frequency, well known and dependent on three terms, can be substituted. The value for absorption coefficient remains the fraction of energy absorbed per absorption event. It is the fraction of trapped zones times the efficiency term.

The formal RT60 equation can be simplified if the absorption coefficient is less than 1/5 by approximation. The complete RT60 equation is written by substituting terms for frequency and fraction of energy. This formal equation can be simplified if the absorption coefficient (F) is less than 1/5 in the log term.

The RT60 equation can be further developed. The room volume (Vr) term is introduced which converts the three mode numbers into wave numbers.

Wave Number Space

Wave number space is a three dimensional coordinate system with A, B, and C axes. Each point (P) in this space defines a resonant mode for the room. This is not a continuous field space. It is more like a crystal; discrete points set apart at specific distances.

The mode point is at the tip of the resultant vector (D) whose magnitude is the sum of the squares of the components. It is also at the far corner of a rectangle whose volume (V) is known by the products of its components.

The frequency and RT60 formulas can be rewritten in terms of this wave number space geometry.

This listening room already has a decay time. Frequently improvement in the decay rate is desired. The minimum upgrade is to trap one zone for each 500 cubic feet of room volume. The resulting RT60 is a simple expression but is only valid for an absolutely rigid room whose only absorption is due to the trapped zones.

Example

Consider a room 18 by 24 by 8 feet high. We can look at mode (2,2,1). The wave numbers (1/9, 1/12, 1/8) are easily calculated along with the volume and diagonal wave number in space. The decay time for that mode is 0.3 sec. This assumes one 100% efficient absorption device per 500 cubic feet of room volume.

How Many Traps

The efficiency term (n) is defined as the ratio of energy absorbed to the energy presented. The ¼ wavelength pressure zone contains a discrete quantity of energy in a definable volume. The trap occupies part of that quadrant with its own volume (V). 80% of the zone’s energy lies within 1/6 wavelength radius from the corner. The ratio of PZT volume to the 1/8 spherical section volume comprises the geometric efficiency (E). This is further reduced by the mechanical efficiency of the trap (a) itself; typically 50%.

The RT60 equation can be fitted with this efficiency term. Additional substitutions and reductions provide the RT60 to have an inverse frequency dependency. Recall the Sabine equations to not be directly frequency dependent. There appears the dimensionless ratio in wave number space of the modal volume to the cubed modal length. This ratio is largest for symmetric modes (1, 1, 7) or (2, 2, 2) and smallest for the eccentric modes as (1, 2, 6). It is always less than unity and a mean value of 1/3 is chosen.

The use of traps sufficient to remedy a room’s poor low end ranges from one trap per 500 cubic feet to one trap per 250 cubic feet of room volume. This simplifies further the RT60 equation. The trap volume can be resolved for the 500 cubic foot ratio to be inversely dependent on both RT60 and frequency.

The typical acoustic efficiency is 50% for these three commercial traps. Their volume levels cross extended through the frequency range call out the RT60 vs. frequency plot for the 250 cubic foot or 500 cubic foot rate. For example, a 4 cubic foot trap provides 2 seconds at 20 Hz, 1 second at 50 Hz and ½ second at 90 Hz RT60 times.

Conversely, for a particular resonant frequency, room volume and required RT60, the number (J) of trapped volumes can be calculated.

Example

A room of 2,000 cubic feet needs an RT60 of 1/2 second at 50 Hz and tubes having a volume of 4 cubic feet each will be used. A total of 7 traps must be placed in the pressure zones of that mode resonance.

By utilizing PZT methods, an absorptive treatment for low frequency resonance can be specified. The (dQ) change in room Q is easily approximated. The volume (Vt) of traps required to produce that change can also be defined.

Example

The 2,000 cubic foot room needed a Q adjustment of 50. The volume of PZT adjustment is 12 cubic feet.

The listening room is the last link in the audio chain. It is an acoustic coupler loaded with resonances. Hundreds of rooms have been developed into satisfactory listening environments by using the 500 cubic feet per trap rule. The average trap volume is 2.5 cubic feet. A correction in Quality of 60 is what the average acoustic treatment produces. Serious listening rooms usually require a correction in Quality of 30. This means the average (Q=40) listening room must have its Q cut in half and a serious room must have a Q equal to 1/3 its untreated Q.

A frequently asked question involves the number of traps required to reduce an existing RT60. PZT allows the answer without resorting to Sabine formulas.

Karma-Audio.dk

Examples

A 2000 cubic foot room has an RT60 of 1.3 sec. at 50 Hz. We wish to reduce it to 0.7 sec. using 4 cubic foot traps. Calculations show 4.4 traps will lower the RT60 as required.

A 2000 cubic foot soft room with an RT60 of 0.5 seconds needs to be reduced to 0.3 seconds. Using 4 cubic foot traps, calculations show 9 are needed.

If RT60 equipment is not available, a slow sine sweep frequency response will suffice. Measure the 3 dB down bandwidth dF. Substitute its relation for initial RT60. The desired RT60 is often specified and doesn’t need conversion to final bandwidth.

Reverb Chamber

Absorption is usually measured in reverb chambers using RT60 values and the Sabine absorption formula. PZT equations can be rearranged into the same format. The distinctive frequency dependence of PZT absorption is clear. This relation connects standard Sabine lab methods to PZT theory.

Conclusion

A listening room does not have an acoustically flat response. Most rooms can play better when their Q is reduced by a factor of 2 or 3. Room color is damped out from the listening ambience. It is the Q not the EQ that distinguishes the listening room from a standard room. Pink noise is an appropriate test signal for EQ settings. Pure tone, not 1/3 octave sweeps or RT60 are required to monitor the room Q.

The Pressure Zone Trap (PZT) approach provides a rational view of discrete absorptive devices in the resonant field. It allows specifications to reduce the RT60, or Q of the room to acceptable levels.

http://www.decware.com/paper43.htm

Der er omkring 45dB "støj"- niveau i en alm. stue.

AKUSTIK

er i denne sammenhæng den lydmæssige oplevelse, man får af

et rum. Føles det rungende (som et kirkerum), ”tørt” (som en

dagligstue), kan man forstå tale i rummet, er der ekkovirkninger?

Et akustisk godt rum er karakteriseret ved en velafbalanceret

akustik, der er afpasset efter rummet og dets anvendelser.

Dette betyder:

at efterklangstiden (så vidt muligt) er den samme i hele

at der er en jævn lydfordeling henover lytteområdet

at baggrundsstøjniveauet er tilstrækkeligt lavt, samtidig med

at lydstyrken af nyttelyden (fx taleren) er tilstrækkelig høj.

toneområdet at der ikke forekommer ekkovirkninger

 i rummet, altså tydelige enkeltrefleksioner.

EFTERKLANGSTID

er den vigtigste størrelse til at beskrive et rums akustik. Groft

sagt er efterklangstiden den tid, det tager lyden at dø ud i rummet.

Hvis du klapper i hænderne eller på anden måde fyrer en

lydimpuls (et skud) af i rummet, vil du fornemme, at det tager en

vis tid, før lyden er væk. Hvis du gør det i en kirke, kan du tydeligt

høre lyden klinge ud i rummet. Den optimale efterklangstid

afhænger af, hvor stort rummet er, og hvad rummet skal bruges

til. For de fleste rum vil efterklangstiden normalt være forskellig

for de forskellige toneområder. Generelt er dybe toner (bassen)

længere tid om at dø ud end lyse toner (diskanten). I et

akustisk godt rum er efterklangstiden nogenlunde den samme i

hele toneområdet. Det er det, der tilstræbes

 ved den akustiske regulering.

dB (DECIBEL)

er måleenhed for lydniveau. Bruges til at angive lydstyrken. 0 dB

er lagt (lidt løseligt) ved den svageste lyd, mennesket er i stand

til at opfatte. Se lydbarometeret for eksempler på forskellige lydniveauer.

Forskellen mellem to lydniveauer udtrykkes også i dB:

er der fx et støjniveau på 100 dB på den ene side af en væg, og

et resulterende støjniveau på den anden side af væggen på 60

dB, så er dæmpningen gennem væggen altså 100-60, dvs. 40 dB.

Væggens lydisolation er dermed 40 dB.

STØJ (AKUSTISK)

er den tilfældige sammenblanding af lyde, som kan virke generende

eller skadelige. Der kan være tale om trafikstøj, maskinstøj,

en dryppende vandhane, nabostøj osv. Mere generelt taler

man om uønsket lyd. Støj har stor betydning for vores velbefindende,

men også i forbindelse med definition af god akustik, idet

taleforståeligheden i et rum kan nedsættes, hvis baggrundsstøjen

er for kraftig. Visse rum, som teatre, kræver meget lav baggrundsstøj

for at fungere, ellers kan publikum enten ikke høre

eller kun vanskeligt forstå skuespillerne.

TALEFORSTÅELIGHED

er som navnet siger et mål for, hvor godt talen opfattes enten i

et rum med taler og lytter, eller via et lydsystem inklusive mikrofon,

forstærker og højttaler(e). Der findes forskellige metoder til

at angive taleforståeligheden. En af de mest anvendte er Speech

Transmission Index, STI, som er en måle/beregningsmetode, hvor

en række parametre for rummet indgår. Det interessante i denne

forbindelse er, at der ved hjælp af STI kan opstilles krav til taleforståelighed

i et rum eller for et lydsystem, således at man som

bygherre får sikkerhed for, at kvaliteten er i orden, når projektet

er færdigt. Kravene til STI vil variere efter den aktuelle situation,

men der er stor erfaring at trække på, når de aktuelle STI-krav

skal opstilles. Efter projektets færdiggørelse kan STI direkte måles

og sammenholdes med de opstillede krav. STI-værdierne vil

ligge mellem 0 og 1, med 0 for den helt uacceptabelt dårlige

taleforståelighed op til 1 for det perfekte anlæg.

Der findes en ”hurtig” version af STI til direkte måling af taleforståelighed

på et eksisterende system. Metoden kaldes STI-Pa

(”STI for PA systemer”) og betjener sig af et bestemt moduleret

støjsignal. Signalet tilføres lydsystemet, og STI-Pa værdien kan

direkte aflæses på et specielt måleinstrument. Også her vil værdierne

ligge mellem 0 og 1.

Det bør bemærkes, at taleforståelighedsmål som AI, SIL, ALcons

og RASTI ikke længere kan betragtes som tidssvarende.

UDARBEJDET AF

EDDY BØGH BRIXEN

EBB-consult www.ebb-consult.com

JAN VOETMANN

VOETMANN · AKUSTIK www.akusikjav.com

http://www.houseofhearing.dk/index.htm

Karma-Audio.dk

Marts 2006

Danmark er stor på "lydkortet"

Danmark har gennem mere end 100 år fostret iderige og foretagsomme enkeltpersoner, der har skabt mindre eller større virksomheder på lydområdet. I dag påstås det, at 80% af al musik vi hører er dansk signalbehandlet.

Eddy Bøgh Brixen

Det kan godt være, at Danmark er et lille land,

men set i en lydmæssig betragtning er størrelsen ingen hindring.

Der har gennem tiden været mange danske hædersfolk,

 som har bidraget til de teknologier, som er vores grundlag i dag.

Historien

Den hurtige version af historien:

HC Ørsted opdagede elektromagnetismen (1820),

Valdemar Poulsen skabte båndoptageren (1898) og buesenderen,

Peter L. Jensen skabte verdens første PA anlæg, Magnavox (1915),

 Axel Petersen og Arnold Poulsen demonstrerede verdens første tonefilm med optisk lyd (1923),

 Holger Lauridsen udtænkte bl.a. MS mikrofonen baseret på det akustiske goniometer (ca. 1951).

 Hertil kom så folk, der forstod at producere:

Fabrikanter for radio & TV (Bang & Olufsen, Linnet og Laursen, Larsen og Høedholt, m.fl.),

 fabrikanterne af akustisk måleudstyr (Brüel & Kjær, Radiometer m.fl.)

og senere alle høreapparatfabrikanterne (Oticon, Danavox (GN Resound) og Widex),

 for slet ikke at tale om det overvældende antal højttalerfabrikanter (Peerless, Vifa, Scanspeak, Dynaudio, Jamo, Dali, m.fl.).

 Nye områder er nu opstået, med udvikling inden for mobiltelefonien (f.eks. Nokia) og med udvikling af software (IO Interactive, Odeon, Loudsoft, m.fl.)

Eddy Bøgh Brixen

Udviklingchef Jan Abildgaard Pedersen fremviser Lyngdorf Audios nye højttalersystem bestående af en mid/high dipol og en "boudary woofer" der placeres utraditionelt i et af rummets hjørner..

Eddy Bøgh Brixen

Oppe i Skive bor firmaet Lyngdorf Audio. Navnet er kendt af alle, der har beskæftiget sig bare den mindste smule med HiFi, for en vis Peter Lyngdorf er manden bag bl.a. HiFi klubben, der har passeret sin 27 års fødselsdag, og bag forskellige brands som Dali og TacT. Nu har han lagt eget navn til en højteknologisk virksomhed, der allerede er kendt og anerkendt i HiFi miljøet, og som snart også vil være det inden for professionel audio.

Digital rumkorrektion

Det er ikke nogen hemmelighed, at der ofte har været dybe grøfter mellem HiFi-freaks og professionelle lydfolk. Set fra den professionelle side er HiFi i bedste fald betragtet som eksotisk fejldisponering, f.eks. noget med at anvende gigadyre monstrøse kabler til højttalere opstillet i håbløs akustik.

Peter Lyngdorfs tanker om at kunne foretage en digital optimering af lyden i rummet har eksisteret i mange år. På et tidspunkt var han indehaver af NAD fabrikken og den amerikanske højttalerproducent Snell Acoustics for at kunne kombinere sin viden med de apparatproducerende virksomheders ekspertise. De første rumkorrektionsalgoritmer blev udviklet i samarbejdet mellem dem og eksterne DSP (Digital Signal Processing) konsulenter, og udviklingen af det første produkt påbegyndt. Det nåede også at komme på gadem, under navnet NAD RCS2.2, men det blev aldrig særlig udbredt.

Senere blev NAD og Snell Acoustics solgt igen, men de akustiske ideer levede videre i firmaet TacT, der blev oprettet i USA. Ud over produktionen af en digital rumkorrektion kom sortimentet til at omfatte digitale effektforstærkere baseret på den viden, som Lars Risbo udviklede i Toccata Technology, der senere blev opkøbt af Texas Instruments.

Hjælp fra Lyngdorf selv

Rumkorrektionen var rimelig, men det optimale opnåede man ikke helt automatisk. Ofte var det faktisk nødvendigt med lidt ekstra konsulentbistand helst fra Peter Lyngdorf selv for at få et godt resultat.

I 2003 trak Peter Lyngdorf sig ud af det amerikanske TacT. Teknologien til digitale forstærkere var i hus, men der var behov for nyt med hensyn til rumkorrektionen - ikke skabt af DSP folk, men af akustikere. Jes Mosgaard blev derfor ansat som administrerende direktør og var også medvirkende til at et nyt firma fik Lyngdorfs eget navn, Lyngdorf Audio.

Rumpilot nr. 1

Forskning og udvikling blev lagt i hænderne på Jan Abildgaard Pedersen, ansat 1. oktober 2005 med medarbejdernummer 8. I dag to år senere har virksomheden 49 ansatte, og man er flyttet til større faciliteter fire gange i løbet af samme periode.

Jan Abildgaard har bl.a. en fortid hos B&O, hvor han stod for udviklingen af ABC systemet (Automatic Bass Control) til B&O's superhit, Beolab 5. Men han var bestemt ikke løbet tør for ideer med det.

Allerede den 5. december 2005 kunne han internt præsentere det, der nu har fået navnet "RoomPerfect". Patentansøgningen blev indgivet 3. januar 2006, og Jan Abildgaard sad faktisk den dag i Gatwick lufthavn og ventede på grønt lys til stige på flyet til USA for at vise systemet på CES i Las Vegas. Han måtte nemlig ikke tage afsted før han havde sikkerhed for at patentansøgningen var indgivet. Her i august 2007 er samme patent så endelig blevet offentliggjort.

RoomPerfect

Men hvad er RoomPerfect egentlig?

Grundlæggende er det en metode til at opnå en perfekt klangbalance for et givet sæt højttalere i et givet lytterum. Man taler om "elektronisk rumkorrektion", og det betyder ikke, som man måske kunne tro, at man korrigerer rummet, men at man så at sige aflæser den indflydelse rummet har på lydbilledet - og tager sine forholdsregler via korrektion i forstærkeren.

I "gamle dage" korrigerede man alene ud fra en måling i lyttepositionen. Dette medførte ofte helt vilde filtreringer, og det kunne derfor lyde rædselsfuldt alle andre steder.

RoomPerfect ser på den energi lydeffekten der bliver postet ind i rummet. Der måles med en mikrofon i et antal positioner. Resultatet giver herefter input til den nødvendige korrektion. Der er sat grænser for hvor meget de involverede filtre får lov til at korrigere for at undgå at dele af systemet drives ud over kanten. Filtrenes båndbredde og øvrige karakteristikker er også valgt under passende hensyn til den psykoakustiske viden: Hvad kan vi høre og hvad er det, vi ikke kan høre.

Resultatet er den optimale korrektion i den givne situation. Man skal så lige vide, at lydkvaliteten øges i takt med at højttalere og rum også har større kvalitet. Man gør ikke en Trabant til Mercedes ved hjælp af lidt elektronik. Ikke engang digital...

RoomPerfect er nu implementeret i produkterne fra Lyngdorf Audio, men ikke kun her. OEM aftaler har sikret, at teknikken (og den tilhørende hardware) kommer til at indgå i andres produkter. Et væsentligt samarbejde er bl.a. indgået med Steinway & Sons. Før var det flygler. Det er det stadig, men nu er det også absolut high-end HiFi - og her har Lyngdorf mere end en finger med i spillet.

Produkter

Flere utraditionelle ideer er kommet til. Det seneste er en højttalerkonfiguration med to subwoofere og to dipolhøjttalere. Det ser meget anderledes ud en det man er vant til, for basserne skal stå op ad væggen. Fordel: Ingen refleksioner fra gulv og nærmeste vægge. Ulempen, at der sættes gang i de stående bølger, bliver elimineret af RoomPerfect. Og dipolhøjttalere: Ingen kabinetrefleksioner og god kobling til rummet.

På forstærkersiden er der allerede flere flagskibe, f.eks. TDAI 2200 (det er en helt ingeniørmæssig betegnelse: Total Digital Amplifier Integrated, 2 x 200 watt).

Et andet produkt er intet mindre end verdens bedste konverter, Millenium ADC. Den har bl.a. dobbelt mono konvertering for at udbalancere støj. Der er 40 forskellige RIAA kurver indbygget. Det betyder at man kan genfinde den korrektionskurve, der blev anvendt på det label i netop den periode. Der er tale om en frontende, som da må være et must for alle vinylelskere og ikke mindst for alle, der arbejder inden for archiving, hvor der skal overspilles og sikres gamle optagelser.

Selv en CD afspiller er det blevet til - en konstruktion hvor al god viden er samlet. F.eks. er der hørbart bedre styr på sample-rate-konverteringen i forhold til flertallet af afspillere på markedet.

Fra HiFi stereo til prof surround

I Skive fortsætter udviklingen på højtryk. Man har ansat endnu en akustiker, nemlig Henrik Mortensen, der var chef akustiker for Jamo i 20 år og nu er blevet det for Lyngdorf Audio.

Virksomheden er også blevet international, via opkøbet af finske Kuusama Design. Indehaveren Juha Kuusama er bl.a. en af stifterne bag Sample Rate Systems OY, som har stor kompetence indenfor surround teknologi. Så i nærmeste fremtid introduceres løsninger for surroundgengivelse.

Inden for professionel audio ved man hvor svært det kan være at optimere et surroundsystem så man kan producere lyd på det. Det bliver interessant at få løsninger, der ikke bare anvendes ved reproduktion, med også kan forbedre lytteforholdene ved produktionen af lyd.

Og så er det jo herligt, at der stadig genereres gode ideer i Dannevang og at der er nogen der har evner og midler til at realisere dem.

Karma-Audio.dk

RoomPerfect"

Lyngdorf korrigerer for akustikken

På mindre end to år har Lyngdorf Audio etableret Nordeuropas største akustikafdeling. Nu er "RoomPerfect" en realitet - et system til korrektion af lytterummets indflydelse på højttalerlyden...

Oktober 2007